{ "version": 1, "updatedAt": "2026-07-30T02:36:34.272Z", "source": "cms-exports", "mediumUrl": "https://medium.com/@OVNICap", "articles": [ { "id": "c5577b25b3f7", "position": 10, "visible": true, "name": "We spent three years quietly testing how European Deep Tech should move to the US.", "slug": "we-spent-three-years-quietly-testing-how-european-deep-tech-should-move-to-the-us", "shortDescription": "The dual-core transatlantic scale method, and how to actually use it", "dateLabel": "06/2026", "dateIso": "2026-06-14T02:47:04", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/1*3KESoIGCSnU9tiiZlMwNsQ.png", "bodyHtml": "

We spent three years quietly testing how European Deep Tech should move to the US. Here is our system.

\"\"

The dual-core transatlantic scale method, and how to actually use it

If you’re in en entrepreneur in a rush, you can try download a working excel version here.

“Dual-core” is the idea that a company can run on two engines at once rather than picking one side of the Atlantic. That question, which side a company should belong to, is suddenly everywhere. That’s because everyone is talking about sovereignty right now, whether in Europe or in the US. The instinct is right but Europe cannot think it can legislate sovereignty into existence or it will be wane into nonrelevance. Rather european sovereignty will be won by building a superior product. And at the moment, a superior product can only be created in the most competitive market on earth, which is the US. That’s the tough reality very few want to hear. And that’s why European deep tech must cross the Atlantic, now more than ever.

When we started OVNI in 2022, we had a thesis we did not want to say out loud: most European deep tech fails on the timing and choreography of its move to the US, not on its science. Companies either cross the Atlantic too early and bleed out, or too late and miss the fundraise window entirely.

We built a system to fix that. We kept it internal for 3 years because we are not consulting experts and we wanted to see if it survives contact with real fundraising dynamics, real founders, and real US funds before we talked publicly about it.

Since moving to San Francisco and opening our US office, we ran the method in the field rather than on a whiteboard. So I am ready to share it online from an entrepreneur’s point of view. What follows is the working version: the scorecards you can fill in and the thresholds you can measure. If you run a frontier company in Europe, you should be able to score yourself before you finish reading (here).

The structural problem, stated plainly

There is a division of labor across the Atlantic that almost no one names directly.

Europe creates. Frontier engineering, applied science, dense technical talent at a lower burn (tech wages for similar tier-1 profiles are lower say many founders), IP and know-how, deep academic and industrial networks. This is real and it is underrated.

The US decides. Category language and product expectations, design partners and budget owners, go-to-market talent, and the growth capital plus Series A comparables that set the price of your company.

The trap is that these two facts pull a founder in opposite directions, and the failure modes are symmetric:

Most “should we move to the US” advice is binary. Stay or go. That framing is the actual problem. The right unit of decision should be the function instead of solely the company. You cannot just decide to move a startup to the US. You move customer discovery, then pilots, then product and go-to-market, then commercial leadership, each one only when proximity to the customer changes the speed and quality of your learning.

That is the whole idea. Everything below is the machinery that makes it operational. Introducing the “Dual-core transatlantic scale method”.

The method is 3 linked instruments

Here are the 3 tools:

The frontier filter. A selection scorecard to determine whether you should even consider a transatlantic story.

The function shift map. Once you’ve validated the frontier filter, this is an operating sequence for which function moves, and when?

The Series A readiness index. A model to quantify your likelihood of raising with a Tier-1 US fund.

\"\"

The reason it works as a system is that the same logic governs all three. You select for a science-led wedge with a U.S. commercial outcome, you operate by moving functions one at a time, and you raise only when the evidence benchmarks on Silicon Valley terms. Selection, support, and fundraising readiness run on one spine.

Layer 1: The frontier filter (score yourself out of 100)

Before you actively pursue a transatlantic story, score the company on five dimensions. These are the 5 criterias we believe actually matters.

\"\"

Inside each dimension we break the points down further, so the score is auditable rather than a vibe. Scientific wedge, for example, splits into strength of proprietary technology (10), engineering and process complexity (8), and external scientific validation (7). U.S. demand splits into an identified buyer segment with documented pain (7), confirmed budget authority and procurement readiness (7), and timing, meaning a real reason to adopt now rather than someday (6).

The threshold is 75 out of 100, with no structural red flag in productizability or dual-core fit. Those two are non-negotiable because they are the ones that quietly kill these companies. A 90 on science with a 5 on productizability is a lab and not necessarily a startup. A company that can only function if the entire technical team relocates is also not what we could call dual-core.

How to read your own score:

If you score yourself honestly and land below 75, the answer is not “go to the US anyway.” It is “keep building until the number is real.”

If you score above 75, start working with layer 2 on your mind.

Layer 2: The function shift map (the order of operations)

This is the part founders most often get wrong, usually under pressure from a board member who wants to “see US momentum.”

We sequence the move across 6 stages, and we are explicit about what lives where:

\"\"

The sequence is fixed: customer discovery, then pilots, then US product and GTM, then commercial leadership. Product management moves before sales scaling, because you cannot scale a motion you have not yet shaped.

And the one rule that overrides everything: never move core engineering to satisfy optics. Move only the functions whose proximity to customers changes how fast and how well you learn. Engineering proximity to a Paris cleanroom matters more than engineering proximity to a Palo Alto conference room. If a move does not increase learning velocity, it is overhead wearing a costume but most US investors are trained to see the disguise.

Layer 3: The Series A readiness index (“permission” to raise)

A US Series A process is expensive in time, focus, and credibility. You get roughly one clean shot at a given set of funds. As a seed fund, we do not let our founders start until the evidence benchmarks on elite terms.

Five components, scored out of a total of 100:

\"\"

The bands tell you what to do with your score:

Notice that commercial pull is weighted as heavily as technical maturity. For deep tech founders, this is the uncomfortable part. The science being further along does not move you up as much as paid pilots and budget owners would. But it is aligned with deeptech expectations (design partners, co-dev agreements, POC,…)

It is important to note that you can probably raise even if you don’t score an 85 but it may just be difficult or you may need to ride a bigger industry trend. We made this scorecard even higher than layer 1, because you won’t find more selective than SF-based funds.

Six operating rules that govern every move

If you want to put the scorecards to the side for a minute, keep these in mind:

\"\"

This turns expansion from a binary “stay versus move” gamble into a structured OS. The founder is never asked to mimic a Silicon Valley template too early. The company earns each next function, geographically and financially.

Conclusion

This method is built for deep tech with a long fuse and not for generic SaaS that could be incorporated in Delaware on day one and lose nothing. If your moat is execution speed rather than science, the dual-core logic adds cost without adding protection.

It will not save a company whose science does not actually compound, and the scorecard is only as good as the honesty of the person filling it in. A founder who scores their own US demand at 18/20 on the strength of three friendly intro calls is lying to themselves, and the framework cannot stop them.

And it is a sequencing system, not a guarantee. Markets shift as mentioned in layer 3, a buyer’s budget freezes, a competitor raises a megaround. The method improves your odds and your timing. It does not remove risk. Anyone who tells you a framework removes risk is selling you something ;-)

Though this article is written with entrepreneurs in mind, we use the method internally to evaluate potential investments (layer 1), to help recent portcos think through their transatlantic expansion (layer 2), and to estimate alongside the entrepreneurs whether they are ready to raise the subsequent round (layer 3).

Try for yourself -> The dual-core transatlantic scale method

The Transatlantic Bridge, the framework I built at Newfund in 2018, is where this all started. Its conviction was simple and, for the time, contrarian: a European startup should build like an American company from day one rather than treating the US as a graduation prize. In practice that meant a set of deliberate moves. Incorporate in the US before you incorporate in France. Put a founder on the team who was educated or built abroad. Hire early for an English-first culture, often by bringing in someone who doesn’t speak French so the whole company defaults to English. Drop the French-language website and the little UK flag standing in for English. Spend real time in New York or San Francisco. Raise from US funds, so American customers, networks, and benchmarks come baked in. These were forcing functions, and they worked. The founders who adopted them moved faster and stopped treating global ambition as something to earn later.

The Bridge was built to solve a general problem: getting European companies across the Atlantic with conviction and speed. As I moved deeper into deep tech, I saw that the same core bet, build for the US early, needed more specialized machinery to hold up. When the product is a photonic interposer or a quantum control stack, the science, the capital intensity, and the structure of US research and procurement shape the path as much as a founder’s ambition does. The Dual-Core Transatlantic Scale Method is the deep-tech evolution of the Bridge: the same founding conviction, extended with the rigor that frontier technology demands.

The Transatlantic Bridge, the framework I built at Newfund in 2018, is where this all started. Its conviction was simple and, for the time, contrarian: a European startup should build like an American company from day one rather than treating the US as a graduation prize. In practice that meant a set of deliberate moves. Incorporate in the US before you incorporate in France. Put a founder on the team who was educated or built abroad. Hire early for an English-first culture, often by bringing in someone who doesn’t speak French so the whole company defaults to English. Drop the French-language website and the little UK flag standing in for English. Spend real time in New York or San Francisco. Raise from US funds, so American customers, networks, and benchmarks come baked in. These were forcing functions, and they worked. The founders who adopted them moved faster and stopped treating global ambition as something to earn later.

The Bridge was built to solve a general problem: getting European companies across the Atlantic with conviction and speed. As I moved deeper into deep tech, I saw that the same core bet, build for the US early, needed more specialized machinery to hold up. When the product is a photonic interposer or a quantum control stack, the science, the capital intensity, and the structure of US research and procurement shape the path as much as a founder’s ambition does. The Dual-Core Transatlantic Scale Method is the deep-tech evolution of the Bridge: the same founding conviction, extended with the rigor that frontier technology demands.

OVNI Capital specializes in bridging the gap between European deep tech innovation and US market leadership. With offices in San Francisco, Berlin, and Paris, we partner with visionary entrepreneurs to build global category leaders by bringing their breakthrough technologies to the US from day one. Our investment strategy is rooted in systematic co-investments with leading US and European funds and leveraging our extensive network of LPs throughout the two geographies.

\"\"

At OVNI Capital, we don’t just fund companies, we empower them to transcend borders and redefine entire industries.

", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/@augustinsayer/we-spent-three-years-quietly-testing-how-european-deep-tech-should-move-to-the-us-c5577b25b3f7", "href": "/insights/we-spent-three-years-quietly-testing-how-european-deep-tech-should-move-to-the-us", "topic": "Deeptech" }, { "id": "f3b042cfff77", "position": 20, "visible": true, "name": "The hidden fragility of AI-era revenue", "slug": "the-hidden-fragility-of-ai-era-revenue", "shortDescription": "In recent months, we’ve seen multiple AI startups , particularly in video generation, creative tools & consumer-facing apps, explode to millions in ARR through viral demos and rapid adoption.", "dateLabel": "02/2026", "dateIso": "2026-02-20T21:14:51", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/1*aa5JbIIy7M1NgvqHtBXRZQ.jpeg", "bodyHtml": "
\"\"

In recent months, we’ve seen multiple AI startups , particularly in video generation, creative tools & consumer-facing apps, explode to millions in ARR through viral demos and rapid adoption. Only to stall or decline sharply a few weeks/months later. Superior open-source or frontier models quickly commoditize core capabilities, major platforms integrate similar features natively or users switch effortlessly as the market matures and differentiation evaporates. These patterns reveal how easily hype-fueled, low-moat revenue can prove fleeting in the current AI landscape.

Now picture this: a cybersecurity company catering to enterprise clients goes from $100k to $1.5M in a year. That’s a 15x jump. Once upon a time, this would be the signal for any serious Series A fund to lead an outsized round. Today it gets filed under good, not great.

Meanwhile, an AI app can ramp from zero to millions in ARR in months and trigger a feeding frenzy. This is a structural change in how venture is being played and it’s quietly increasing the risk that investors are overpaying for fast, disloyal revenue while underfunding companies with the kind of durable compounding that actually drives DPI.

Why the bar moved (and why it became so binary)

1/ Liquidity pressure is forcing VCs into only obvious winners modeOne of the most under-discussed drivers of today’s behavior is that the venture ecosystem has been running through a long stretch of strained liquidity. PitchBook notes that net cash flows to LPs have been negative by ~$169B since 2022, and that 2025 exit value was projected to remain below $300B. That’s not just macro, it changes partner psychology and the tolerance for anything that doesn’t scream breakout.

\"\"

2/ Fundraising tightened, increasing herding behavior

In the same PitchBook outlook, US venture fundraising through Q3 2025 was ~$45B, described as the lowest since 2017.

When fewer firms are raising easily, more firms are possibly trying to mark their relevance by being attached to the perceived consensus. That’s how you get binary underwriting and consensus piling into the same trade.

3/ AI is absorbing the oxygen by the numbers

Crunchbase data shows AI captured close to 50% of all global startup funding in 2025, up from 34% in 2024, with $202.3B invested in the AI sector in 2025.

Capital concentration does something subtle: it trains investors to treat non-AI as career risk. That’s how you end up with a 15x year being meh.

\"\"

4/ AI also made revenue acceleration easier to manufacture

This matters: in many AI categories, the product can be shipped faster, distribution can be viral, and monetization can be usage-based. That combination can create steep early revenue ramps even when long-term retention hasn’t been proven.

\"\"

Bessemer’s State of AI popularized the the idea that the new elite companies may hit ~$40M ARR in year 1 and ~$125M in year 2. When this becomes the benchmark (even implicitly), everything else looks slow.

The uncomfortable truth: in AI, some of the fastest revenue is disloyal

The core issue isn’t AI is hype. AI is real, valuable, and here to stay. The issue is that a meaningful slice of AI startup revenue behaves less like traditional SaaS and more like rented demand:

easy to start, easy to switch, easy to price-compress, and sometimes costly to serve.

Here are the mechanics behind that disloyalty.

1/ Switching costs are often low because the underlying capabilities are commoditizing fast

When the cost/performance curve moves this quickly, defensibility based on “we built it first” evaporates.

Stanford’s AI Index highlights just how extreme the cost curve has been: the cost to query a model at GPT‑3.5 level dropped from ~$20 per million tokens (Nov 2022) to ~$0.07 per million tokens (Oct 2024) — a roughly 280× decline. That kind of drop doesn’t just help margins. It also lowers the barrier to entry for competitors and accelerates feature parity.

In categories like coding assistants, copilots, summarization, meeting notes, customer support agents, many products share similar primitives. If your differentiation is “we wrapped the best model nicely” then your customers can often migrate when the next model leap happens, a platform vendor bundles the feature, or someone undercuts pricing.

Get Augustin Sayer’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

2/ Some GenAI ARR has weaker margin quality than classic SaaSClassic SaaS has historically enjoyed high gross margins and predictable renewals once embedded. Many AI businesses don’t have that profile yet.

Bessemer points out that many GenAI supernovas have gross margins around ~25% (often negative), even while scaling rapidly. That single data point should change how investors interpret “ARR.” If the cost of serving revenue remains meaningfully variable (inference, tool calls, human-in-the-loop, data licensing), then the revenue is less profitable, pricing pressure hurts more, and the business can’t just “SaaS-multiple” its way into inevitability.

\"\"

Fast growth + low gross margin is not the same asset as fast growth + 75–90% gross margin.

3/ Enterprises are increasingly deciding to build for certain feastures (and especially when the core is commoditized)Menlo’s enterprise AI report shows a notable build vs buy shift: 47% of AI solutions developed in-house vs 53% sourced from vendors, versus 80% relying on third-party GenAI software in 2023. That is a direct warning sign for wrapper revenue: if the value is generic and the primitives are cheap, enterprises will often internalize it particularly when security, privacy, and integration matter.

\"\"

The same report also notes what buyers say they care about most when selecting GenAI tools: easily quantifiable ROI (30%) and customization by org/industry (26%), while cheaper registers at ~1%. My take: the long-term winners are not the cheapest or the flashiest, they will be the ones that become contextual infrastructure inside the enterprise.

4/ Many AI deployments still aren’t producing measurable business impact yetThis is where revenue quality becomes tricky: a lot of early AI spending is experimentation.

\"\"

A widely circulated MIT Project NANDA report (preliminary findings) argues most GenAI initiatives haven’t translated into measurable bottom-line impact and highlights a big gap between pilots and scaled production, while also documenting limitations in methodology and representativeness. Even if you treat the headline numbers skeptically, the directional point matches what many operators see: POCs create spend but only sustained workflow change will yield to contract renewal.

Why VCs should be careful: explosive growth can be a mirage in the AI stack

When investors over-index on exponential ARR, three errors become common:

Error #1:Some AI revenue behaves more like paid acquisition spend (turn it on/off), not like a multi-year embedded contract. If you apply SaaS-style valuation instincts to non-SaaS-like retention, you overpay.

Error #2: A viral loop, a great launch, a model jump, or a platform moment can create spikes. But spikes are not moats.

Error #3:

In categories like cybersecurity, dev infrastructure, data tooling, industrial software, climate/energy, robotics, bio , real deployment is messy. Sales cycles are longer. Integrations are deeper. Procurement is slower. But once a product becomes part of the operating system, the revenue behaves differently.

And the market data still supports this: SaaS Capital’s retention benchmarks show that for higher-ACV businesses, median gross retention is ~93% and top performers can reach ~118–120% net revenue retention at higher ACVs.

\"\"

The contrarian conclusion

Venture has always been about outliers. But in the current cycle, too many investors are chasing the easiest outlier to recognize: explosive AI revenue growth.

The risk is that we’re turning venture into a bad trade: paying peak prices for revenue that’s cheaper to replace every quarter, while starving the companies building the boring, embedded, compounding value that creates real customer lock-in (and real DPI).

In 3–7 years, the scoreboard won’t reward who grew fastest from zero. It will reward who built the revenue that stayed.

", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/@augustinsayer/the-hidden-fragility-of-ai-era-revenue-f3b042cfff77", "href": "/insights/the-hidden-fragility-of-ai-era-revenue", "topic": "AI" }, { "id": "d63aaa0e9613", "position": 30, "visible": true, "name": "Europe’s startups don’t lack talent, they lack a domestic scale engine (for now)", "slug": "europes-startups-don-t-lack-talent-they-lack-a-domestic-scale-engine-for-now", "shortDescription": "Europe can invent. Europe can build. Europe can even create category-defining companies.", "dateLabel": "02/2026", "dateIso": "2026-02-02T23:55:44", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/1*DjJOQsYp-o0SIR6FpscPyA.png", "bodyHtml": "

Europe’s startups don’t lack talent, they lack a domestic scale engine (for now)

\"\"

Europe can invent. Europe can build. Europe can even create category-defining companies.

But absent three structural ingredients,

a genuinely unified market (especially in services),

a more favorable tax-and-incentives environment, and

large, risk-tolerant, asset-backed pension capital

the European Union will remain structurally predisposed to rely on the United States to scale its most ambitious startups.

\"\"

That statement is neither Europe-bashing nor resignation. It’s an attempt to describe the mechanism, using data, and to explain why “going transatlantic” is often the rational equilibrium, even for founders who would prefer to stay entirely European.

And, crucially: in the meantime, transatlantic scaling can still create high-paying jobs in Europe. The value that shifts toward the U.S. is not (only) a moral failure, it’s a market outcome. The U.S. is the place where scale is routinely financed, priced, and exited, because the U.S. built the deepest scale engine.

Europe can build its own. But until it does, the U.S. will remain Europe’s outsourced scaling layer.

1. The scale problem is not “Europe is small.” It’s that Europe is not one market where it matters most.

The EU is often described (correctly) as a massive demand pool: “more than 440 million consumers.” So why doesn’t that translate into U.S.-style scaling?

Because the constraint is not raw population. It’s the effective market a startup can access with one go-to-market motion: one product compliance posture, one set of contracts, one sales playbook, one labor framework, one set of customer expectations, one set of distribution rails.

The mismatch becomes obvious when you look at where Europe’s economy actually lives:

Let’s pair that with the EU’s own integration KPI:

Put it simpler, in 2022, EU countries sourced roughly 26% of their goods from other EU countries, but only about 7.5% of their services. That is not a philosophical problem. It’s a scaling math problem.

Most modern VC-backed companies have, economically speaking, services businesses (software, marketplaces, fintech, digital health, AI tooling, media, logistics layers, etc.). When service trade is far less integrated than goods trade, Europe may be structurally optimized to produce innovation but not to compound it quickly across borders.

One provocative way this fragmentation shows up is via tariff-equivalent estimates of intra-European barriers. The IMF has pointed to internal barriers that can be equivalent to large tariffs. On the order of ~44% for goods and ~110% for services — as a way to communicate how costly fragmentation can be. That exact quantification is debated (and should be treated skeptically).

\"\"

But here’s the key: even if you discount the headline percentages, the direction is not controversial. Services are less integrated than goods, and startups are disproportionately services-like. So founders do what founders always do: they go where the one-market assumption is closer to true.

2. Taxes aren’t just about government size. They are a scaling friction, especially for hiring and equity.

When entrepreneurs ask to lower taxes, it’s easy to mishear it as a political slogan. For startups, it’s usually something more technical:

A single number captures part of that cost: the tax wedge (income tax + employee and employer social contributions, net of benefits) for a single worker at average earnings.

In 2024, the OECD reports:

When Europe’s two largest economies also have some of the highest tax wedges, scaling becomes more capital-intensive from day one: startups must spend meaningfully more to deliver the same net compensation, and the cost of building large teams rises faster than in lower-wedge systems.

\"\"

Two observations matter more than any single country comparison:

The EU does not have one tax wedge; it has a distribution. That means a scaling company faces not just “higher or lower” taxes but variance, plus administrative and legal overhead, as it expands across member states.

\"\"

For scale-ups, labor economics matter earlier than corporate profit taxes. Startups usually don’t pay much corporate income tax while they are loss-making. But they do pay for labor, and labor is the main spend line in knowledge-intensive growth.

This is not an argument against the European social model. It’s an argument that, if Europe wants European scaling, it needs to reduce the marginal friction of scaling: simplify cross-border employment, make equity incentives easier, and avoid turning expansion into a compliance project.

Absent that, the U.S. looks attractive not because founders love America, but because scaling cost is legible inside one system.

3. The deepest reason: Europe is not short of saving, it is short of risk-bearing, long-duration domestic capital.

Here is the under-discussed reality: Europe’s constraint is not talent. It’s balance sheets.

A venture ecosystem does not become self-sustaining at scale until it has three domestic ingredients:

Deep pools of long-duration capital (pensions, insurers, endowments)

Large public markets (or equivalent late-stage private markets)

A coherent set of exits that recycles capital and talent back into the system

The U.S. has built these layers over decades. Europe has built some of them unevenly, and others not at all.

Start with the retirement capital base

The U.S. retirement market is enormous: U.S. retirement assets totaled $48.1 trillion at the end of Q3 2025. In Europe, by contrast, funded occupational pensions remain far smaller at the EU level: European IORPs (Institutions for Occupational Retirement Provision) held €2.69 trillion in investments as of Q4 2024.

\"\"

This gap is not simply cultural, it is structural. Many European countries still rely heavily on pay-as-you-go public pension systems, which are a political choice, but also a capital-formation choice: contributions are largely used to finance current retirees rather than accumulated into large, long-duration investment pools. In 2022, EU pension expenditure was 12.2% of GDP, and 27.1% of the EU population were pension beneficiaries.

Europe is not “too poor” to finance innovation. It has deep savings. But in the absence of large domestic pools of risk-bearing retirement capital consistently allocating to venture and growth, European savings often end up financing innovation indirectly, through U.S. public markets and U.S. corporate equity, where liquidity and scale are unmatched.

The result is not simply capital leaves. It is a more subtle equilibrium: European savers capture a slice of U.S. scaling upside, while European startups often must cross the Atlantic to access the scaling infrastructure that converts innovation into global platforms.

Now connect pensions to venture capital, directly

An EU-commissioned analysis under the StepUp StartUps initiative makes the mechanism explicit:

This is the heart of it: the U.S. has a domestic institutional bid for venture and growth that Europe has not replicated at EU scale.

\"\"

So when a European startup hits the scale-up phase , when rounds become large, timelines long, and the cost of being wrong high, the gravitational pull is predictable.

The transatlantic outcome isn’t just “fundraising.” It’s who captures the scaling premium at exit.

Europe’s scale-ups are not doomed to fail. In many cases, they achieve exits at comparable rates. But the buyer and listing venue matters , that’s where value is often repriced.

The EIB finds that, at exit:

So the issue is not “Europe cannot exit.” It’s who buys:

\"\"

That is the scaling premium migrating: not necessarily the jobs, not necessarily the invention but the marginal unit of value created at scale, which is increasingly priced and captured in the U.S. ecosystem.

This is not a moral indictment. It is the predictable consequence of where the deepest growth capital and exit markets are.

Why this is not unpatriotic and why it can still be good for Europe in the near term

If your goal is “European companies should scale in Europe,” it can feel uncomfortable to admit how structural the reliance on the U.S. is. But denial doesn’t create alternatives. Systems do.

\"\"

A more constructive framing is:

Europe is strong at creating frontier capability: engineering, research, industrial know-how, and increasingly deep tech. The European Commission explicitly highlights assets like “highly qualified engineers” and “excellence in research.”

The U.S. is strong at financing and pricing scale: late-stage equity, mega-funds, liquid public markets, and a unified commercial environment.

Transatlantic scaling can be a division of labor rather than a defeat, especially if Europe keeps R&D, product, and significant operations anchored locally.

It is also worth saying plainly: value capture is not binary. Even when a company scales via U.S. capital markets, Europe can still capture high-skill employment (engineering, product, research, operations), supplier ecosystems, founder liquidity that becomes angel/seed capital locally, and geopolitical leverage through ownership of strategic technologies.

\"\"

At the same time, it is honest to acknowledge the trade-off: the marginal scaling premium, the step from “great company” to “global platform” is more likely to be priced, financed, and exited in the U.S. until Europe builds comparable machinery.

The bottom line: Europe will keep producing breakout startups. The question is where the breakout compounds.

Europe will continue to create exceptional companies. The evidence already shows that.

But unless Europe unifies the market where startups live (services), reduces the marginal friction of hiring and equity at scale, and grows large pools of patient, risk-bearing pension capital, the EU will remain structurally destined to lean on the U.S. for scaling.

That reliance is not shameful. It is reality.

And in the near term, the best founders will do the rational thing: build in Europe, scale through the U.S., and if they can bring the compounding back home.

Because the real goal is not ideological purity. The goal is a European scale engine so strong that transatlantic scaling becomes a choice not an inevitability.

\"\"
", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/@augustinsayer/europes-startups-don-t-lack-talent-they-lack-a-domestic-scale-engine-for-now-d63aaa0e9613", "href": "/insights/europes-startups-don-t-lack-talent-they-lack-a-domestic-scale-engine-for-now", "topic": "Venture" }, { "id": "b505409b99f0", "position": 40, "visible": true, "name": "2035: The Great Reversion?", "slug": "2035-the-great-reversion", "shortDescription": "In 2035, intelligence will be free. Not cheap. Free.", "dateLabel": "12/2025", "dateIso": "2025-12-23T09:28:13", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/1*samT2hu97lYkiU8LvVOjOQ.png", "bodyHtml": "

2035: The Great Reversion?

\"\"

In 2035, intelligence will be free. Not cheap. Free.

When thinking costs nothing, being smart stops mattering. And when efficiency costs nothing, wasting time becomes the ultimate luxury. I often mention on podcast that ten years from now we’ll revert to being Romans.

If you asked a futurist in 2025 what 2035 would look like, many would have described a world of seamless digital integration: humans merged with the cloud, living in the metaverse, worshipping the god of efficiency.

I think the reality will be far more human than anyone expected. The collapse of Schumpeter’s Law and the arrival of zero-marginal-cost intelligence won’t just change our economy. It will invert our social hierarchy. The great mistake of the 2020s was thinking AI would make us less human. It is doing the opposite.

Here are my five simple, contrarian takes on what the Western world could actually look like a decade from now.

1. The death of smart as a status symbol

\"\"

For the entirety of the 20th and early 21st centuries, intelligence was the primary currency of status. We revered the doctor, the software engineer, and the lawyer because they possessed a cognitive monopoly on complex tasks.

But in 2035, displaying high cognitive intelligence is about as impressive as displaying the ability to lift a heavy rock was in the industrial age. When a digital agent can diagnose a rare disease or draft a complex merger agreement for $0.001 of energy, being smart is no longer a differentiator.

The contrarian reality: The new elite are not the knowledge owners. That term is now an oxymoron. The new elite are the “Empaths” and the “Makers”.

2. Provably human is the new organic

\"\"

In 2025, we worried about deepfakes. By 2030, the internet had become the dark forest, a place so flooded with AI-generated noise and infinite content that truth became the most expensive asset.

This triggered a massive cultural recoil. Just as we once paid a premium for organic food to avoid industrial chemicals, in 2035 we pay a premium for organic experiences to avoid synthetic intelligence.

The contrarian reality: We have seen the rise of analog-only zones.

3. The end of the job and the rise of the portfolio citizen

(To understand this part, I recommen d you read my previous article here)

\"\"

The old Marxist fear was that if capital (machines) replaced labor, the worker would starve. But as I wrote back a few weeks back, the system couldn’t survive zero demand.

We didn’t get the dystopian breadlines; we got the Universal Basic Equity (UBE).

The contrarian reality: In 2035, you don’t ask someone “What do you do?” You ask, “What is your stack?”

4. The rise of conspicuous inefficiency

\"\"

My previous hypothesis that value would flee to the physical interface because robots couldn’t handle the messiness of the real world was too optimistic. It turns out, once you give an AI a dexterous body, it solves the doing bottleneck just as fast as it solved the thinking one.

So, in 2035, why are there still human bartenders, human gardeners, and human butlers?

The contrarian reality: Human labor has transformed from a factor of production into a luxury good.

In the 20th century, buying a handmade item was often about higher quality. In 2035, a robot-made table is objectively superior: stronger joinery, perfect symmetry, zero waste. Therefore, hiring a human to build a table, or pour a drink, or drive a car is an act of conspicuous inefficiency.

5. From execution to curation

\"\"

Finally, the definition of creativity has flipped. In the past, having an idea was cheap, and executing it was expensive. Ideas are worthless, execution is everything, VCs used to say.

In 2035, execution is an API call. You can build an app, a movie, or a supply chain by whispering into your lapel.

The contrarian reality: We are living in the era of taste.

Conclusion: The great exhale

\"\"
Me trying to exhale in December 2025

The world of 2035 is not the cold, robotic dystopia we feared. In fact, by automating intelligence, the spreadsheets, the legal reviews, the debugging, we inadvertently forced ourselves to become more human. How is that for a positive spin?

We stopped trying to be machines. We let the machines be machines, and we remembered that our comparative advantage was never our ability to calculate. It was our ability to care, to touch, and to dream.

The robot took the job. But it gave us back our life.

OVNI Capital specializes in bridging the gap between European deep tech innovation and U.S. market leadership. With offices in San Francisco and Paris, we partner with visionary entrepreneurs to build global category leaders by bringing their breakthrough technologies to the U.S. from day one. Our investment strategy is rooted in systematic co-investments with leading U.S. and European funds and leveraging our extensive network of LPs in North America.

\"\"

At OVNI Capital, we don’t just fund companies; we empower them to transcend borders and redefine industries.

", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/@augustinsayer/2035-the-great-reversion-b505409b99f0", "href": "/insights/2035-the-great-reversion", "topic": "Venture" }, { "id": "85689d8ad733", "position": 50, "visible": true, "name": "Why AI may be Marx’s inadvertent vindication", "slug": "why-ai-may-be-marxs-inadvertent-vindication", "shortDescription": "This is an observation of how AI’s underlying economics may accidentally recreate the dynamics Marx theorized, and obviously not an…", "dateLabel": "12/2025", "dateIso": "2025-12-08T18:13:25", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/1*V9Uj5ox-snjr_OsSlpuCuA.jpeg", "bodyHtml": "

Why AI may be Marx’s inadvertent vindication

This is an observation of how AI’s underlying economics may accidentally recreate the dynamics Marx theorized, and obviously not an endorsement.

\"\"

History has a sense of irony that is borderline mathematical. For the last century and more, the greatest engine of wealth creation, western democratic capitalism, has defined itself in diametric opposition to the collective ownership of the means of production (socialism). We built the modern world, or most of it, on the sanctity of the individual actor, the entrepreneur who takes risks, and the laborer who sells his time for a wage. We believed, with different fervor among western countries, that the market was the ultimate judge for value and that competition was the only driver of progress. But as we stand closer to the AI Breakpoint (AGI?), where the marginal cost of intelligence collapses to zero, we must confront a terrifying yet exhilirating conclusion: Silicon Valley is unintentionally, constructing a system whose mechanics resemble what Marx described, though stripped of any ideological intent or collectivist design.

This is not a pivot of ideology; it is a pivot of mechanics. To understand why, we must look at the invisible loop that has powered the last century of growth. Capitalism has always relied on a symbiotic cycle: the worker produces the goods, earns a wage, and then becomes the consumer who buys the goods. Henry Ford understood this perfectly: he didn’t just build cars, he built a wage structure that allowed his employees to buy them.

But the AI technology stack will be severing this link. By decoupling economic growth from human labor, we are dismantling the mechanism that turns workers into customers.

Consider the math: If the marginal cost of intelligence trends toward zero, we unlock infinite supply (production). But if that same intelligence replaces the human earner, we create zero solvent demand (consumption). This is the paradox: Private ownership of the means of production becomes mathematically incoherent in a world where the means of consumption have been engineered out of existence.

The crisis of the consumer

Let’s start with the consumer. Karl Marx’s 19th-century critique of capitalism disastrously failed because he fundamentally underestimated the elasticity of human utility and the power of technological deflation. He believed capital would hoard wealth until the proletariat starved, which would in turn spark revolutions. He was wrong, and proven wrong endless times, because Schumpeter was incredibly right: technology made goods cheaper and created new, higher-paying jobs. The Fordist compromise held the world together for a century: we paid the workers enough to buy the cars they built. This virtuous cycle ensured that supply always met a solvent demand.

However, the AI Breakpoint shatters this compromise. When a car is built by a robot, designed by an LLM, and logistically managed by an autonomous agent, who is paid? If AI drives the marginal value of human labor to zero, it effectively demonetizes the consumer. In a world where software writes software and robots manage the physical interface, the supply side of the economy approaches infinite abundance (is energy production the next frontier?), while the demand side, human purchasing power derived from wages, approaches zero. This creates a capitalist’s dilemma. You cannot have a mass market without a mass labor force (yet). If venture capitalists fund the companies that replace 99% of the workforce, they are simultaneously deleting their future customers. We are building an engine of infinite supply for a world of zero demand.

The new means of production

\"\"

In the industrial age, the means of production were scarce, rivalrous physical assets: blast furnaces, railroads, and arable land. If one capitalist occupied the factory, another could not. In the AI age, the means of production are compute, energy, and data. My “theory” on the cognitive monopoly suggests that value is fleeing to proprietary data and massive compute clusters. This consolidation creates a power law on steroids, moving us toward a reality where a single entity could theoretically possess the intelligence to diagnose every disease, write every law, and design every building.

In this scenario, the distinction between private capital and state utility evaporates. If a single closed-loop system controls the zero-marginal-cost infrastructure of existence, allowing that entity to operate solely for profit is a systemic risk. It isn’t just unfair in a social justice sense, it is an economic dead end. If the cost of production is zero, the price must trend toward zero. But if the price is zero, there is no profit. If there is no profit, the capitalist incentive structure collapses. We are left with a system that looks less like a competitive market and more like a public utility. The logic of the technology demands that the dividend of this infinite intelligence be distributed, not because of charity, but because the economy requires liquidity to function. Do you see where I’m going with this?

The end of the labor theory of value

We are moving from a society based on the labor theory of value where things are worth the human effort required to produce them to an energy theory of value, where value is determined by the joules and compute required to manifest them. In this new world, work ceases to be the primary mechanism for distributing resources. This terrifies conservatives because it detaches survival from effort and it terrifies liberals because it detaches power from organized labor (if it hasn’t already). But technology does not care about our political convictions.

We are entering a paradox: As tools like ‘Lovable’ make software production free, the ability to build is being democratized, but the ability to profit is being centralized. To navigate this, we must first strip away the toxic/deadly branding of Communism. The soviet experiment didn’t fail because sharing is bad, it failed because it tried to centrally plan scarcity. No human bureaucracy could manage the complex pricing of bread and shoes across eleven time zones without creating poverty and death. It was a failure of data processing.

AI solves the data problem and ushers in abundance. But this abundance forces us to question the moral foundation of the billionaire. Historically, we allowed the capitalist to keep the surplus because they took on execution risk. They did the hard, messy work of organizing human labor and supply chains. But when the worker becomes a GPU cluster, that execution risk evaporates. The capitalist is no longer a general contractor organizing a workforce, they become a toll collector, simply charging rent on a digital brain that was trained on the collective knowledge of humanity.

Conclusion: From UBI to UBE?

The synthesis of these paragraphs is not a violent revolution, but a quiet and technical restructure. However, we are wondering how the future is going to look. Many in Silicon Valley are talking about Universal Basic Income (UBI) as the solution. But to me, UBI is a trap. It is a stipend for survival paid by the machine-owners to the masses to prevent them from revolting. It maintains some of the dynamics of the old world.

If we truly accept that the means of production (compute and data) are becoming the only source of value, then the answer is not a salary , it is equity. We should move toward Universal Basic Equity (UBE).

The trillion-parameter models that will drive this new economy were not created in a vacuum. They were trained on the collective output of human history: our books, our open-source code, our art, and our scientific papers. We, all humans alive (and future humans), are the Limited Partners (LPs) in this new asset class. Therefore, the dividend of the AI age cannot be a government handout funded by taxes. It must be a structural ownership stake in the compute infrastructure itself. Every citizen effectively holds a share in the National Compute Reserve. If the machine earns, we earn. If the economy grows, our dividend grows.

This is not a call for collectivism, but a recognition that advanced capitalism may need to evolve its ownership logic to remain viable. As I’ve said before, the efficient frontier is gone and the infinite frontier has arrived. The only question left is: Who is on the cap table?

OVNI Capital specializes in bridging the gap between European deep tech innovation and U.S. market leadership. With offices in San Francisco and Paris, we partner with visionary entrepreneurs to build global category leaders by bringing their breakthrough technologies to the U.S. from day one. Our investment strategy is rooted in systematic co-investments with leading U.S. and European funds and leveraging our extensive network of LPs in North America.

\"\"

At OVNI Capital, we don’t just fund companies; we empower them to transcend borders and redefine industries.

", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/entrepreneurial-resolutions/why-ai-may-be-marxs-inadvertent-vindication-85689d8ad733", "href": "/insights/why-ai-may-be-marxs-inadvertent-vindication", "topic": "AI" }, { "id": "b6a6dc02574b", "position": 60, "visible": true, "name": "Venture Capital in the age of zero-marginal labor", "slug": "venture-capital-in-the-age-of-zero-marginal-labor", "shortDescription": "This past week, I attended for the first time a Jeffersonian dinner (single-speaker format) focused on futuristic themes.", "dateLabel": "11/2025", "dateIso": "2025-11-24T16:54:08", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/1*B8vzj9FK0_5zo9r_Q-HhRA.jpeg", "bodyHtml": "
\"\"

This past week, I attended for the first time a Jeffersonian dinner (single-speaker format) focused on futuristic themes. It made for quite an exciting discussion. Towards the middle of the dinner, we had exchanges around the marginal value of human labor going to zero. Being the econ major that I am, it led me to thinking about Schumpeter’s law and its impending death.

Intro: The sunset of the twin laws

For the past 50 years, the global economy, and the venture capital asset class that fuels its innovation, has been guided by two invisible handrails. These handrails functioned as the “laws of motion” for modern growth.

The first was physical and technical: Moore’s Law. This was the observation that the density of transistors on a microchip would double every two years, guaranteeing that compute would become exponentially cheaper and more powerful. It was the supply-side engine of the digital age.

\"\"

The second was socioeconomic: Schumpeter’s Law. Derived from Joseph Schumpeter’s theory of “creative destruction,” this law offered a comforting economic equilibrium. It held that while technological innovation destroys old industries and jobs, it inevitably creates new, higher-value human tasks in their place. The blacksmith becomes the machinist; the elevator operator becomes the prompt engineer. This law ensured that the marginal value of human labor remained positive and generally increasing.

\"\"

Today, we stand at a precipice where both laws are decoupling. Moore’s Law is dying a death of physics, crashing against the hard limits of the atom and thermodynamics, turning scaling from a “free lunch” into a capital-intensive grind (hello Ncodin, hello Isentroniq!). But far more consequentially for the investor and humans around the universe, Schumpeter’s Law is dying a death of economics.

As Artificial Intelligence drives the marginal cost of cognitive labor toward zero, the historical guarantee that human labor will always find a higher-value refuge is expiring. We are entering an era where the marginal value of human labor, for the first time in history, is trending toward zero.

\"\"

I. The mechanism of the old world: Why Schumpeter worked

To understand why the investment landscape is shifting, we must understand why Schumpeter’s Law held true for so long. Why didn’t the steam engine or the spreadsheet make humans obsolete?

The answer lies in The Cognitive Monopoly.

For the entirety of human history, no matter how powerful our machines became, they lacked the ability to navigate unstructured environments, understand context, or make decisions based on incomplete data. A tractor could plow a field, but it couldn’t decide which field to plow, nor could it negotiate the price of the wheat. “Intelligence” and “Adaptability” were scarce resources that only the biological substrate of a human could provide.

Consequently, automation was always a complement to labor, not a substitute. When VCs funded SaaS companies in the 2010s, they were funding tools that made humans more efficient. Salesforce didn’t replace the salesperson; it gave them a dashboard to sell more. The “unit of production” remained the human, augmented by software. This created a virtuous cycle: Capital deepened, productivity rose, and wages followed (imperfectly), because the human operator was the indispensable pilot of the machine.

II. The breakpoint: the zero marginal cost of intelligence

The arrival of Generative AI and autonomous agents marks the end of the Cognitive Monopoly. We have decoupled “intelligence” from “consciousness.” We can now replicate the economic output of a knowledge worker without the biological overhead of a human.

This breaks the Schumpeterian cycle because AI is shifting from Labor-Augmenting to Labor-Replacing.

\"\"

Consider the unit economics of a standard Series B startup. The largest line item on the P&L is almost always “Salaries and Benefits” (often 60–80% of OpEx). This cost exists because human intelligence is expensive to rent. Humans need sleep, health insurance, psychological safety, and training.

However, Sam Altman’s “Moore’s Law for Everything” thesis suggests that the price of intelligence is falling to the price of energy. If an AI agent can write code, answer support tickets, or analyze legal contracts for $0.05/hour with 24/7 availability, the market-clearing price for human labor in those domains collapses.

Schumpeter’s Law fails here because there are no “safe harbors” left.

Where do we flee when the machine can think? There is no higher cognitive ground (unless leisure is?). When the marginal cost of intelligence is zero, the mechanism of “creative destruction” still destroys the old jobs, but it no longer requires humans to build the new ones. The “creation” part of the cycle is executed by the machines themselves.

\"\"

III. The crisis of moats: where does the value go?

If intelligence is a commodity, and any founder can spin up an AI engineering team and an AI sales team via an API, where is the competitive advantage?

Get Augustin Sayer’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

Schumpeter’s Law provided a natural moat: it was _hard_ to hire 500 great engineers. That friction protected incumbents. If that friction disappears, VCs must hunt for new forms of scarcity. The “Alpha” in venture could be moving to three specific areas:

1. Proprietary data and the feedback loop

When the model (the brain) is a commodity, the context (the memory) is the gold. Returns will flow to companies that sit on unique, non-public data streams.

2. The physical interface

As the digital world becomes frictionless (and therefore lower margin), value will flee to the physical world, which remains stubborn, scarce, and hard.

Our Operating Partner discussed this precise subject in one of our recent podcasts with (Red Glass Ventures):

3. Trust and brand

In a world of zero-marginal-cost content, the internet will be flooded with AI-generated noise. “Truth” becomes the most expensive asset.

\"\"

IV. Power law on steroids

Venture Capital has always been a power-law game, where one winner pays for ninety losers. The death of Schumpeter’s Law will exacerbate this to an extreme degree.

In the labor-constrained world, there was room for the “Number 2” or “Number 3” player. If Salesforce was too expensive, you could hire a cheaper team to build a CRM for SMBs. But if the cost of software production is near zero, the “Best” product costs the same to produce as the “Mediocre” product.

This leads to Hyper-Winner-Take-All markets. The best AI doctor, the best AI lawyer, and the best AI tutor can serve the entire global population instantly.

V. The idea age

\"\"

While the death of Schumpeter’s Law sounds ominous for the worker (it presents massive societal challenges regarding distribution) it is the most bullish signal in history for the creator (and could be for the investor).

We are transitioning from an era of Execution Risk to an era of Imagination Risk. For the last 50 years, VCs often had to pass on brilliant ideas because “the go-to-market is too hard,” “the margins are too low,” or “building the engineering team would take too long.”

As the marginal value of human labor drops to zero, the friction between “having an idea” and “making it real” evaporates.

Conclusion: the new allocation

Moore’s Law is dead because we ran out of atoms. Schumpeter’s Law is dead because we are running out of human utility. But from the ashes of these laws rises a new paradigm.

We are possibly entering a period of super-abundance. The constraints on economic growth are no longer the number of working-age adults or the speed of training them. The constraint is simply how much energy we can generate and how boldly we can dream.

For the Venture Capitalist, the mandate is clear: Stop looking for the next platform that optimizes human labor. Start looking for the engines that replace it. The job of the investor is no longer to fund the management of scarcity, but to fund the design of abundance. The returns will not just be financial; they will be civilizational. The efficient frontier is gone; the infinite frontier has arrived.

OVNI Capital specializes in bridging the gap between European deep tech innovation and U.S. market leadership. We partner with visionary entrepreneurs to build global category leaders by bringing their breakthrough technologies to the U.S. from day one. Our investment strategy is rooted in systematic co-investments with leading U.S. and European funds and leveraging our extensive network of LPs in North America.

\"\"

At OVNI Capital, we don’t just fund companies; we empower them to transcend borders and redefine industries.

", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/entrepreneurial-resolutions/venture-capital-in-the-age-of-zero-marginal-labor-b6a6dc02574b", "href": "/insights/venture-capital-in-the-age-of-zero-marginal-labor", "topic": "Venture" }, { "id": "68c43101ec9e6cdffa02f66f", "position": 70, "visible": true, "name": "See-Write-Do Systems", "slug": "see-write-do-systems", "shortDescription": "Cameras are about to stop just recording — and start deciding what work gets done.", "dateLabel": "09/2025", "dateIso": "2025-09-12T00:00:00", "thumbnail": "https://firebasestorage.googleapis.com/v0/b/ovni-website.firebasestorage.app/o/insights%2Fgenerated-posters%2Fsee-write-do-systems-11b992f7168f.jpg?alt=media", "bodyHtml": "

We have put cameras everywhere. They are in trucks, on factory floors, at construction sites, in hospitals, in stores. For the most part they just sit there recording. Nobody cares until something goes wrong.

But models can now take a live video feed and turn it into language:

The reactive paradigm is ending: The only reason we don’t have 1,000x more cameras today is that reviewing footage took forever and cost too much and that bottleneck is about to disappear. This leads to the proactive paradigm: every critical job, every inspection, every high-risk workplace is going to have a camera on it.

AI Cameras Aren’t New

\"\"
Figure 1. AI Cameras timeline

We’ve been trying to make cameras smart for a long time, at least 15y.

In the 2010s, most of the technology was rules-based. A camera could flag “motion detected” by comparing one frame of video to the next. Security teams quicky learned to ignore alerts as a change of light, a passing bird or a branch moving in the wind could set them off.

The next wave (mid 2010s) leaned on classic computer vision. Algorithms like Viola-Jones which had been good enough for quite some time for consumer face detection were brought into CCTV and access control. They could confirm that a human was in the frame of that a face matched a template but the output was limited. Bounding boxes and binary signals didn’t translate well to the complexity of a construction site/warehouse/busy street.

By the early 2020s, startups began to specialize. Some offered helmet detection on construction sites, others pitched traffic analystics for city governments or intrustion detection for factories. They used deep learning models like YOLO and Faster R-CNN, which represented a leap over rules-based methods. But the accuracy was still inconsistent, false positives were piling up and the tools rarely integrated with the systems companies already used to manage work. An alert was still just an alert with no real way to place it in a bigger picture.

What Changed in the Past Two Years

Two big shifts have changed the equation:

You can see the shift in the data: Submissions to CVPR, the world’s leading computer vision conference, have grown nearly 3x since 2019, topping 13,000 in 2025.

\"\"
Figure 2. CVPR Conference Submissions (2010–2025)

Those two shifts are the proactive paradigm: cameras can now process an information on the spot, decide if it matters and trigger action in real time.

The First Generation of “Actionable Vision” Companies

Once you accept that cameras can now process and act, you start spotting the pioneers everywhere. On highways, in warehouses, in hospitals, on job sites. This is what the first market map looks like (non exhaustive).

\"\"
Figure 3. Quick map of the ecosystem

We’ve spent fifteen years trying to make cameras smart but the real breakthrough wasn’t in the lens , it was in the stack around it. i) Multimodal models gave cameras a language, ii) Edge compute made real-time inference cheap and iii) Agent frameworks are (hopefully) about to plug the output into workflows.

The result is a new class of systems: See-Write-Do ==> Cameras that process, decide, and act.

The first generation of companies is already here, let’s see who will build the next ones and in which verticals/layers. :)

Happy to discuss with anyone working on this topic.

", "authorSlug": "thomas-renaudin-aoh3t", "mediumUrl": "", "href": "/insights/see-write-do-systems", "topic": "Computer Vision" }, { "id": "e7d1877dd518", "position": 80, "visible": true, "name": "Demystifying Startup Term Sheets for Founders", "slug": "demystifying-startup-term-sheets-for-founders", "shortDescription": "When I raised my first round of funding as an entrepreneur, I remember staring at a term sheet full of legal jargon and feeling completely lost.", "dateLabel": "03/2025", "dateIso": "2025-03-26T06:13:40", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/0*58PnQKdJHJtb0AYT.jpg", "bodyHtml": "

When I raised my first round of funding as an entrepreneur, I remember staring at a term sheet full of legal jargon and feeling completely lost. Terms like _liquidation preference_ and _anti-dilution_ were being thrown around, and I was Googling their meanings at midnight. If you’ve been there, you know the feeling. At our last OVNI offsite (by Annecy) and related to an ongoing deal where the founder was genuinely curious about many clauses and their implications, we realized there had to be a better way to understand and navigate these documents without needing a law degree.

\"\"

One fifth of our LPs are based around the Geneva area. So our last offsite was in Annecy with events organized in Chambéry, Genève and Lausanne.

Fast forward to today, we helped create BuildYourTermSheet.com, a tool that lets anyone generate a venture capital term sheet by simply toggling clauses on or off. This article is the backstory of why we built it, and a friendly guide to demystifying the most common term sheet clauses in plain English. My goal is to bring more transparency, education, and empowerment to startup founders and early-stage investors alike, especially those without a legal background.

\"\"

The Backstory: Frustrations with Fundraising Legalese

Every founder knows that raising money comes with a stack of paperwork. When I first encountered a term sheet, it felt like reading a foreign language. I had great lawyers, but I still felt uneasy that I didn’t fully grasp what I was agreeing to. Was I giving up too much control? What did all these preferred stock terms actually mean for me and my team down the line?

I wasn’t alone. Talking to other startup founders and even some new angel investors, I kept hearing the same pain points:

These conversations struck a chord. Venture deals are already stressful; not understanding the terms adds unnecessary anxiety. I wanted to change that. I envisioned a tool where you could play with the building blocks of a term sheet, see the effect of each clause, and learn as you go. No dense 50-page legal manuals, no assuming you’re an expert — just a simple, interactive way to build your own term sheet and actually understand it.

Transparency and Empowerment: The Goal

The mission behind BuildYourTermSheet is simple: make venture financing transparent and approachable. For too long, the knowledge of term sheet details has been locked behind law firm doors or buried in thick documents. That knowledge gap means founders often feel disempowered in negotiations, sometimes agreeing to terms they don’t fully comprehend.

By putting a term sheet builder in the hands of founders (and first-time investors), we’re aiming to level the playing field. Education is at the core of this project. Each clause in our tool comes with a plain-English explanation. Toggle a clause on, and you don’t just see legal text appear — you also get a short description of what it means and why it matters. Toggle it off, and you can see how the term sheet might look without that clause.

\"\"

The idea is that _transparency leads to better conversations_. When both founders and investors understand the terms clearly, negotiations become less about power imbalances and more about finding fair middle ground. We’re not lawyers and this tool isn’t a substitute for proper legal advice. But it empowers you with knowledge, so you can have informed discussions with your lawyers and investors rather than feeling in the dark.

\"\"

Building a Term Sheet by Toggling Clauses

So how does BuildYourTermSheet.com work? We designed the interface to be as straightforward as possible. Imagine all the common clauses of a venture capital term sheet laid out as a checklist. Next to each clause name, there’s a toggle switch (an on/off slider).

When you turn a clause “on,” that provision gets included in your draft term sheet. Turn it “off,” and it’s removed. As you toggle, the tool updates the term sheet document in real-time, reflecting your choices. This way, you can experiment: _What if I remove the anti-dilution clause? What if I add a second board seat for investors?_ You instantly see the differences in the document.

\"\"

In the screenshot above, you can see how intuitive the interface is. For example, to include a Pro-Rata Rights clause, you just flip the switch next to “Pro-Rata Rights” to on. A snippet of text explaining pro-rata rights will appear in the term sheet preview, and a tooltip might even give you a quick analogy about keeping your slice of the pie. If you’re unsure about a term, there’s an info icon you can click for a one-paragraph explanation in plain language.

\"\"

We built it this way to make the learning experience seamless. Rather than reading a long post about term sheets (ironic, I know, since you’re reading this one!), you learn by doing. You can literally construct a term sheet piece by piece, and every time you toggle something, you’re educating yourself about that clause.

Another benefit is collaboration. Founders can sit down with their team or mentors and play with the tool together. Early-stage investors can use it to draft a founder-friendly term sheet, perhaps turning off certain aggressive clauses to see what a “lighter” version looks like. It sparks conversation: _“Do we really need this clause? What’s standard?”_ and _“What happens if things go wrong — how does this clause protect us?”_ In other words, it’s not just a tool for creating a document, but for prompting meaningful discussions about the deal.

Get Augustin Sayer’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

In the end, just download your term sheet as a .docx, edit it, and share it.

\"\"

Common Term Sheet Clauses Explained

Let’s break down some of the most common venture capital term sheet clauses you’ll encounter. I’ll keep it simple and use analogies where possible, just like how our tool explains them. Think of this as Term Sheets 101:

Whew, that’s a lot of terminology, I know. The good news is that armed with a basic understanding of these concepts, you’ll feel a lot more confident when you see an actual term sheet. And remember, term sheets are usually just a starting point for negotiation — almost everything is up for discussion if something feels off. The tool we built is there to help you see what’s standard and what options you have, so you can enter those negotiations informed rather than intimidated.

Friendly Advice from a Fellow Founder (_albeit of a VC fund_)

Having been through this process (and countless hours tweaking this tool), I want to leave you with a few friendly tips:

  1. Don’t be afraid to ask questions. Whether you’re using a tool like BuildYourTermSheet or reviewing a term sheet from an investor, if something is unclear, ask for clarification. It doesn’t make you look inexperienced; it makes you look thorough and committed to understanding your business.
  2. Know your must-haves and deal-breakers. Every startup is different. Maybe maintaining control of the board is a must-have for you, or maybe you’re okay giving one board seat away but not two. Maybe you can live with a standard 1x liquidation preference (most can) but would push back on anything higher. Use your knowledge to decide what terms truly matter for your company, and which ones you have flexibility on.
  3. Use tools and templates to your advantage, but also get a good lawyer. Tools like BuildYourTermSheet are great for education and drafting, and standard templates (offered by accelerators or law firms) exist for a reason. But when it comes to signing a real deal, having an experienced startup lawyer review everything is invaluable. Think of the tool as a way to become an _informed_ client — you’ll save time (and legal fees) if you already understand the basics and only need to dive into the tricky parts with your lawyer.
  4. Keep the conversation open with your investors. In an early-stage deal, both sides ultimately want the company to succeed. If you approach the term sheet discussion with a collaborative mindset (and a shared understanding thanks to transparency), it sets a positive tone for the partnership. Investors actually appreciate when founders understand the terms; it makes the process smoother and faster for everyone.

Conclusion: Making Fundraising a Little Less Scary

Term sheets will probably always have a bit of intimidating legalese (they are legal documents, after all), but they don’t have to be a total mystery. With a little knowledge and the right tools, any founder or new investor can grasp the fundamentals. I helped build BuildYourTermSheet.com because I believe that understanding these documents is the first step toward negotiating fair deals and building trust on both sides of the table.

At the end of the day, a term sheet is not just a document — it’s the blueprint of a future partnership. When you know what each clause means, you’re not just protecting your interests; you’re also building credibility by showing you’ve done your homework. My hope is that with greater transparency and education, we’ll see fewer founders blindly signing on the dotted line and more founders confidently saying, “I understand what I’m signing, and I’m okay with it.”

Happy fundraising, and remember: knowledge is power, especially when it’s your company’s future on the table. Feel free to play around with the tool, share it with a fellow founder, and let me know if it helped you navigate your term sheet with a bit more confidence. After all, we’re all learning and building this startup rocket ship together — one term sheet at a time!

OVNI Capital specializes in bridging the gap between European deep tech innovation and U.S. market leadership. We partner with visionary entrepreneurs to build global category leaders by bringing their breakthrough technologies to the U.S. from day one. Our investment strategy is rooted in systematic co-investments with leading U.S. and European funds and leveraging our extensive network of LPs in North America.

\"\"

At OVNI Capital, we don’t just fund companies; we empower them to transcend borders and redefine industries.

", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/entrepreneurial-resolutions/demystifying-startup-term-sheets-for-founders-e7d1877dd518", "href": "/insights/demystifying-startup-term-sheets-for-founders", "topic": "Venture" }, { "id": "11a5ef219d69", "position": 90, "visible": true, "name": "From Zero to Deep", "slug": "from-zero-to-deep", "shortDescription": "The venture capital world is having its “iPhone moment.” Forget apps and SaaS platforms — the real money is now chasing fusion reactors…", "dateLabel": "01/2025", "dateIso": "2025-01-28T09:51:49", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/1*sXq20piXfXn9oFMCIcw5PQ.png", "bodyHtml": "

From Zero to Deep

\"\"

The venture capital world is having its “iPhone moment.” Forget apps and SaaS platforms — the real money is now chasing fusion reactors, quantum computers, and cancer-detecting blood tests. While deep tech pulled in $79 billion last year, the real story’s in the returns: these science-powered ventures are outperforming traditional tech funds by a staggering 70%. This isn’t just a new trend — it’s VC’s complete reinvention.

\"\"

Traditional tech markets have hit a technical and economic ceiling

Today’s conventional tech sector faces deeper issues than simple market overcrowding, revealing fundamental limits that are compelling venture capital to adapt. In the past, building a billion-dollar business by enhancing software was feasible. Now, those days are waning. As an example, enterprise software providers have seen their cost recovery timelines for new customers more than double — rising from roughly 11 months in 2019 to 24 months in 2023. Margins are also under strain, as core expenditures like cloud infrastructure have hit a plateau and aren’t likely to drop any further.

Artificial intelligence underscores the mounting constraints in traditional tech ventures, as seen with GPT-4’s staggering $100 million training cost, mostly for powering massive compute clusters. Yet OpenAI can only charge $20 monthly for ChatGPT because rivals like Deepseek rapidly introduce similar tools, forcing prices downward. This is no mere momentary hiccup but a fundamental issue: every doubling of model performance demands an exponential hike in compute resources, squeezing returns on purely software-based breakthroughs. While these AI advances remain impressive, they highlight the diminishing capacity of the standard tech model to sustain growth. In contrast, deep tech’s foundational scientific leaps create protected value, paving a more durable path for venture capital. It’s clearly the next frontier.

By contrast, deep tech ventures derive their value from fundamental scientific leaps that aren’t easily copied. Quantum computing offers a clear illustration: PsiQuantum is using photons instead of superconducting materials, allowing it to operate at “very cold” temperatures (-269°C) rather than the “nearly impossible” -273.135°C. Though that difference may seem minuscule on paper, it’s the difference between a practical technology and one that remains largely theoretical , much like moving from room-sized early computers to manageable desktop PCs. Crucially, these deeper breakthroughs create durable competitive advantages that months of coding can’t replicate.

Energy innovation is seeing a similarly transformative shift. Modern Electron has figured out how to convert heat directly into electricity with an impressive 45% efficiency, more than doubling previous benchmarks. Imagine reclaiming nearly half of the wasted heat in industrial processes and repurposing it for practical power. It’s akin to capturing half the water that typically evaporates in a reservoir. These sorts of advances, rooted in the laws of physics and often safeguarded by patents, exemplify the kind of profound innovation that pure software solutions can’t match.

Deep tech returns are fundamentally different due to technical moats

Deep tech’s success isn’t merely about boosting profits; it’s about creating breakthroughs so profound they’re exceptionally hard to copy, which is increasingly the only reliable path to venture-scale returns. From 2018 to 2023, every standout deep tech startup made at least three major scientific advances, each taking years and millions of dollars to achieve, naturally insulating them from rivals in a way traditional tech rarely can.

Take Upside Foods, a pioneer in cellular agriculture. They overcame one of lab-grown meat’s biggest hurdles: scaling animal cell production. Think of it like engineering a greenhouse that grows plants seven times faster while using 82% less water and fertilizer. Upside now produces meat for $3.30 per pound, compared to the $14.50 competitors charge — a leap from a high-end novelty to a viable alternative to conventional meat. More importantly, each breakthrough builds on the last, compounding the challenge for anyone trying to catch up. Venture capitalists call this a “deep moat”, a sustainable edge that only grows stronger over time.

Form Energy offers another striking example with their iron-air batteries. While most companies fine-tune lithium-ion tech, Form Energy took an entirely different route. By using iron, one of Earth’s most abundant materials, they reversibly rust and un-rust metal to store energy, slashing grid-scale costs to just $20 per kilowatt-hour. It’s like inventing a water tank 90% cheaper than standard models, capable of powering an entire city nonstop. Beyond the cost reduction, their competitive moat springs from fundamental chemistry, advanced manufacturing, and control over key raw materials.

Deep tech is targeting trillion-dollar technical limits

What sets deep tech apart and why it’s slated to dominate venture capital is its ability to tackle major constraints in colossal industries that software alone can’t solve. Semiconductors illustrate this perfectly. We’re nearing the physical limit of how small we can make chips using existing lithography. Emerging methods involve self-assembling molecules, imagine Lego pieces that arrange themselves into flawless patterns , potentially allowing features a mere nanometer in width (about five atoms across). Photonics is another avenue to solve this pain. This could unlock $500 billion in new chip production while achieving the kind of fundamental innovation that code can’t replicate.

Fusion energy, meanwhile, is becoming reality at Commonwealth Fusion Systems, which has engineered magnets strong enough to contain plasma at temperatures hotter than the Sun’s core. They’re leveraging new superconductors that operate at higher (though still cryogenic) temperatures. Although building their first fusion plant will cost $3.5 billion, it could yield virtually limitless clean energy at $45 per megawatt-hour, competitive with fossil fuels minus the emissions. This neatly captures why deep tech is the future of VC: solving the world’s toughest challenges calls for breakthroughs rooted in hard science, not incremental software tweaks.

Biotechnology underscores the same trend. Exact Sciences developed a $200 blood test that spots five forms of cancer with 94% accuracy. Rather than relying on multiple expensive screenings, a single blood draw can catch different cancers early, when treatment is most effective. The enduring advantage here goes beyond the technology itself: patents, regulatory hurdles, and clinical validation fortify its competitive moat.

The list of exciting examples goes on. Ginkgo Bioworks offers yet another solution: they cut the time to engineer new microorganisms from two years to four months. It’s basically programming biology like software, except these “programs” are living factories that can produce complex chemicals at a fraction of current costs. Targeting a $2 trillion market that traditional chemistry now serves, Ginkgo isn’t just optimizing processes, it’s rethinking entire industries from the ground up. That’s precisely why deep tech beckons to investors: transformative scientific leaps open doors that agile coding alone can’t.

My takeaway

The shift toward deep tech requires venture capital firms to fundamentally change how they operate, and this transformation is already underway. Technical due diligence, the process of verifying if a technology actually works can now costs up to $175,000 per investment and requires PhD-level experts. It’s like moving from reviewing a company’s financial statements to having to verify if their underlying scientific discoveries are real and valuable. But this increased complexity and cost is precisely why deep tech will dominate returns. It creates barriers to entry not just for startups, but for investors themselves, leading to better returns for those who can successfully make the transition.

The era of the generalist venture capitalist is ending not because specialized knowledge is trendy, but because it’s becoming the only reliable path to superior returns. As traditional tech advantages are increasingly competed away in months rather than years, the future belongs to investors who can identify and support the fundamental technical breakthroughs that create lasting value. This isn’t just a prediction — it’s the necessary outcome of how technology value creation is evolving.

OVNI Capital specializes in bridging the gap between European deep tech innovation and U.S. market leadership. We partner with visionary entrepreneurs to build global category leaders by bringing their breakthrough technologies to the U.S. from day one. Our investment strategy is rooted in systematic co-investments with leading U.S. and European funds and leveraging our extensive network of LPs in North America.

\"\"

At OVNI Capital, we don’t just fund companies; we empower them to transcend borders and redefine industries.

", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/@augustinsayer/from-zero-to-deep-11a5ef219d69", "href": "/insights/from-zero-to-deep", "topic": "Deeptech" }, { "id": "bd96b2f335af", "position": 100, "visible": true, "name": "Why startups should think of VCs as customers", "slug": "why-startups-should-think-of-vcs-as-customers", "shortDescription": "Startups today are operating under a dangerous illusion. While founders obsess over product-market fit, user feedback, and customer…", "dateLabel": "11/2024", "dateIso": "2024-11-26T08:28:54", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/0*MPlE5QOBAU8inHTq.png", "bodyHtml": "

Why startups should think of VCs as customers

\"\"

Startups today are operating under a dangerous illusion. While founders obsess over product-market fit, user feedback, and customer satisfaction, they’re often blind to an equally crucial truth: venture capitalists aren’t just capital providers — they’re also an odd type of customer whose needs must be understood and served — for the benefit of the startups’ continued growth and success. This isn’t a popular view in an ecosystem that worships at the altar of user-centricity, but it’s a reality that separates truly disruptive companies from those that merely show promise. Before you balk at this theory, let’s see why together.

The uncomfortable truth: VCs as customers, not capital

The traditional startup playbook suggests building purely for users, achieving product-market fit, and then approaching VCs when growth capital is needed. This sequential thinking, while intuitive, fundamentally misunderstands the nature of company building in today’s market. Venture capitalists aren’t passive capital providers waiting to fuel your growth , they can be sophisticated market participants whose needs, preferences, and success metrics could influence your company’s DNA from inception.

Understand how different this mindset is from the current founder orthodoxy. We celebrate stories of founders who ignore investor advice, who bootstrap until they’re undeniable, who build purely for users. While these tales are inspiring (Mailchimp, Basecamp, Atlassian, Zoho), they often mask a more complex reality: truly disruptive companies aren’t just built for users — they’re built for the sophisticated requirements of both users and venture capitalists simultaneously.

When Stripe decided to focus exclusively on developers, it wasn’t just a product decision. It was a deliberate choice that aligned with VC preferences for developer-led adoption in enterprise software. When Uber aggressively expanded into new markets despite massive losses, it wasn’t just about growth . It was about creating the kind of network effects and market dominance that VCs knew would lead to long-term value creation.

The capital imperative: why traditional bootstrap thinking falls short

This dual-client reality stems from a simple but often overlooked truth: the very nature of disruption requires speed and scale that only significant capital can provide. While bootstrapping can create profitable businesses, it rarely leads to the kind of enterprises that reshape industries (7% of unicorns are bootstrapped today) purely because internet has brought and intensified international competition to every single industry.

The capital imperative isn’t just about growth . It’s about the fundamental nature of disruption in modern markets. Whether it’s building payment infrastructure, revolutionizing transportation, or creating new computing platforms, these endeavors require massive capital not as an accelerant, but as a core ingredient. Without it, even the most brilliant innovations remain local successes rather than global transformations.

Similar to Uber mentioned earlier, take the example of Airbnb. The core concept of home-sharing could have been built as a bootstrapped business in a single city. However, the company’s true disruption came from its ability to create a global network effect, standardize the experience across markets, and build trust systems at scale. This required not just billions in capital, but building the company in a way that aligned with VC understanding of marketplace dynamics from the very beginning.

VCs as market oracles and strategic accelerants

Venture capitalists should bring more than money — they bring pattern recognition developed through seeing thousands of startups and their outcomes. Not all of them do but the ones you want supporting you must. When top VCs consistently emphasize certain metrics or strategies, it’s not arbitrary, it’s because they’ve seen countless companies succeed or fail based on these factors.

The best VCs function as market oracles, with their investment criteria serving as sophisticated predictive models for business success. When Sequoia Capital backed Apple in 1978, it wasn’t just providing capital — it was validating a new vision for personal computing. When Andreessen Horowitz heavily invested in crypto and web3 startups, it wasn’t just deploying funds, it was signaling a fundamental belief in the decentralization of the internet.

Beyond capital and vision, VCs serve as strategic accelerants through their network effects. When Benchmark invested in Uber, it wasn’t just the $13.5M Series A that mattered . It was their role in helping recruit key hires like CTO Thuan Pham and several board members, connections that proved crucial for international expansion. Similarly, when Sequoia led DoorDash’s $535M Series D, they didn’t just provide capital — they brought crucial insights from their investments in Asian delivery giants like Meituan (which preceded Sequoia’s investment in DoorDash by four years). That’s what you should expect from VCs.

Don’t build for VCs but use them as feedback loop to understand where the market is going. 50% of a company’s sucess will always be time to market, so if you really don’t buy into this theory of VCs as customers, then you better be cash resilient and expect to weather some drought before the potential flood of success.

The final word

The future of company building lies in recognizing that VCs aren’t just capital providers . They’re strategic partners whose needs and insights could shape company strategy from day one. This isn’t about compromising your vision — it’s about building companies that are designed to maximize impact through the powerful combination of user value and venture scale. In an era where speed to market and network effects often determine winners, the ability to align both user needs and VC feedback isn’t just an advantage . It’s increasingly the price of admission for building truly transformative companies.

OVNI Capital specializes in bridging the gap between European deep tech innovation and U.S. market leadership. We partner with visionary entrepreneurs to build global category leaders by bringing their breakthrough technologies to the U.S. from day one. Our investment strategy is rooted in systematic co-investments with leading U.S. and European funds and leveraging our extensive network of LPs in North America.

\"\"

At OVNI Capital, we don’t just fund companies; we empower them to transcend borders and redefine industries.

", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/entrepreneurial-resolutions/why-startups-should-think-of-vcs-as-customers-bd96b2f335af", "href": "/insights/why-startups-should-think-of-vcs-as-customers", "topic": "Venture" }, { "id": "6ebeef45c776", "position": 110, "visible": true, "name": "Stop half-measuring: why day-one U.S. presence is your only shot at building a global winner", "slug": "stop-half-measuring-why-day-one-u-s-presence-is-your-only-shot-at-building-a-global-winner", "shortDescription": "Most advice about U.S. expansion focuses on timing: “When should you enter the U.S. market?” This is the wrong question. The data shows…", "dateLabel": "11/2024", "dateIso": "2024-11-18T12:59:18", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/1*wAGLxGEP0fhiKEdyZHqVDA.png", "bodyHtml": "

Stop half-measuring: why day-one U.S. presence is your only shot at building a global winner

\"\"

Most advice about U.S. expansion focuses on timing: “When should you enter the U.S. market?” This is the wrong question. The data shows that in today’s global technology market, you’re either born with a U.S. presence, or you’re fighting an uphill battle that becomes steeper with each passing month. The success stories are clear: companies that treat the U.S. as their home market from day one build category leaders. Those that delay build acquisition targets.

Market Velocity: Speed Kills Competition

The raw numbers tell a compelling story. The U.S. software market stands at $315.24 billion versus Europe’s $230 billion. But this surface-level comparison misses the crucial point. The real story is in the growth rates: for the next 4 years alone, U.S. enterprise software spending increased at 13.5% annually, while Europe should be lagging at 5.0%. This differential isn’t just a statistic — it’s a kingmaker. And that’s even before you incorporate the geographical/cultural complexities of expanding throughout Europe.

Consider what this growth gap means in practical terms. A SaaS company starting in the U.S. market enjoys access to 50% of global SaaS revenue, while their European counterparts compete for a very fragmented 26%. But the implications run even deeper than market size. U.S. enterprises make purchasing decisions in weeks rather than months. They sign larger initial contracts. They’re willing to bet on innovation rather than waiting for proof in smaller markets.

\"\"
Source: Omnius 2024 SaaS Report (available here)

The U.S. market’s superiority manifests in three critical ways:

First, in sales cycle speed. U.S. enterprises have standardized purchasing processes for technology. They understand SaaS. They have dedicated software procurement teams. A sale that takes six months in Europe often closes in six weeks in the U.S.

Second, in contract values. From previous experience in enterprise sales, I would say that average enterprise SaaS contract in the U.S. is at least 2x larger than its European equivalent (not perfect statistics). I wouldn’t say this is because U.S. companies are richer — it’s because they understand the value of moving fast and committing resources to digital transformation.

Third, in market feedback velocity. U.S. customers give direct, immediate feedback. Most business users will put in their credit card to try your product and cancel quickly (one or two months top) if they’re not satisfied. This implied feedback loop accelerates product-market fit in ways that polite European rejection emails never will.

Capital Advantage: The Multiplier Effect

The velocity advantage directly translates into capital superiority. In 2023, U.S. venture funding reached $170.6 billion compared to Europe’s $45 billion. But focusing on these totals misses the structural advantage U.S.-present companies enjoy. The average Series A in the U.S. reached $18M in 2023, while European Series A rounds averaged just $8.5M (Q4 2023). That’s also an amazing opportunities for investors of European seed startups that then raise in the US.

\"\"
The US can be a formidable capital accelerator

This capital gap creates a compounding advantage that few founders understand until they’re suffering from it. Remember my article on trading the stock market last week? Well according to Bessemer Cloud Index 2023, SaaS companies in the U.S. trade at 12x revenue multiples versus 6.8x for European peers. This isn’t market inefficiency — it’s rational pricing of growth potential and market access.

Monday.com’s journey illustrates this perfectly. By establishing their commercial headquarters in New York while keeping R&D in Tel Aviv, they didn’t just access U.S. customers — they accessed U.S. valuations. Their public offering valued the company at $7.5B despite their Israeli origins. Compare this to European companies that often struggle to break $1B valuations even with similar revenue.

The fundraising advantage manifests beyond just amounts and valuations. When Israeli tech companies raise funding, 71% include U.S. investors. These aren’t just any investors — they’re the ones who understand specific verticals deeply and can make valuable introductions to potential customers and acquirers.

The implications extend to exit opportunities. In the last five years, U.S.-based tech acquisitions have averaged purchase prices 3.2x higher than European acquisitions in similar verticals. This isn’t coincidence — it reflects the strategic value of having an established U.S. presence and customer base.

Talent & Organization: The Winning Formula

Here’s where market velocity and capital advantage come together to create organizational imperatives. Most European companies make a critical mistake: they try to run U.S. sales from Europe, or they wait too long to establish senior presence in the U.S. The Israeli ecosystem, with 94 unicorns from a country of just 9 million people proves there’s a better way.

\"\"

For us, the winning formula is clear: Keep R&D in Europe, where deep technical talent is abundant and efficient, but move commercial operations and key decision-makers to the U.S. immediately. The data supports this approach: 86% of Israeli software companies that reached >$100M valuation had U.S. operations within their first two years.

This split structure works because it leverages the best of both worlds. European engineering talent is exceptional and often more cost-effective. European R&D teams show higher retention rates and deeper technical expertise. But U.S.-based commercial teams understand the market innately. They have the networks. They speak the language of U.S. enterprises not just literally, but culturally.

This isn’t just about sales — it’s about building a U.S.-native go-to-market motion. When product decisions, customer development, and go-to-market strategy happen close to your largest market, velocity increases exponentially. You start building what U.S. customers will buy, not what European engineers think they should use.

The Path Forward: Time is Your Enemy

The market is moving too fast for half-measures. Every month spent “preparing” for U.S. expansion is a month your future competitors spend building their lead. Every quarter focused solely on European customers is a quarter you fall behind in the market that matters most.

The evidence is irrefutable: Market velocity creates capital advantages, which in turn enable the right organizational structure. The Israeli ecosystem didn’t just stumble upon success — they engineered it by understanding this chain reaction. During the time European founders spend debating when to enter the U.S. market, their future competitors are already building there. During the months spent “preparing” for U.S. expansion, American companies are closing deals, raising capital at higher valuations, and building insurmountable leads.

This isn’t about expansion — it’s about survival. Your choice today determines if you’re building the next category leader or its future acquisition target. The U.S. market moves too fast, the capital gap is too wide, and the talent advantage is too significant to overcome with a delayed entry strategy. The next Monday.com, Wiz, or SimilarWeb won’t start as a European company that eventually succeeds in America. They’ll be born as an American company that happens to have exceptional European DNA. The only question is: will that be you?

OVNI Capital specializes in bridging the gap between European deep tech innovation and U.S. market leadership. We partner with visionary entrepreneurs to build global category leaders by bringing their breakthrough technologies to the U.S. from day one. Our investment strategy is rooted in systematic co-investments with leading U.S. and European funds and leveraging our extensive network of LPs in North America.

\"\"

At OVNI Capital, we don’t just fund companies; we empower them to transcend borders and redefine industries.

", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/entrepreneurial-resolutions/stop-half-measuring-why-day-one-u-s-presence-is-your-only-shot-at-building-a-global-winner-6ebeef45c776", "href": "/insights/stop-half-measuring-why-day-one-u-s-presence-is-your-only-shot-at-building-a-global-winner", "topic": "Venture" }, { "id": "209bdf8de0ae", "position": 120, "visible": true, "name": "The psychology of missing out: Why your sold stocks always seem to soar", "slug": "the-psychology-of-missing-out-why-your-sold-stocks-always-seem-to-soar", "shortDescription": "To better understand early-stage investments, I have consistently invested in the stock market (90% of which in the US). My experience working at a hedge fund before eventually founding OVNI Capital has reinforced the value of this approach. It provides us with critical insights ", "dateLabel": "11/2024", "dateIso": "2024-11-15T08:47:29", "thumbnail": "https://miro.medium.com/v2/resize:fit:1600/1*4GAD7as5s2FZlm1FmbkkIg.png", "bodyHtml": "

The psychology of missing out: Why your sold stocks always seem to soar

To better understand early-stage investments, I have consistently invested in the stock market (90% of which in the US). My experience working at a hedge fund before eventually founding OVNI Capital has reinforced the value of this approach. It provides us with critical insights into valuations (potential exits), market trends, and M&A opportunities, all of which inform our investment strategy in real-time.

\"\"

It’s a phenomenon so common it might as well be Newton’s 4th Law of Motion: the moment you sell a stock, it starts climbing like it’s trying to reach the sky. Meanwhile, that promising company you’ve been eyeing for months takes off without you, and your existing portfolio seems permanently stuck in the red. This is not just your imagination — it’s a fascinating intersection of behavioral psychology, cognitive biases, and the peculiar way our brains process financial decisions.

The science behind the sensation

A few years ago, Dr. Terrance Odean, a professor of finance at the University of California, Berkeley, conducted a groundbreaking study analyzing 10,000 brokerage accounts over a seven-year period. His findings were striking: the stocks investors sold generally outperformed the stocks they bought by approximately 3.8% annually. This phenomenon, which he termed the “disposition effect,” reveals our tendency to sell winners too early and hold onto losers too long.

Understanding loss aversion

At the heart of this psychological maze lies loss aversion, a concept introduced by Nobel laureates Daniel Kahneman and Amos Tversky (I strongly recommend reading the book Thinking Fast and Slow for more on the subject). Their research demonstrated that the pain of losing is psychologically about twice as powerful as the pleasure of gaining. This fundamental bias creates a perfect storm when it comes to investment decisions:

We rush to lock in gains (selling winners too early),

We refuse to accept losses (holding losers too long), and

We obsess over missed opportunities (amplifying the perception of “what could have been”)

\"\"

The selective memory effect

Research published in the Journal of Behavioral Finance found that investors tend to remember their missed opportunities and losing investments more vividly than their successes (hence the meme at the beginning of the article).

Dr. Sarah Williams, a behavioral economist at Stanford University, explains: “Our brains are wired to give more weight to negative experiences as a survival mechanism. In the context of investing, this means we’re more likely to remember and ruminate over our losses and missed opportunities than our gains.”

We humans are excellent at finding patterns that confirm our existing beliefs while ignoring evidence that contradicts them. This confirmation bias creates a powerful feedback loop:

You sell a stock,

You continue to track its performance,

If it goes up, this confirms your belief that “stocks always go up after I sell”,

If it goes down, you’re less likely to notice or remember it, and

The cycle reinforces itself.

The numbers tell a different story

A bit different analysis than the one mentioned in the first paragraph, when researchers at Vanguard analyzed investor behavior over a 30-year period, they found something surprising: investors’ perception of their “missed opportunities” was significantly skewed. While study participants reported feeling like they “always” missed out on big gains after selling, the data showed that:

Breaking the psychological cycle

Understanding these psychological patterns is the first step toward better investment decisions. Here are evidence-based strategies for managing these biases.

1. Implement systematic decision making

Investors who follow systematic trading rules perform better than those who rely on intuition:

2. Practice information hygiene

Investors who limited their portfolio monitoring to scheduled intervals experienced:

3. Keep a trading journal

Investors who maintain detailed records helps combat memory bias:

The feeling that sold stocks always go up is a textbook example of how our psychological biases can distort reality. By understanding these biases and implementing systematic approaches to combat them, we can make better investment decisions.

\"\"

The key is not to eliminate these feelings — that’s impossible. Instead, we need to acknowledge them while building systems that help us make decisions based on data rather than emotion.

Final practical steps for investors

To summarize and give some practical takeaways, here are my top 3:

Implement a waiting period: Research shows that mandatory “cooling-off” periods before making investment decisions can reduce emotional trading by up to 35%,

Use technology wisely: Modern portfolio tracking tools can help maintain objectivity by providing comprehensive performance data rather than relying on memory (hello SeekingAlpha + Jstock),

Develop a support system: Investors who discuss decisions with a trusted advisor or investment group (create a whatsapp group with friends if you haven’t already) make more rational choices.

Conclusion

The next time you feel that familiar pain of regret watching a sold stock soar, remember: you are experiencing a well-documented psychological phenomenon, not a personal curse. Your brain is playing tricks on you, highlighting the memories that confirm your beliefs while conveniently forgetting the many times your selling decisions were perfectly timed.

By understanding these psychological forces and implementing systematic approaches to combat them, we can work toward becoming more rational, successful investors. After all, the market does not have it out for you personally — it just feels that way because of how your brain processes information about gains, losses, and missed opportunities.

As Warren Buffett famously said, “The stock market is a device for transferring money from the impatient to the patient.” Perhaps it’s time we added: “and from the emotional to the systematic.”

Note: While this article draws from academic research and expert insights (thank you Perplexity for the sources), investors should always conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results.

", "authorSlug": "augustin-sayer", "mediumUrl": "https://medium.com/entrepreneurial-resolutions/the-psychology-of-missing-out-why-your-sold-stocks-always-seem-to-soar-209bdf8de0ae", "href": "/insights/the-psychology-of-missing-out-why-your-sold-stocks-always-seem-to-soar", "topic": "Markets" }, { "id": "67d990ba54a06d958945fb90", "position": 130, "visible": true, "name": "Sell the work/service/product", "slug": "sell-the-work-service-product", "shortDescription": "Vertical integrators don’t sell tools to incumbents — they sell the final product and take the market.", "dateLabel": "11/2024", "dateIso": "2024-11-14T00:00:00", "thumbnail": "https://firebasestorage.googleapis.com/v0/b/ovni-website.firebasestorage.app/o/insights%2Fgenerated-posters%2Fsell-the-work-service-product-79f840cb67ed.jpg?alt=media", "bodyHtml": "

We are more and more interested in vertical integrators — companies that leverage a tech-first, vertically integrated model to transform traditional physical industries.

→ Instead of simply selling tech solutions to established players, these companies build and deploy their own technology, allowing them to compete directly with them and tap into much larger markets.
→ Basically, you don’t sell hardware/software anymore, but you sell the final product/service. While some include only actors who use proven technologies to deliver the product, I’ll also include those who have developed new technologies and built physical factories around them.

Much has been written on this trend (see here and here), so I’ll only highlight some promising ideas/companies making strides in this space. The goal is to help people ideating on this (whether by copying or spotting unmet opportunities). If you’re excited about these opportunities and the tech risks they bring, feel free to reach out at thomas@ovni.capital.

Across these examples, the key technologies tend to focus on:

Here are some ideas/examples:

1. Advanced Cement & Concrete Production

2. Smart Home Energy Management

3. Precision Aquaculture

4. Advanced Manufacturing

5. Smart Mining/exploration & Refining

6. Advanced Nuclear Reactor Design

7. Intelligent Building Materials / Construction companies

8. Intelligent Infrastructure Inspection

9. Smart Lumber Manufacturing

There are unlimited opportunities, from new component manufacturing to advanced defense systems, blending hardware and software to deliver better products faster.

I’d love to hear any additional ideas you might have. Again, ping me at thomas@ovni.capital

", "authorSlug": "thomas-renaudin-aoh3t", "mediumUrl": "", "href": "/insights/sell-the-work-service-product", "topic": "Deeptech" }, { "id": "67dae206d04d2e2333444eb0", "position": 140, "visible": true, "name": "Software-Defined Everything", "slug": "software-defined-everything", "shortDescription": "Hard industries once dismissed as software-resistant are being rebuilt as software-defined systems.", "dateLabel": "08/2024", "dateIso": "2024-08-21T00:00:00", "thumbnail": "https://firebasestorage.googleapis.com/v0/b/ovni-website.firebasestorage.app/o/insights%2Fgenerated-posters%2Fsoftware-defined-everything-600e8b05ff56.jpg?alt=media", "bodyHtml": "

Software-Defined Everything.

Over a decade ago, a16z predicted that software would revolutionize every industry. While this has largely come true for many white-collar jobs across enterprise and consumer, crucial sectors (eg. manufacturing, construction, defense, healthcare…) have been slow to adapt, often dismissed by software investors as ‘software-resistant’ and ‘bad categories’. That’s changing fast. A new wave of entrepreneurs is proving that even these tough industries can go software-driven.

Welcome to the software-defined paradigm!

The Limits of Software-Only Solutions

Software has been the golden child of the tech world for over a decade — and for good reason. It’s scalable, offers high margins with low marginal costs, is flexible, provides recurring revenue, and avoids the headaches of managing physical inventory. But as we have pushed deeper into critical industries, we have bumped against many limitations of software-only plays:

  1. Integrating software with existing hardware/processes is a nightmare,
  2. User in traditional industries often prefer tangible interfaces (eg. a construction worker might be more comfortable with a physical control panel than a touchscreen app),
  3. When it comes to critical operations, there is an inherent trust in physical, tangible solutions. A factory manager might be more comfortable with a physical robot they can see and touch rather than a cloud-based AI system controlling their production line,
  4. Data collection issues (eg., in industries like manufacturing or agriculture, you need sensors and physical devices to gather real-world data, in real time). Software alone can’t bridge this gap between the digital and physical world.

Actually, I think I could write 100s of points on why it’s extremely difficult to build VC-backable companies targeting critical industries with a software-only value proposition — from the different setups needed for various factories, even within the same account; to operators who can’t afford to cut their factories’ continuous production lines, and don’t have time to set up complex software. So, let’s just move on and understand why software-defined products are gaining momentum. :)

Software-Defined Models, Definition+ Momentum

Definition

Lately, we’ve come across several entrepreneurs who are tackling age-old industrial problems by blending hardware with software and selling the hardware part. They’re using software — often powered by AI and data — to optimize traditional hardware, physical systems, and offline processes.

These models address several pain points that software-only solutions often miss, but three really stand out when we talk to clients and founders:

  1. They overcome interoperability issues by providing a standardized abstraction layer above diverse hardware components (facilitating integration between legacy and new systems), and
  2. They tackle the data collection issues by enhancing real-time data collection capabilities and providing a more flexible and standardized approach to integrating diverse sensors and data sources.
  3. Their customer base are more inclined to buy hardware wether than software, and it’s a true GTM hack.

We’ve noticed that integrating hardware speeds up the sales cycle (some industries are more likely to adopt new tech when it comes in a familiar, tangible form), and even helps create new categories within certain industries. The thing is, this isn’t exactly a new concept… so why is it gaining such momentum now?

Momentum

We have hit a sweet spot where technology, market demand and invesment potential all align.

Tech:

→ We are at a point where AI & Data Analytics have become incredibly powerful and accessible. We can now tackle complex, real-world problem.
→ The hardware part is not as intimidating as it used to be. Nowadays, hardware has become more of a commodity — just think of what you can do with a simple Raspberry Pi!

Market Demand:

→ People operating in the built world crave tech solutions, they know that many processes are still in need of optimization. Classical software just don’t allow them to do so. Software-defined is just another way to help these sectors adopt the tech, just as Vertically Integrated Companies do by integrating new technologies with more agility and beating incumbents.

Investors Appetite:

→ The real magic and the IP that VCs drool over lies in the software part. This mean that many investors an not scared of the hardware part which is commoditized and are more inclined to invest.

Concrete Examples

Several major tech companies have adopted this software-defined approach (i.e. Palantir, Tesla, Check Point…). Same for scale ups like Armis, Anduril, Flock Safety or even Hadrian that introduced the concept of Software-Defined Factories.

Many early-stage companies are exploring software-defined solutions, so I wanted to highlight 5 use cases that showcase its potential.

ReOrbit Space — Launching Software-Defined Satellites

\"\"

Traditional satellites transmit data in simple point-to-point paths, unlike the complex, interconnected networks on the ground. Software-defined satellites are revolutionizing this by enabling a more sophisticated, network-like data flow in space. ReOrbit Space is developing modular spacecraft with autonomous features and a software-defined design for greater flexibility (eg. a modular spacecraft could be easily reconfigured to switch between carrying communications equipment, scientific instruments, or Earth observation sensors, depending on mission needs.). In this case, software-defined design enable real-time adjustments to the satellite’s operations and data handling, improving mission efficiency and reducing operational costs.

Spore Biotechnologies — Pathogen detection in FMCG factories

\"\"

Spore.Bio is building a device that detects pathogens instantly on the factory floor. They use ML to assess bacterial concentrations in food, beverages, pharmaceuticals, and cosmetics, providing immediate alerts to quality managers. Previously, quality testing required sending samples to external labs, taking 5 to 20 days and incurring significant costs. Spore.Bio’s software-defined approach is the perfect example of data collection acceleration. They offer near-instant feedback, enhancing traceability, and enabling manufacturers to address contamination sources promptly. This not only cuts down on cross-contamination risks and expenses but also establishes a new standard in quality control.

Resolve Stroke — Software-Defined Ultrasounds

\"\"

Resolve Stroke is utilizing a software-defined approach to transform 15+ years of research into a product that maximizes the potential of ultrasound technology within commoditized devices. They package state-of-the-art ultrasounds and deep learning techniques into traditional hardware, and this represent the perfect example of IP laying in the software part and enabled by new hardware generation. This approach allows Resolve Stroke to pioneer a new category of data sets specifically tailored for the medical field. This woulnd’t be possible without the hardware part.

Trout Software — Hardware to build more secure facilities

\"\"

Trout develops a hardware solution called CyberSwitch that enables industrial sites to gain a comprehensive understanding of their networks, structure them efficiently, and implement advanced security measures through a user-friendly interface. Similar to Resolve, the core intellectual property is in the software, with the hardware being more of a commodity. This gives Trout a competitive edge in their go-to-market strategy. While industrial cybersecurity has traditionally been a luxury for large enterprises, the simplicity offered by CyberSwitch allows Trout to address the needs of smaller accounts that are often underserved by most OT players.

Deep Mine — Software-Defined Drones

\"\"

Deep Mine is changing detection through the use of sophisticated deep learning models within drones. Their approach employs Neural Networks to analyze Ground Penetrating Radar images, significantly improving detection accuracy. Their drone improves data collection, while the algorithm facilitates data augmentation and analysis. Again, perfect use case where the core IP resides in the software, but the client wants you to sell hardware.

The software-defined approach can deliver different types of advantages, from improved UX and GTM acceleration to unlocking new value propositions. I’m bullish on seeing other like-minded companies entering critical industries with this approach. If you are building with this thesis in mind, get in touch thomas@ovni.capital !

", "authorSlug": "thomas-renaudin-aoh3t", "mediumUrl": "", "href": "/insights/software-defined-everything", "topic": "Deeptech" }, { "id": "67db19f2e30c009ad68c476c", "position": 150, "visible": true, "name": "MEV Supply Chain", "slug": "mev-supply-chain-blockchain-infrastructure-providers-benefiting-from-the-emergence-of-real-world-assets", "shortDescription": "As real-world assets move on-chain, infrastructure providers sit at the center of a growing MEV market.", "dateLabel": "01/2024", "dateIso": "2024-01-01T00:00:00", "thumbnail": "https://firebasestorage.googleapis.com/v0/b/ovni-website.firebasestorage.app/o/insights%2Fgenerated-posters%2Fmev-supply-chain-blockchain-infrastructure-providers-benefiting-from-the-emergence-of-real-world-assets-96ad34297ba4.jpg?alt=media", "bodyHtml": "

Did you ever wonder how your favorite Decentralized Finance (DeFi) applications and on-chain protocols keep running and being accessible 24/7? It’s largely thanks to Infrastructure providers. These so-called nodes, and validator nodes, are respectively responsible for keeping a copy of public ledgers (e.g., blockchains) and validating every operation that is submitted (token exchanges, smart contract¹ interactions, etc.) to a blockchain. Given their critical positioning, infrastructure providers play a major role in the liveness and safety of blockchain Decentralized Applications (DApps)².

Originally discovered by an anonymous developer in 2014, it was only in 2019, that the Ethereum ecosystem started to witness ‘unusual trading activities’ — arbitrage bots (we’ll refer to them as searchers in this article), started to monitor pending operations⁴, from Ethereum mempool, in an attempt to exploit profitable opportunities created by them. Phil Daian was the first to coin this very opportunistic behavior as MEV, or Miner Extractable Value, in his research paper.

MEV historically refers to the profits validators (mostly miners at its inception) can generate from reordering, censoring, or front-running transactions on a blockchain. The MEV market has exploded over the last three years along with the growth of DeFi, generating over $1 billion in revenue for Ethereum miners in 2021 alone.

The momentum in this market is staggering. As more DApps, DEXs⁵, and protocols launch on different chains, the potential for MEV profits is growing. Besides this growth, the space has professionalized, to structure the MEV toward a structured Supply Chain, welcoming specialized actors, playing different roles in this emerging value chain:

On users’ hand, this might lead to slower, and trickier operation settlement and execution uncertainty — front/back-run transactions, censored operations, and lots of money being lost behind the scenes.

The risks are real. As the MEV market balloons, concerns over the centralization of blockbuilders, the emergence of private networks, and market manipulations are mounting. There’s a possibility that MEV could undermine the decentralized nature of blockchains if left unchecked. The future of this market remains uncertain, but one thing is clear — MEV is a big business opportunity.

Quick Definitions

MEV (Maximal Extractable Value): Previously known as Miner Extractable Value (when Ethereum were a Proof of Work blockchain). It is the maximum profit that can be made by including, excluding, changing the order of transactions in blocks, and arbitraging (front/back-run, sandwich attacks, etc.) operations in a mempool.
RWAs (Real World Assets): designates any assets, being physical, or digital, of which the value on a blockchain, stems from their existence outside — in the “real world”.
Blockchain Infrastructure Providers: any computing infrastructure that is responsible for playing a role in a blockchain (node, validation node, etc.)
Peer-to-peer networks: a group of computers, each of them being a node and sharing information, files within the group.

I. Real-World Assets’ state of the art

The tokenization of real-world assets (RWAs) has emerged as the next narrative of the bull market. From real estate and the digitization of financial assets to the rise of tokenized artworks, RWAs have experienced a significant growth. This growth is primarily facilitated by blockchain’s capability to enable the fractionalization of assets and provide trustless, secure registration, monitoring, and exchanges of tokenized asset fragments. Ultimately, it has the potential to make liquid assets considered as illiquid in traditional finance. Fundamentally, the tokenization of RWAs transforms the value and ownership of tangible assets, making their digital representations tradable and transferable.

The integration of on-chain RWAs addresses various existing challenges, including:

In the past two years, companies such as Ondo Finance, Mapple, or Centrifuge have started to offer tokenized RWAs to B2B clients leading to a surge in on-chain volumes. Examining these volumes, here are various application fields of RWAs:

\"\"
Market Mapping by Galaxy

The recent surge in institutional interest aligns with the shift from isolated enterprise blockchains to more interoperable tokenization trials on public networks and protocols. Major institutions such as Goldman Sachs or Franklin Templeton, along with consortiums of actors, are spearheading institutional use cases.

In terms of technology, as the cryptocurrency market matures, there’s a recognition that specific standards must be established to meet the requirements of RWAs and ensure compliance with existing regulatory frameworks. The recent introduction of novel token standards, such as ERC 3643, provides an openly accessible set of smart contracts crafted for creating, managing, and transferring tokens. This framework is well-suited for tokenized assets requiring advanced security controls.

II. The MEV rationale

In blockchain softwares, validators (or miners) serve as securing entities, ensuring the validation of on-chain operations and maintaining the network’s liveness by creating blocks.

It is worth mentioning that there is no theoretical guarantee that an initiated operation will be executed exactly as submitted by a user or any arbitrary initiator. This is because block producers (validators and block builders) select operations from the public mempool, arrange them, and include them according to their preferences. By default, block producers often order operations based on the highest operation fees to maximize their profits — operations require a minimum amount of fees, depending on their complexity, and their subsequent execution time. Consequently, block producers can exploit their advantageous positioning to unilaterally reorder operations, creating what is known as Maximal Extractable Value.

To illustrate such a phenomenon, consider an arbitrage opportunity arising on a decentralized exchange. Two possible outcomes may occur:

  1. Searchers notice the opportunity and adjust their operation fees to be the first to include their bundle of transactions⁶ in a block, thus profiting from the opportunity.
  2. Block producers replicate the searcher’s trade, censor its operation, and execute it themselves.

In cases where block producers choose not to exploit the opportunity, searchers may compete, creating a bidding market where they aggressively outbid each other to win the bid, eventually boosting block producers’ revenue.

Currently, MEV is mostly associated with searchers who significantly influence the order of operations in a block by submitting complex bundles of transactions and modifying the operation fees paid to block producers. Validators are the primary beneficiaries of MEV while still performing the same validation duties.

Common types of MEV include:

Optimally, MEV contributes to enhancing the efficiency of DeFi markets by creating financial incentives to address price inconsistencies or discrepancies. Still, MEV also poses risks to the network consensus, especially when unexpected attacks like Time-bandit occur.

III. The Growth of Real-World Assets on Blockchains Driving Transaction Volumes

a) Surging On-Chain Transaction Activity

The acceleration of real-world assets tokenization on public blockchains is driving an increase in on-chain transaction volumes. As more people buy, sell, and trade RWAs on-chain, the number of transactions will skyrocket.

Since their inception, digital asset tokenization hasn’t significantly taken off, the highest volumes being still dominated by stablecoins and NFTs. This is nonetheless fundamentally changing, as we witness an increase of interest, toward a momentum among financial institutions, retail users, as well as institutional investors.

\"\"

Different reasons may explain this recent surge in interest. Higher interest rates encourage tokenization use cases as a hedge against market volatility, certain classes of real-world assets suffering from limited liquidity and accessibility become openly tradable on-chain, and potential long-term cost reduction becomes an appealing factor for world governments and regulators. Hence, sovereign bonds, money market funds, and repurchase agreements are increasingly being offered, benefiting from enhanced liquidity, lower entry barriers and transferability restrictions, and assets’ divisibility, thanks to their tokenization.

b) Creation of New MEV Opportunities and their Mitigation

This influx of new on-chain activity presents an opportunity for arbitragers, market makers, and infrastructure providers (Block builders, Validators) to capitalize on inefficiencies, price discrepancies, and transaction volume. While searchers constantly monitor mempools and chains for arbitrage opportunities, looking to take advantage of users’ activity; their trading activity, when optimal, allows market makers to provide liquidity to the market, tightening spreads and stabilizing prices.

Infrastructure providers are positioned to be the biggest winners, as the more on-chain activity and volume increase, the more blocks tend to be profitable — infrastructure providers validating and including hundreds of complex operations in their blocks.

Incidently, some forms of arbitrage, such as front-running, or sandwich attacks are considered predatory. Arbitrageurs monitor pending transactions in the mempool for large orders and trade against initiating parties, harming liquidity providers and traders. Solutions like Flashbots, created a suite of tools to mitigate MEV. For instance, they implemented MEV-boost to manage private transactions until their delivery to selected validators, shielding them from front-runners. Other initiatives, such as MEV-Blocker tends to redistribute MEV to end users who create arbitrage opportunities for searchers.

Blockchain researchers are constantly seeking to implement strategies to curb malicious MEV and protect users while still allowing beneficial arbitrage. Some solutions include:

Validators and builders are incentivized to include the highest bidders, redistributing some profits to a broader range of actors. Solutions like this could make it simple for average users to participate in and benefit from MEV, rather than just the tech-savvy few.

Other options focus on limiting harmful MEV practices in the first place, such as front-running or censorship. For example, zero-knowledge proofs and other privacy-enhancing technologies (FHE) may make transactions opaque to MEV extractors, reducing their ability to manipulate order. Sequencing services can also help by obscuring the connection between transactions in a block, though MEV will likely find new vulnerabilities to exploit.

The growth of RWAs demonstrates the potential of public blockchains but also underlines the need to mitigate risks like predatory MEV. With the right solutions, blockchains can reduce these negative externalities, build trust in their networks, and achieve mainstream adoption. The future is bright for on-chain markets that get the balancing act right.

IV. The MEV Supply Chain Participants and Their Roles

When it comes to the MEV supply chain, there are several key participants involved and the roles they play. Let’s break down who’s who.

\"\"

a) Block Builders: Entities that run semi-proprietary algorithms (improved versions of public validator code), and compete in the market to create the most profitable blocks, on behalf of validators. They accept operations from MEV searchers, run their engine to craft the most profitable blocks, and send these blocks to relayers.

b) Relayers: Intermediaries between block builders and validators (aka proposers). They allow validators to offer their block space, given they are the ones selected by the consensus layer to produce new blocks. These entities have been controversial in the past, notably due to censorship concerns — Blocks submitted being deliberately ignored. We can distinguish two main categories of relays: censoring ones (that comply with legal constraints, such as blacklisted contracts and addresses) and non-censoring ones (that doesn’t operate any filtering).

c) Searchers: Also known as arbitrageurs, they write proprietary code and program bots to identify MEV opportunities, by carefully monitoring the public mempool as well as private pools of operations, where they compete to submit their bundles.

d) Peer-to-peer private networks: Decentralized connections among nodes, ensuring secure, direct communication and data sharing with a focus on transactions’ privacy. As of today, up to 15% of Ethereum blocks included transactions come from private networks.

e) Users: You and me and anyone else transacting on the blockchain. Unfortunately, as the MEV supply chain currently functions, regular users are often left vulnerable to front-running, censorship, and unfair transaction reordering by extractors looking to profit from MEV. User-focused mitigation strategies aim to give users more control and protection over their transactions.

f) Validators / Proposers: They are responsible for verifying and proposing new blocks that will be added to the blockchain. Validators can choose the most profitable block from multiple relays.

g) Indirect participants: Applications, protocols, and DApps that generate operations, influence operational fees, and validators’ execution rewards. They have the power to mitigate MEV for end-users.

V. The Risks of MEV Centralization Across Chains and Domains

The centralization of MEV across chains and domains poses serious risks that could undermine the resilience and censorship resistance of public blockchains. As the MEV market matures, certain participants are well-positioned to dominate the supply chain.

Large staking pools and validators have a built-in advantage for capturing MEV on proof-of-stake blockchains like Ethereum 2.0. They have more opportunities to propose blocks, giving them more chances to include high-value / high-rewards MEV transactions. This concentration of power threatens decentralization and could allow a few major players to exert control over transaction ordering and block production.

The emergence of specialized actors also introduces risks. Block Builders who develop advanced software and optimization algorithms to maximize MEV extraction may gain control over a large portion of the MEV market. If builders start to collaborate or merge, it could lead to a highly centralized block-creation process dominated by a few powerful entities. Such a scenario poses a serious threat to censorship-resistance.

Cross-domain MEV, where transactions on one blockchain can be front-run or manipulated by actors on another chain, represents an emerging threat, made possible by new efficient bridging solutions. As blockchain ecosystems become increasingly interconnected, MEV extracted from one domain could be used to attack another, creating a ripple effect across chains. This could undermine the security and resilience of the entire blockchain industry if left unaddressed.

Mitigating these risks and preserving the blockchain Ethos will require a combination of social consensus, cryptographic solutions, and protocol-level changes. Raising awareness about the threats of MEV centralization and rallying community support for solutions is a first step. Technological fixes like single-slot finality, verifiable delay functions, and zero-knowledge proofs offer promising ways to mitigate the MEV at a protocol level. Smoothing MEV redistributions across validators, staking pools and retail users could also help decentralize the supply chain and balance such phenomena.

Protecting retail users is critical to enabling mainstream adoption. By proactively addressing centralization risks and finding the right trade-offs between security, decentralization, and MEV opportunity, blockchains can build a more robust, resilient infrastructure for users and developers alike.

VI. Real-World Assets tokenization creating new MEV opportunities and Boosting Infrastructure Providers’ rewards

Bringing real-world assets onto blockchain infrastructures offers numerous advantages. By bridging Traditional Finance with Decentralized primitives, it enhances liquidity provision, streamlines exchanges’ settlement processes, and dismantles existing financial silos. This solution against opaque and inefficient financial systems, characterized by unfair transaction fees and time-consuming processes, represents a significant step towards transparency and efficiency.

The recent surge in tokenizing tangible real-world assets presents a lucrative opportunity for MEV infrastructure providers, particularly benefiting block builders and validators. The strategic move of directly incorporating tangible assets into blockchain infrastructures aligns with the broader objectives of digitizing and decentralizing ownership. This shift is expected to create larger liquidity pools, a diverse range of tokens backed by real-world assets offered directly to retail users, more advanced arbitrage opportunities, and blocks that tend to be more profitable than ever before.

Tokenizing real-world assets opens up opportunities to amplify on-chain operations, leading to increased fee generation, a surge in transactions, and consequently, augmented revenue for MEV supply chain participants — block builders, validators, and searchers. It’s worth noting that MEV is already a billion-dollar market, and the value capture potential for actors in the supply chain is poised to exceed $100 billion in the near future.

Looking forward, the potential arising from RWA tokenization emerges as a pivotal development for these actors. It not only significantly enhances liquidity, creating a more seamless environment for asset exchange and trade, but also contributes to the overall expansion of blockchain market volumes.

Conclusion

In the world of blockchain infrastructure technologies, where RWAs meet the nascent MEV Supply Chain market, we are at a crossroads of new emerging risks. The MEV market, an already billion-dollar industry, relies on the intricate dance between various participants: traders, infrastructure providers, and users. Yet, with this growth opportunity comes threats to decentralization and monopolistic positions.

Mitigating these risks demands a holistic approach blending technology, social consensus, and protocol-level changes. Democratizing MEV access and minimizing its negative externalities suggest collaborative solutions, aligning with a user-centric approach across blockchain infrastructures.

Looking ahead, the combination of MEV and RWA tokenization foretells a future where blockchain not only expands but highly specializes in complex software infrastructures, requiring professional expertise. The prospect of MEV participants reaping rewards in the era of RWA tokenization underscores the creation of new revenue opportunities.

Navigating the MEV and RWA landscape requires a delicate balance — embracing blockchain’s transformative potential while vigilantly guarding against centralization pitfalls. The future hinges on a collective commitment to a resilient, decentralized, and inclusive blockchain ecosystem.

¹ Smart contract: a piece of code that executes predefined functions on the blockchain.
² DApps or Decentralized applications: a set of smart contracts running to offer use cases and solutions.
³ Mempool: a storage area within the blockchain node, where transactions initiated on the blockchain can be stored for a short period of time, until their validation.
⁴ Pending operation: an operation from the mempool, that is being treated, but not yet included in a block.
⁵ DEX or Decentralized Exchange: a DApp representing a peer-to-peer marketplace where transactions occur directly between end-users. It fosters financial transactions that aren’t officiated by banks, brokers, or any other intermediary.
⁶ Bundle of transactions: transactions that are grouped together and executed in the order they are provided. Often used by MEV searchers, they may contain other users’ pending transactions from the public mempool.
⁷ L2 or Layer-2: technology built on top of a base blockchain (a Layer-1 network) that helps to extend the capabilities (mainly the scalability) of the underlying base layer network.

Working on MEV use cases? Feel free to ping us thomas@ovni.vc & daniel@venture-partner.xyz

", "authorSlug": "thomas-renaudin-aoh3t", "mediumUrl": "", "href": "/insights/mev-supply-chain-blockchain-infrastructure-providers-benefiting-from-the-emergence-of-real-world-assets", "topic": "Crypto" }, { "id": "67ecf93c484f2ab1509907a6", "position": 160, "visible": true, "name": "Revolutionizing Medical Imaging", "slug": "revolutionizing-medical-imaging-investing-in-resolve-strokes-ultrasound-breakthrough-2", "shortDescription": "Why we backed Resolve Stroke’s ultrasound breakthrough for point-of-care neuroimaging.", "dateLabel": "09/2023", "dateIso": "2023-09-18T00:00:00", "thumbnail": "https://firebasestorage.googleapis.com/v0/b/ovni-website.firebasestorage.app/o/insights%2Fgenerated-posters%2Frevolutionizing-medical-imaging-investing-in-resolve-strokes-ultrasound-breakthrough-2-466e57baa8e8.jpg?alt=media", "bodyHtml": "

As generalist investors, it is always challenging to differentiate between groundbreaking advancements and incremental innovations when it comes to Deep Tech.

Throughout the fundraising process, Aritz (Co-Founder & CEO) provided a compelling vision of the potential impact their solution could bring to the market through novel data sets. It was truly impressive to discover the talent within France’s ultrasound technology sector and the potential for innovation within this ecosystem.

The outcomes of a world-class ecosystem

Indeed, the renowned LIB (Laboratoire d’Imagerie Biomédicale), one of the world’s advanced laboratory when it comes to ultrasounds, have dedicated fifteen years to developing ultrafast sonography. By combining this technique with microbubbles and ultrasound probes, they found a way to visualize intricate details of the brain’s blood vessels and gather insights into blood flow dynamics.

No other non-invasive medical imaging technique has achieved this level of detail before. This new method is called Ultrasound Localization Microscopy.

Seizing this opportunity, Resolve Stroke emerged as a spin-off from the LIB, driven by a talented team comprising Aritz Zamacola (CEO), Vincent Hingot (CTO), and Olivier Couture (CSO and co-inventor of the ULM Tech). Their mission is to harness this technology and apply it effectively in the field of medicine.

Meet Resolve Stroke

Resolve Stroke has developed both a technology and a device to democratize ULM tech in a medical environment. It is called the 3D ULM Technology.

Technology
The company is developing a 3-level proprietary layer on top of the ULM technology:

1. 3D: Current ULM solutions are 2D. The company has focused on 3D renderings in order to reduce the level of expertise required to generate actionable data.
2. Human application: Historically, labs have trained the ULM technology with animals. The company is training the model on humans.
3. Rendering Speed: The entrepreneurs have focused on rendering results in 20 mins vs industry standard of 6 hours without losing quality.

\"\"

This 3D ULM technology is a type of ultrasound that provides detailed information about the blood vessels in the brain, such as blood flow velocity, vessel size, and potential abnormalities. It can also create maps of the cerebral vessels and perfusion indicators to assess tissue health. Resolve Stroke can use this for urgent and non-urgent cases, including acute mental change assessment, ICU follow-up, and stroke workflow. Finally, it also allows the startup to have a 100μm vascular resolution (vs industry standard of 1000μm).

Revolutionizing Point of Care Imaging in Neurocritical Care

A growing pain

Every year, 15 million people suffer from strokes, making it one of the leading cause of death and the first cause of disability worldwide.

In the United States, a stroke occurs every 40 seconds on average. It stands as one of the country’s leading cause of death, claiming over 140,000 lives each year. Moreover, it leaves a significant number of individuals with enduring disabilities, resulting in substantial healthcare expenses amounting to billions of dollars.

Timely diagnosis is crucial when it comes to strokes, as they are difficult to detect and can have devastating consequences. In stroke management, patients must undergo imaging to confirm and characterize the stroke before treatment with thrombolysis (drug) or thrombectomy (surgery) can begin.

Despite the existence of effective treatments, patients are too often deprived of them due to lack of time. Neither MRI technology nor CT scanners are capable of addressing this pain:

1. MRI technology: due to the complexity of the process, patients often experience longer wait times, with an average backlog of 4–6 hours.

2. CT Scanners: While they can provide more quick and detailed images, the toxicity of x-ray exposure and iodine-based contrast agents imposes similar limitations to MRI technology.

Ultrasound technology has become a popular option for point-of-care imaging due to its safety, compactness, and affordability. It is capable of capturing images of anatomy and blood flow in the largest blood vessels. However, the imaging quality of conventional ultrasound remains limited in neuroimaging due to poor image quality through the skull bone and has therefore made it impossible to treat strokes.

Market Dynamics

Out of every 100 patients, only three receive adequate treatment for strokes.

Proper diagnosis is lacking, and victims struggle to reach hospitals where they can receive the necessary care. This is the happening because of the difficulty in streamlining the process after a patient shows stroke symptoms

Still, the point of care imaging market for neurocritical care is evolving as existing devices (Start-ups & Incumbents) are educating the market on the importance of point of care imaging in ICUs and Mobile Stroke Units for prehospital stroke management.

1. Market Frustration: Existing devices are bulky, expensive, and have lower resolution compared to conventional devices, creating frustration in the market.
2. Increase in Demand: The prevalence of strokes is driving an increase in demand for point of care imaging solutions in ICUs and Mobile Stroke Units.
3. Network Densification: The need for a denser network of more compact and accessible devices with direct connectivity to expert stroke centers is increasing to streamline workflows and improve patient outcomes.

As the market continues to grow, there is an increasing demand for more accessible and higher resolution devices. The fundamental challenge lies in establishing connections between radiologists, emergency physicians, and primary care practitioners and the most suitable specialists for specific diseases. This connection is essential to enable effective care coordination, ensure optimized and consistent care pathways, and reduce variability in treatment.

Empowering Medical Innovation with New Data Paradigms

Resolve Stroke stands out from the crowd by focusing on pioneering a new category of data sets exclusively for the medical field. While other companies prioritize enhancing existing data sets, Resolve Stroke’s unique approach has the potential to disrupt the industry and attract new players who can leverage their specialized datasets for new use cases and drive a new wave of innovation.

Resolve Stroke’s Go-To-Market is on stroke management, but the impact of their technology extends far beyond this use case alone. By analyzing blood flow in the brain, Resolve Stroke’s technology can offer insights into an individual’s overall health and well-being.

In fact, there are numerous potential applications for their imaging technology. Imagine the possibility of assessing kidney health without the need for invasive biopsies or providing enhanced care before and after surgical procedures. Even diabetes assessment could be made more efficient and accurate.

By pioneering this new category of medical data sets and pushing the boundaries of what’s possible, Resolve Stroke is not only revolutionizing stroke management but also paving the way for advancements in various other areas of healthcare.

With its technology, Resolve Stroke is opening doors for new players in the industry who can leverage their solutions to excel in their respective domains.

\"\"

It’s an exciting time for Resolve Stroke and we couldn’t dream of a better team to tackle this massive opportunity!

Go Aritz, Vincent & Olivier!

", "authorSlug": "augustin-sayer", "mediumUrl": "", "href": "/insights/revolutionizing-medical-imaging-investing-in-resolve-strokes-ultrasound-breakthrough-2", "topic": "Health" }, { "id": "67ecfcc3c636180733e85bb9", "position": 170, "visible": true, "name": "Direct Air Capture", "slug": "direct-air-capture-where-weve-been-and-where-were-going", "shortDescription": "Carbon removal is moving from climate ambition to industrial engineering — and capital is following.", "dateLabel": "09/2023", "dateIso": "2023-09-05T00:00:00", "thumbnail": "https://firebasestorage.googleapis.com/v0/b/ovni-website.firebasestorage.app/o/insights%2Fgenerated-posters%2Fdirect-air-capture-where-weve-been-and-where-were-going-31e8b25b3499.jpg?alt=media", "bodyHtml": "

Over the past century and more, our primary method of obtaining energy has been heavily reliant on carbon-based resources such as oil, coal, and natural gas. We extract these fuels from underground, burn them to generate energy, and unfortunately, release carbon into the atmosphere in the process. For over 100 years, we have been stuck in this cycle of taking carbon from the ground and depositing it into the atmosphere.

When we take into account both CO2 and CO2 equivalents, the average annual emissions of these greenhouse gases range from 40 to 50 gigatons. Now, looking ahead to 2050, our goal is to reduce these emissions. But not only that, we need to remove a staggering 10 Gt/year.

Let’s put this challenge into perspective: we’re essentially tasked with building an entirely new industry from the ground up within just 20 years. And this new industry needs to be at least one-quarter as developed as what we’ve achieved over the past century. That is a big challenge, and this is why people are excited about making quick progress.

Enter Direct Air Capture or DAC — an approach that has been qualified as necessary by the IPCC 2021.

In this article, we will explore the history of DAC, how the different technologies work, the current state of the industry, and why DAC might be necessary in the fight against climate change. The road ahead for DAC isn’t without challenges, but the potential benefits to our planet are huge.

Join us to learn all about the past, present and future of pulling carbon out of thin air!

What Is Direct Air Capture?

Just a quick reminder: while we will explore DAC in this article, there are many ways to remove carbon from the atmosphere. Broadly, solutions fall into three categories:

  1. Biological Approach: Enhance natural removal pathways.
  2. Engineered Approach: Use technology to extract CO2 from the air.
  3. Hybrid Approach: Combine elements of both engineered and biological methods.

DACs fall into the Engineered approach category. They provide a way to remove CO2 from the atmosphere to address climate change.

\"\"
Figure 1. Explanation by Sylvera

Four crucial aspects define the success of DACs:

  1. CO2 traps: Implementing chemicals that act as effective traps, binding to CO2 molecules (often liquid solvents & solid sorbent but we will delve into this later).
  2. Air Circulation: Finding ways to move and expose the air to these traps, using minimal energy. While wind power may be suitable for certain locations, others may require using fans to engineer air movement through the sorbent or solvent material.
  3. Extracting: Applying minimal energy to collect the trapped CO2 in purified form (most energy intensive step in DAC operation). After engineering the traps, we can extract CO2 from the dilute mixture, breaking the bonds between CO2 and the trap to obtain purified CO2.
  4. Storing: Storing the captured CO2 is essential. To have a positive impact on the climate, CO2 must be stored geologically for extended periods, lasting decades to thousands of years. This ensures its complete removal from the global carbon cycle. That’s why DACS (Direct Air Capture and Storage) are truly the negative emission technology, not just DAC. Storage is key.

All this need to be achieved while being economically (low costs) and technically (high volume of carbon capture) viable, without consuming to much energy in the process.

Two approaches stand out: liquid solvent and solid sorbents

Both methods revolve around engineering materials or molecules that act as traps to capture CO2 from the air. These solvent/sorbents are designed to selectively adhere to CO2 while avoiding the binding of nitrogen, oxygen, or water, which are abundant compounds in the air:

\"\"
Figure 2. Carbon Credits
\"\"
Figure 3. Carbon Credits

Currently, the three biggest actors use one of these two approaches (Climeworks, Carbon Engineering and Global Thermostat).

A lot of companies use amines because they are extensively studied for acid/base reactions and already employed in CO2 capture in gas purification plants. These are well-understood, commercially available (although it is debatable), relatively low-cost, and highly selective for CO2. This selectivity is essential given the need to distinguish CO2 (0.04% of the atmosphere) from nitrogen, oxygen, and water, which make up over 99% of the atmosphere.

At the end of the day, it’s all about finding a way to decrease costs and increase capture amount with clean energy. We will talk about this later, but there is room for plenty innovation depending on the location of the DAC.

Milestones in the Development of DAC Technology

The history of DAC spans over two decades, with key milestones shaping its development (see below).

\"\"
Figure 4. DACs timeline

So first generation of DACs came in 2009 and demonstrated the feasibility of using machines to extract amounts of CO2 from the air, marking a huge technical milestone. But they faced challenges due to their energy-intensive desorption step, which resulted in high energy consumption per tonne of captured CO2. Also, many experts in gas separation believed that DAC would be too costly, stating that it could not cost less than $1000/tons.

Considering this, there has been an upswing in the advancement of second-generation DAC processes and companies from 2012 to 2020. Their main objective was to minimize the cost of DAC technology. But while these R&D efforts enabled to bring down costs (on the order of $700/tons right now), we are still far away from what is needed for the technology to have a real impact on climate change (both due to economic, technologic and energy issues). The real challenge now begins: DAC scalability.

DACs: Scalability issues

\"\"

As discussed, DACs face three primary challenges that hinder their scalability: I) cost-related issues, II) the need to capture carbon at a large scale, and III) energy consumption issues to make it work.

Cost Issues…

Out of all carbon capture applications, capturing CO2 from the air is the most costly. This is because the concentration of CO2 in the atmosphere is much lower compared to sources like flue gas from power stations or cement plants. So DAC requires more energy and incurs higher costs compared to these other applications.

Currently, DACs applications come with a price tag ranging from $600 to $1000 per tonne of CO2 captured from the atmosphere. When we take into account that permits on the European Union’s carbon market reached $106.57 per tonne recently (early 2023), the cost of capturing CO2 is six times higher than the selling price at its peak. This makes it challenging to establish a viable business model for DACs at the moment. According to BCG, costs must decrease to below $200 per metric ton, ideally closer to $100 per metric ton, by midcentury to have a significant impact on global climate goals (don’t really understand how this would be viable though, as it would only match the carbon price at its peak. If you have a POV on this, please reach out).

…Carbon capture at large scale issues..

As of now, DAC-based technologies remove less than 0.01 million tonnes of CO2 / year. Still, to align with the NZE Scenario, a scale-up is required targeting approximately 70 MtCO2/year in 2030 and around 10GtCO2/year in 2050, equivalent to Indonesia’s total energy-related CO2 emissions in 2021.

… and Energy consumption issues…

Current DAC installations have a very large energy footprint. A significant amount of energy (and consequently, CO2 emissions) is needed to potentially capture a small amount of CO2. For critics of DAC, this is the final straw.

Yup, much like numerous other industrial procedures, DAC demands a substantial energy input. The most advanced methods available today would necessitate over 100% of the globe’s renewable energy capacity to eliminate the maximum annual CO2 tonnage required (roughly 40GtCO2/y).

Jean- Marc Jancovici stated that “If we wanted to capture, using this kind of DAC device, the entirety of our annual emissions, we would need to dedicate all of the yearly electricity and all of the oil consumed in the world each year to it. Therefore, the energy would only be used to recover the CO2 emitted into the air due to energy consumption.” Quite an ironic situation.

At gigatonne scale, DAC will also faces limits beyond energy and cost — water, land, materials, and supply chains. Neglecting these could worsen environmental problems.

… But the game is not over yet

\"\"
Figure 5. Life cycle GHG emissions in kg CO2-eq. per ton of gross CO2 removal with the DAC plant as well as carbon removal efficiencies [%] for different system layouts in selected countries (Source: https://pubs.acs.org/doi/10.1021/acs.est.1c03263#).

In some cases, we can remove more CO2 than we create through DACs. The carbon removal efficiency can vary a lot (widely from 9% to 97%) depending on the geography, energy type, technology used, etc…

Despite all the issues, brilliant minds worldwide are actively trying to tackle these challenges. Their motivations aren’t necessarily technolutionist, as some claim. Many are simply guided by the IPCC’s assertion: carbon removal (at scale!) is essential. Starting the cutbacks a few years back could have rendered these technologies redundant, but now it is too late. For them, with the urgency of having surpassed the point of no return, any argument against its feasibility should trigger extensive research akin to the Human Genome Project.

Next steps for the DAC industry

The driving force behind this deep dive is the recent collaboration of leading teams aiming to forge the third generation of scalable DACs.

More and more brillant teams are entering the space

As always, illustrating logos on a map is a good method for comprehending an environment so let’s do it (see below). Not comprehensive — if you know of other teams taking the plunge, give us a shout!

\"\"
Figure 6. DACs Mapping

As of now, the International Energy Agency reports 130 DAC plants in global development, comprising 27 commissioned and 18 completed. These are all small-scale facilities, collectively capable of removing around 11,000 tons of CO2 annually.

\"\"
Figure 7. Direct air capture expansion projects of selected companies (Capacity in kt CO2/year)

Initial DAC tech demonstrated machine-based CO2 removal, but heat-driven desorption inefficiencies led to high energy use per ton of CO2. After several lab experiment, the 3rd generation of DAC aim to solve this by employing different strategies:

A winner-takes-all market ?

With all these new start-ups popping, it is easy to wonder: “Which tech will really win the big prize? How do I know I’m working on the next big thing?”

Well, it seems DAC isn’t a winner-takes-all tech field. It’s not about one big solution. We’re talking about 3, 6, 8, 10, even 20 different innovations that matter. Why? Because what works best varies by location.

\"\"
Figure 8. DAC and CCS plants around the globe

Look at it from a chemical engineering angle: DAC pulls in air wherever it is set up. Up north near the Arctic? You are dealing with cold. Down in the Middle East? Heat is the issue. So, the smartest, most cost-effective DAC designs differ a lot. Forget about that one-size-fits-all magic answer.

That being said, one approach seems to be more & more adopted. A lot of teams are trying to scale DAC thanks to process optimization. They create machines that use established and familiar technology, and the waste heat of existing infrastructures like industrial plants or buildings to lower energy consumption during desorption. So it really seems to be a matter of process integration now. This will likely turn into a supply chain issue very soon.

The next challenge: scaling Supply Chains

One big hurdle in pushing forward the progress of DAC is setting up efficient supply chains. Many startups use existing market components, gradually improving cost-efficiency.

As we get better at this and try to make things cheaper (going from $500 to $300 to $100 per ton), we’ll start dealing with more advanced technologies. This means more special gadgets, custom gas-solid contactors, and solutions that will need completely new supply chains.

Right now, we’re not making these custom parts, but soon, we might have to make millions of them every year. The challenge might be getting these supply chains up and running and teaming up with bigger companies that have the money to invest in making these parts / the infrastructures for creating these elements.

Conclusion

In the world of DAC, we are walking a tightrope between speed and patience, as

It is a balancing act. We can’t get frustrated if change is not immediate, but we can’t afford to slow down either. Policymakers need to push for speed, and investors need to understand that this is a long-haul commitment, not a quick fix.

So what’s next?

  1. Incentives are crucial, from both private and government sources,
  2. Welcoming new minds, especially young talent in fields like chemistry and engineering, will keep innovation alive,
  3. Everyone’s involvement counts, whether it’s tech experts or community members in the areas where DAC projects are happening (yes, local communities might not always be keen on having these big installations nearby).

I hope you liked reading this article! Your thoughts are really important to us. If you know of any teams working in this area, feel free to reach out.

‍

At OVNI, we invest in pre-seed/seed stages and partner with founders who have global ambitions from day one. If you are a founder in this space or know someone who is, feel free to contact me at thomas@ovni.vc.

", "authorSlug": "thomas-renaudin-aoh3t", "mediumUrl": "", "href": "/insights/direct-air-capture-where-weve-been-and-where-were-going", "topic": "Climate" }, { "id": "67ecfd6f17e207672077de62", "position": 180, "visible": true, "name": "The Rising Quantum Computing Industry", "slug": "the-rising-quantum-computing-industry-current-state-and-future-prospects", "shortDescription": "Beyond the generative AI wave, quantum computing is assembling a real industrial stack.", "dateLabel": "05/2023", "dateIso": "2023-05-12T00:00:00", "thumbnail": "https://firebasestorage.googleapis.com/v0/b/ovni-website.firebasestorage.app/o/insights%2Fgenerated-posters%2Fthe-rising-quantum-computing-industry-current-state-and-future-prospects-4c81ca63f15b.jpg?alt=media", "bodyHtml": "

As investors, we are always on the lookout for the next big thing in technology. And with the growing buzz around generative AI, it is natural to wonder what the future holds. But as we have been meeting with quantum computing start-ups lately, we have become increasingly intrigued by the potential of quantum technologies.

According to Gartner’s hype cycle, quantum technologies are currently in the innovation trigger stage. This means that while they are not quite ready for prime time yet, they are showing promise and attracting attention from investors, researchers, and innovators (especially in France).

\"\"
Hype cycle from Gartner

The aim of this post is to gain an understanding of the current state of the quantum computing field and the challenges that must be addressed to achieve maturity as an industry.

Part 1: Technologies driving the market and growth signals (from investments to partnerships)
Part 2: The software challenges to adopt the hardware
Part 3: The latest innovation made by Big Tech on the subject

Hope you will enjoy the reading!

What Is Quantum Computing? A Primer on How Quantum Computers Work

Quantum computing has been in existence since the late 20th century, but it’s only in recent years that research and development have picked up steam. Unlike traditional computers that operate on binary bits of either 1 or 0, quantum computing utilizes qubits that can represent both 1 and 0 simultaneously — a phenomenon called superposition.

\"\"
Schema created by “Le Lab Quantique” which we found really relevant to better understand the feature

This unique feature of quantum computing allows for multiple calculations and processes to be performed at once, making it ideal for handling complex simulations and calculations. For instance, quantum computing can accelerate drug discovery processes by analyzing numerous compounds at once to identify those with the highest efficacy. This could impact a huge number of sectors and unlock new use case for several industries.

The Growth of the Quantum Computing Industry

Quantum computing is a technology that has the potential to revolutionize many fields, but it requires technological breakthroughs in both hardware and software development. In the quantum computing field, the crucial technical challenge is the development of stable and reliable qubits, which are the fundamental building blocks of quantum computers.

Understanding the key players and their position in the technology stack is a challenge in emerging markets. One classic solution is to display logos, such as the map provided by The Quantum Insider (credit to Alex Challans for the effort):

\"\"
Market maping of the Quantum Computing industry made by the Quantum Insider, May 2022

By connecting this map with the recent report by “Le Lab Quantique,” we can identify key technologies that are driving the industry forward:

Superconducting circuits are currently the most prominent technology used to perform quantum computing. From Google, IBM, Intel, Rigetti, Alice&Bob to D-Wave, it has been the privileged qubit technology that most of the actors have chosen to bet on. This choice is not arbitrary at all, and one of the intuitive reasons for which superconducting circuits make the ideal qubit candidate is that superconductivity is essentially a Macroscopic Quantum Phenomenon (arises from the collective behavior of many particles at once).

Other technologies such as cold atoms, photonics, silicon, and carbon nanotubes are also being developed to yield qubits. These technologies are being explored to improve the quality and scalability of qubits.

Despite slightly lower fundraising pace in 2022, with an estimated $2.2 billion invested in quantum(see below),

\"\"
Global quantum investments in $M (Capgemini)

there has been a surge in corporate partnerships and collaborations aimed at bringing quantum computing out of research labs and into commercial applications:

  1. IBM and JPMorgan Chase: IBM has partnered with JPMorgan Chase to develop quantum algorithms for financial services, with a focus on portfolio optimization, fraud detection, and Monte Carlo simulations.
  2. Volkswagen and D-Wave: Volkswagen has collaborated with D-Wave to investigate how quantum computing can optimize traffic flow and reduce congestion in cities.
  3. Honeywell and Cambridge Quantum Computing: Honeywell has partnered with Cambridge Quantum Computing to develop quantum software tools for commercial use, including quantum machine learning and quantum chemistry.
  4. Microsoft and Airbus: Microsoft has teamed up with Airbus to explore the use of quantum computing in airplane optimization, with the goal of reducing fuel consumption and emissions.
  5. Google and Volkswagen: Google and Volkswagen are working together to develop quantum algorithms for electric vehicle batteries and to optimize traffic flow in cities.

In addition to hardware development, various error-correction methods have been investigated to protect quantum information from environmental noise and other sources of errors that cannot be wholly eliminated by hardware. The development of quantum software and algorithms enabling full advantage of the capabilities of different quantum architectures seems to represent a barrier to widespread technology adoption.

The Quantum Software and Tools Market Is Growing Rapidly

The development of quantum software and algorithms is a key aspect of quantum computing. Although hardware is crucial, it’s the software that unlocks the full potential of diverse quantum architectures.

This is why a number of startups are popping up all over the world that specialize in developing quantum software solutions for a variety of use cases. This market includes software and tools for quantum simulation, quantum optimization, quantum machine learning, and quantum cryptography.

\"\"
Some start-ups operating in the space

But one of the most significant challenges in quantum software development is the need to create algorithms that can run on quantum hardware. This is because quantum computers operate differently from classical computers, and traditional algorithms are not suitable for quantum computing. Therefore, new algorithms and programming languages are being developed to enable the creation of quantum software.

Open Source Quantum Software Development and Incumbents tackling the opportunity

There is an extensive open source community dedicated to advancing the development of quantum software tools:

\"\"
How Qiskit works (Source: IBM)

The aim is to encourage the advancement and evaluation of quantum computer algorithms in the short term. This provides a cost-effective opportunity for newcomers in the quantum industry to start without having to immediately invest in developing their own proprietary solutions.

Key Quantum Computing Companies and Their Progress

By now, you might be wondering what big companies are tackling quantum technology. Quite a few incumbents are making progress. Leaders like IBM, Microsoft, Google and others are striving in quantum hardware such as with quantum error mitigation and dynamic circuits. Governments worldwide are also strategically investing and encouraging quantum research hubs (interesting to see how France is well positioned in that regard).

\"\"
Overview of public funding in quantum technologies (source: https://qureca.com)

We found it relevant to share with you the latest developments from the three main incumbents working on this topic:

Microsoft

Microsoft stands out as a leader in the field of quantum computing, with the goal of building large-scale systems that will empower industries to solve real-world problems using quantum computing. They made progress in 2022 with the development of topological qubits through their work on the topological phase of matter.

IBM

As discussed, IBM is another big player in the quantum computing space. It has made its quantum computer accessible via cloud-based architectures, adding a layer of sophistication to its technology stack. They made strides with Quantum Osprey processor (three times its Eagle version), a major breakthrough in the pursuit of quantum advantage. With 433 qubits, it has set a new benchmark as the most powerful general-purpose quantum computer based on superconducting technology.

Google

Google has made its own push into quantum computing, with the goal of developing an advanced form of artificial intelligence (AI) based on the principles of quantum mechanics. Its innovation in this area includes Google’s Sycamore processor — the first chip built using scalable superconducting qubits with error correction capabilities — as well as its Quantum AI Lab (highlights of the advancements made in 2022), which brings together experts from diverse fields like physics, mathematics, engineering and computer science to work on areas like AI algorithms that could potentially be powered by advanced quantum computers. Latest technological advancement: https://www.nature.com/articles/s41586-022-05434-1

The Future of Quantum Computing and What’s Next

Looking to the future, it’s clear that quantum computing will become increasingly important, with governments and private organizations alike investing huge resources into R&D.

It is likely to enable a new wave of technology applications and use cases, such as:

The potential of quantum computing will also increase efficiency in sectors like biomedicine, energy optimization and climate change analysis. In many ways, it could be a decisive factor in helping us address some of the most pressing global problems of our time.

The key metrics further demonstrate the exponential potential of this technology:

Still, there is a gap between achieving quantum supremacy in theory and utilizing quantum computers practically for optimal machine learning and optimization. Although research is underway, there are considerations to be addressed such as error correction, problem transfer efficiency, and programming complexity.

It appears that quantum computers won’t be a one-size-fits-all solution, and the average person may not experience a “wow” factor like what happened with AI. They won’t be able to replace digital computers for many tasks. Instead, it is more likely that they will complement classical computing by providing new capabilities and applications.

The hype surrounding quantum technology is not at its peak yet, but according to the Gartner curve, it is slowly getting there. Some vendors are already taking advantage of this hype-debut by prematurely promoting the technology for short-term gains, which could result in an early “quantum winter” in about 2 years from now. That’s why Specific hackathons focused on quantum computing are essential for companies to cultivate specialized talent for their specific use cases. This approach helps prevent any disappointment that may arise from the gap between promised results and actual outcomes.

We hold the belief that entrepreneurs are in the best position to identify opportunities for innovation and creation within a given industry. If you have any opinions regarding the future of quantum tech and the areas that will thrive, please do not hesitate to share them with us. Your feedback will always be highly appreciated.

Brief highlight of the cryptographic threats from the perspective of Gaspard Billaud (Cryptologist engineer at Thales), who assisted us in creating this article from the ground up.

If the arrival of the quantum computer promises many revolutions in different sectors, the technological breakthrough that it represents is perceived as a threat to cryptography. Indeed, a quantum computer that would be able to run the shor algorithm (1994) would be able to compromise the security of most information systems. The quantum computer is therefore seen as a “threat” in the world of cryptography.

In 2016, this threat prompted NIST, a U.S. organization in charge of standardizing cryptographic schemes, to launch an international competition for the standardization of post-quantum (quantum computer-resistant) cryptographic algorithms.

Four first algorithms have already been standardized, but the competition is still ongoing (next round on June 2023).

‍

At OVNI, we invest in pre-seed stages and partner with founders who have global ambitions from day one. If you are a founder in this space or know someone who is, feel free to contact me at thomas@ovni.vc.

", "authorSlug": "thomas-renaudin-aoh3t", "mediumUrl": "", "href": "/insights/the-rising-quantum-computing-industry-current-state-and-future-prospects", "topic": "Quantum" }, { "id": "67ecfe3ed9e759345c4aa746", "position": 190, "visible": true, "name": "Cloud Infrastructure Wars", "slug": "cloud-infrastructure-wars-lessons-for-genai-companies", "shortDescription": "The winners in generative AI may be decided by who can keep inference prices low.", "dateLabel": "04/2023", "dateIso": "2023-04-04T00:00:00", "thumbnail": "https://firebasestorage.googleapis.com/v0/b/ovni-website.firebasestorage.app/o/insights%2Fgenerated-posters%2Fcloud-infrastructure-wars-lessons-for-genai-companies-b12165d625bc.jpg?alt=media", "bodyHtml": "

Generative AI is taking the world by storm, but companies’ success may depend on one key factor: the ability to keep prices low. By looking at the cloud infrastructure wars of the past, we can uncover relevant lessons that businesses should consider when navigating their pricing strategies.

With this post, we aim to provide a comparison of the market pressures faced by generative AI companies, drawing parallels with cloud providers and how those pressures will influence their future decisions:

No matter where you stand on this issue, there’s no denying that generative AI companies must make smart decisions in order to succeed in this market — the kind of decisions that will have lasting implications for the industry.

Let’s dive in.

Generative AI companies are competing fiercely in a business landscape where the stakes have not been this high for a long time. Just like the cloud computing industry in the early 2000s, there is a “prisoner’s dilemma” currently in play: each company vies for the largest piece of the market share, market price and margins are pushed down as competition increases, which can prevent any one competitor from maximizing profits.

\"\"
Figure 1. The Prisoner’s Dilemma. Alternative text for the Prisoner’s Dilemma can be accessed here.

Let’s take a closer look at how cloud infrastructure wars of the past can provide insight on what we can expect from Generative AI companies going forward.

The Cloud War: What Happened and What We Learned

The cloud computing war of the last decade pitted tech giants like Microsoft, Amazon, Google and IBM against each other and gave companies countless options to choose from in terms of infrastructure, cost, features and more. The result of this competition was the ever-decreasing cost of cloud services as the giants pushed each other to offer better deals to attract customers. This created a prisoner’s dilemma-like situation in which no company wanted to be the first to raise prices due to fear of losing market share.

In addition, clients benefited from increased transparency as it became easy for companies to compare cloud providers and identify the most appropriate service for their needs. This also led to open source cloud-native projects such as Kubernetes that enabled customers to easily move services between clouds with minimal effort.

A. The Background of Cloud Wars

The cloud computing industry began to take shape in the early 2000s as companies realized the benefits of shifting their IT infrastructure to the cloud. AWS, launched by Amazon in 2006, was one of the first major players in the market, offering cloud services such as storage, computing power, and databases. Other players in the market at that time included Microsoft with its Azure platform and Google with its Google Cloud Platform (GCP).

\"\"
Figure 2. Cloud Ecosystem & Key players

As the market for cloud services grew, competition among the key players intensified, with each provider looking to differentiate itself and capture market share. Pricing was one of the areas where competition was particularly fierce. AWS, Azure, and GCP were engaged in a price war, each provider trying to undercut the other in terms of pricing.

This resulted in significant price cuts over the years, with AWS reducing the price of its EC2 instances by more than 80% since its launch in 2006 and Azure reducing its prices by up to 90% since its launch in 2010. This led to a massive increase in the adoption of cloud services, with businesses of all sizes migrating their IT infrastructure to the cloud to take advantage of its benefits.

The cloud wars were a pivotal moment in the development of the cloud computing industry, resulting in aggressive pricing strategies and the development of new and innovative services, ultimately making cloud services more affordable and accessible for businesses of all sizes.

B. What We Learned From the Outcome

A few lessons can be drawn from what happened in the cloud, especially from big players. For example:

To stay relevant in the industry, it was imperative for each of these companies to establish a distinctive brand identity that set them appart from their competitors. They had to constantly be thinking about how they could differentiate themselves from their competitors and provide unique value to their customers.

So, whether they’re using the latest tech, coming up with creative ways to use AI, or just providing exceptional customer service, GenAI companies need to find special ways to stand out if they want to stay ahead of the competition. It’s all about thinking strategically and finding that special something that sets you apart from the rest.

GenAI: Moving Towards an Unprecedented Price War

A lot of Generative AI companies will face the same dilemma that cloud infrastructure companies faced years ago. Prices for AI algorithms and products are being pushed down aggressively, as larger companies invest heavily in natural language processing, computer vision, and various other related technologies.

With this increased competition, it has become harder for smaller firms to compete in terms of cost.

To make matters worse, many larger firms are willing to offer generous discounts and bundled services to larger customers, creating an unlevel playing field. This leaves startups and other small firms competing from a position of weakness, as they lack the resources necessary to offer similar deals. As a result, many GenAI companies will be forced to lower their prices or risk losing out on potential business opportunities.

Moreover, the cost of developing new AI algorithms is increasing exponentially, making it difficult for smaller firms to remain competitive. This is especially true given the increased complexity of NLP and CV algorithms, which require significant investment in research and development before they can be used commercially. For example, ImageNet-trained models typically require tens of thousands of images for training — an expensive endeavor for smaller firms with limited resources.

A. How Venture Capital Is Driving the Price War

The venture capitalists are driving the price war in GenAI, and it’s not necessarily a bad thing (for customers). Thanks to funding, GenAI companies have additional resources and manpower to scale quickly and invest heavily in research and development. This drives top players to be more competitive with their pricing, which can lead to lower prices for customers.

\"\"
Figure 3. Investor interest in Generative AI since 2017 (Source: CB Insights)

But Venture Capital comes with its own risks. Many startups rely heavily on raising capital to continue their operations, which will create instability as investors may become wary of investing in a highly competitive market.

Investors may fear that their investments will not yield any returns as GenAI companies continually lower their prices, or that their investments will be wasted as these companies struggle to create a sustainable income stream.

VCs may find difficult to predict which company will emerge as a leader in the industry due to its nascent stage of development. More than two-thirds of generative AI companies have not yet raised a Series A round, indicating a nascent market and a lack of clarity on where the true value of the technology lies.

\"\"
Figure 4. Percent of companies by latest disclosed round (Source: CB Insights)

B. What Startups Could Do to Compete

As with the Cloud Infrastructure wars, Generative AI startups need to be aware of the prisoner’s dilemma faced by competitors. The ability to make swift decisions and maintain agility will be critical to stay ahead of the competition.

There are several steps startups can take to gain a competitive edge in this industry:

  1. Invest in scalable technologies that can handle large data sets: Generative AI start-ups should invest in scalable technologies that can handle large data sets. This will allow them to process and analyze vast amounts of data quickly and efficiently, giving them an advantage over competitors who are unable to handle large data sets.
  2. Utilize open source technology whenever possible: Open source technology are a valuable resource for generative AI start-ups, enabling fast iteration in the product. But the core techno/data sets should be proprietary. Arbitrage between what’s proprietary/what’s not will be crucial.
  3. Take a customer-centric approach, providing support for current customers as well as creating new markets: This means understanding customer needs, providing excellent customer service to build strong relationships and brand loyalty and finding use cases that go beyond the typical “xxx empowered by genAI” thing.
  4. Employ a lean system for data science development and deployment: Developing a flexible and adaptable approach to data science, using agile methodologies and continuous integration and deployment to quickly iterate and improve products and services.
  5. Establish clear pricing models and marketing plans to ensure maximum profit potential: Have a deep understanding the market and competition, and develop pricing and marketing strategies that effectively position products and services to attract and retain your customers.
  6. Unique data sets for better outcomes: Consider the uniqueness and diversity of data sets generated by their algorithms as a crucial factor. The capability of producing high-quality and varied data sets can distinguish startups in the genAI market and provide more dependable outcomes for their customers.
  7. Tackle a very niche market and develop it: Carve out a niche market by targeting specific industries or use cases, like vertical software did. Develop a deeper understanding of customer needs, provide tailored solutions and build strong brand recognition within that specific market.

Few companies that are leveraging these strategies: Adept, PhotoRoom, Lalaland.ai, Raidium, Mini Studio, Dust.tt, Replicate.ai, Anthropic, Eleuther.ai

‍

At Ovni, we invest in pre-seed stages and partner with founders who have global ambitions from day one. If you are a founder in this space or know someone who is, feel free to contact me at thomas@ovni.vc.

", "authorSlug": "thomas-renaudin-aoh3t", "mediumUrl": "", "href": "/insights/cloud-infrastructure-wars-lessons-for-genai-companies", "topic": "Infrastructure" }, { "id": "67ecfdf68f7da91ab696776b", "position": 200, "visible": true, "name": "Unveiling our Term Sheet 1.0", "slug": "unveiling-our-term-sheet", "shortDescription": "We’re publishing our Term Sheet 1.0 to make how we invest clearer for founders.", "dateLabel": "03/2023", "dateIso": "2023-03-27T00:00:00", "thumbnail": "https://firebasestorage.googleapis.com/v0/b/ovni-website.firebasestorage.app/o/insights%2Fgenerated-posters%2Funveiling-our-term-sheet-72c3a387b8ba.jpg?alt=media", "bodyHtml": "

At OVNI Capital, we strive for transparency. And we leverage our proprietary technology to provide founders with assistance in making initial hires, connecting with essential stakeholders, and obtaining valuable insights to expedite their commercial progress.

Our aim, similar to YC’s SAFE creation, is to promote quick business discussions before the round even closes.

With this goal in mind, we are happy to offer you one of the most essential components that shape the relationship between a VC and a founder: the Term Sheet.

\"\"

The Term Sheet is a non-binding agreement that lays out the terms of the investment, including the amount of funding, the valuation of the company, and any other conditions that need to be met before the investment is made. Sometimes it is also called a LOI (i.e “letter of intent”), but don’t worry too much about the jargon.

While term sheets usually aren’t legally binding, in the venture capital world, people take these very seriously. It’s common practice for VCs to stick to the terms outlined in the term sheet, even if circumstances change down the line. This is because VCs don’t want to damage their reputation by going back on their word.

As a founder, that’s something you should keep in mind in case a VC tries to change the terms later. Don’t let them use the excuse that the market has changed to lower the value of your company.

Our sole focus: facilitate your Series A round

We call ourselves venture capitalists for a reason. We believe in power law: few will succeed and return the fund, it’s our job to make the process seamless and help you reach Product-Market-Fit.

Here is the term sheet, feedback highly appreciated!

‍

Working on something new ? Feel free to reach out: dealflow@ovni.vc

", "authorSlug": "augustin-sayer", "mediumUrl": "", "href": "/insights/unveiling-our-term-sheet", "topic": "Venture" } ], "orderedIds": [ "c5577b25b3f7", "f3b042cfff77", "d63aaa0e9613", "b505409b99f0", "85689d8ad733", "b6a6dc02574b", "68c43101ec9e6cdffa02f66f", "e7d1877dd518", "11a5ef219d69", "bd96b2f335af", "6ebeef45c776", "209bdf8de0ae", "67d990ba54a06d958945fb90", "67dae206d04d2e2333444eb0", "67db19f2e30c009ad68c476c", "67ecf93c484f2ab1509907a6", "67ecfcc3c636180733e85bb9", "67ecfd6f17e207672077de62", "67ecfe3ed9e759345c4aa746", "67ecfdf68f7da91ab696776b" ] }