Everyone is rushing into growth capital. They're running toward a closing window.
In the second half of 2025, I noticed a pattern in our fundraising conversations for Fund IV. One after another, institutional LPs, mainly from Europe, told us some version of the same narrative: they had made a strategic decision to move away from seed funds and concentrate their venture allocation entirely on growth VCs.
It was frustrating at times, but I acknowledged their pursuit to maximize returns by capturing "the bigger opportunity" growth capital presented — especially after years of scarcity in Europe. However, something about this narrative was off. I was not convinced.
Initially, I thought that it was my unwavering belief in seed as the true venture — a bias that probably blinded me to the reality that seed is going out of fashion… again. Then I thought it was the 2021–2022 LP fatigue from overexposure to seed funds with no results to show.
But I had my aha moment a few weeks ago, when I sat down with my team to review the quarterly performance of our Fund III portfolio. What we saw in the data of the last cohort of investments stopped me cold, and suddenly, the awkwardness I'd been carrying through those LP meetings had a name.
The institutional capital is moving toward growth just as growth is ceasing to exist in the form they expect to find it.
The window opened — then AI walked in
The diagnosis that drove that LP consensus was correct at the time. European software scaleups were genuinely starved of growth capital. The Series B and C gap was real; you could see it in the deal count data. While seed activity remained resilient through the 2022–2024 correction, late-stage and growth-stage funding collapsed. Late-stage European VC fell from $76 billion in 2021 to $25.7 billion in 2023, a 66% drop in two years. The need was acute. The calls from founders, investors, and policymakers were justified.
But execution in Europe is phenomenally slow. There is reliably a 2–3 year lag between the moment a structural gap is identified and the moment capital is organized to fill it. By the time the funds are raised, the strategies articulated, the investment committees aligned - the world has moved on. The gap that needed filling in 2022 is now being filled in 2026 by funds that will deploy through 2027–2030.
Except for the software companies, they will find that in 2027–2028 will not look like the companies that needed funding in 2022. The reason is straightforward: the advancement of AI fundamentally changed what it costs to build, run, and scale a software company by a magnitude.
A new physics of company building
The AI-native company operates under a different set of laws. Where a pre-AI SaaS company at the growth stage might employ 150–200 people, an AI-native equivalent reaches the same revenue with 35–40. Where it previously took a few years to find product-market fit, AI-native teams are doing it in months. Where growth used to require raising a Series B to fund a sales motion, the new model reaches profitability within the first year of operations.
A few trivia facts about the superiority of AI-native companies vs SaaS ones:
- 4x faster growth
- 7–8x fewer employees per dollar of revenue
- net revenue retention of 120–132% versus 108% for mid-market SaaS peers.
- leading AI-native companies already exceed $1M in revenue per employee compared to a $175,000–615,000 range for scaling SaaS firms.
Cursor and Lovable were the first to shock the market with their exponential growth and lean operations. It was extraordinary in the beginning, and many thought this space was reserved only for the exceptional ones. But then others followed. Now it has become the expected baseline for AI-native companies. However, the market is still catching up to what that means.
I'm not speaking theoretically. We see it increasingly in our own portfolio. Native Teams, a 2022 investment, reached the same revenue milestones as Payhawk, one of our best-performing companies from the 2018 vintage, in a shorter period and with a few times less capital, provided only by the seed investors. As a bonus, they hit profitability in 2025 as well.
Fund III data makes the case even more intriguing: the last cohort of investments is either cashflow positive or on the path there within their first year of operations. That number changes the whole conversation about what seed capital can actually achieve.
Initially, we thought that Native Teams were the exception. But apparently, this is becoming a pattern, and the implications are fundamental. This opens an entirely new universe of strategic options for growth for AI-native companies not present in the entire venture history.
A company that reaches profitability on seed funding may never need a growth round.
Or it may choose, at the moment of its choosing, on its own terms, to raise one opportunistically, when the market is favorable, and the strategic logic is clear. Or it may go for alternative funding options like venture debt or commercial debt. That is a very different animal from a company that requires a Series B to survive. The growth funds being raised today are optimized for the latter category. But the category is shrinking.
The CEE accelerant
Nowhere is this dynamic more acute than in Central and Eastern Europe — and this is no coincidence. It is structural.
Capital efficiency has always been part of the CEE founder's DNA. Although CEE startups comprise 20% of the European ecosystem value, they attract only 3% of European venture capital. Due to limited access to institutional capital and thinner private LP networks, CEE VC funds are smaller in size and fewer in number, with smaller checks. So over the years, CEE founders learned how to be resourceful and do more with less to survive and potentially thrive. CEE has the biggest percentage in Europe for bootstrapped startups.
Over the years, many great companies have been created in CEE — Avast, Telerik, Siteground, Payhawk, UiPath, Blueground, and Docplanner, to name a few. What struck me every time wasn't their ambition but their instinctive resourcefulness. None of these founders built on the assumption that unlimited capital was available, since it was not. It was the major growth restriction, and they had to overcome it in the smartest possible way.
Now, AI has taken that instinct and given it an engine. The result is a compounding effect. CEE's natural tendency toward capital efficiency, amplified by AI, accelerates an already-accelerating global trend. Companies here are not just reaching profitability faster than the European average. They are redefining what "growth stage" means in practice. In many cases, they are skipping it entirely.
AI just killed the VC "factory line", an obsession of a generation of founders and investors. Today, a seed round of $3–10 million is no longer just enough to find product-market fit. It is enough to reach it, pass it, hire the first revenue team, hit sustainable unit economics, and begin compounding. The CEE startups are thriving in this new world.
We are watching the seed stage absorb what used to be the job of the growth stage. The shelf life of a company's dependency on external capital has compressed from a decade to three or four years. The old growth round is being displaced — from below, by leaner companies that don't need it, and from above, by something altogether different.
The super growth round is not a growth round. It's a category winner picker game
What remains of the growth stage is undergoing its own transformation. The large rounds being written today — $300 million, $500 million, $1 billion+ — are not growth capital in the traditional sense. They are not financing the next phase of a business that needs to hire and scale. They are category-declaration events.
Their purpose is threefold:
- First, to signal to the market that this company is the chosen winner in its space, the round is as much a press release as a financing.
- Second, to scare competitors into capitulation or consolidation.
- Third, to fund M&A and marketing attrition, not product development.
This is a fundamentally different activity from venture capital. It is closer to private equity, to investment banking, to the kind of structural capital allocation that large institutions do when they are picking platform companies rather than backing entrepreneurs. The investors writing these checks are not seed VCs who moved downstream. They are sovereign wealth funds, large crossover funds, and the US mega-firms that effectively became RIAs in recent years.
The middle — the traditional growth fund playing the Series B/C game — is being squeezed from both ends. I've been saying this in LP meetings for the better part of a year, and I'm starting to see others in the ecosystem name it publicly, too. From below, by the AI-native seed company that never shows up needing their capital. From above, by the super-growth round that requires a check size they cannot write, and a network they do not have.
What this means for LPs
If you are an LP currently evaluating a European growth fund, this is the question you should be asking: which companies does this fund expect to invest in? If the answer is software companies at Series B and C, press harder. Because the software companies raising traditional Series B and C rounds in 2026 and 2027 are disproportionately going to be the ones that couldn't reach profitability on their seed capital — the ones that failed the new efficiency test.
That is adverse selection at the portfolio level before a single check is written. The better frame for LPs with a genuine appetite for growth-stage returns is this: the growth opportunity now lives inside large seed funds.
That's precisely how we've structured our next fund — €100 million+, with co-investment capacity built in from day one — not as a feature, but as the core of the LP value proposition. A seed fund with the portfolio quality and LP relationships to offer co-investment at the moment a company decides to raise a larger round - on the company's terms, at the company's chosen moment is a structurally superior access point to growth returns than a dedicated growth vehicle grinding through a Series B pipeline.
The co-investment right attached to a top-quartile European seed fund is, increasingly, the growth allocation. LPs who understand this early will have access. Those who wait for the dedicated growth fund to prove itself will find that the best companies never needed it.
Rewriting the scorecard
All of this comes at a cost to our industry. It requires us to redefine the way we measure success along the way before DPI kicks in.
Traditional VC KPIs at the seed stage are Series A/B rate (and respective mark-ups) — how many of your companies raised a priced institutional round? It is a reasonable proxy for quality in a world where the path to value runs through successive financing rounds — the VC “factory line”. But in a world where the best companies skip growth rounds entirely, the Series A/B rate tells you very little. A high Series A rate at a seed fund might now indicate that the portfolio needed follow-on capital, which is precisely the wrong signal.
Internally, we've already started shifting how we talk about portfolio performance. The question we ask about a company isn't "what round are they raising next?" It's "do they need to raise at all?"
The KPI that actually matters in the AI era is simpler and harder: how many of your companies reached profitability independently, with high retention? Not with the help of the next round. Not through acquisition. On their own. That is the measure of whether you backed companies with real models, real margins, and real futures.
The contrarian bet
So my contrarian bet is seed. Specifically seed funds with:
- the discipline to back AI-native companies
- portfolio construction to own meaningful stakes (at least 15%), and
- the LP relationships to offer co-investment at scale
- focus on CEE, especially, where the capital efficiency advantage is structural, and the AI-native transition is happening faster than anywhere else in Europe.
Seed is not just the new early stage. In the age of AI, seed is the new growth.
I'm writing this because I believe it — and because I'm putting our next fund behind it. That's the only honest way to make a contrarian argument: not just to say it, but to stake something on it.
The window for that realization is also closing — just more slowly than the growth capital window. The funds building toward it now will be the ones LPs wish they had backed in 2030.
A clarification note: my thoughts consider software companies not so much deeptech or capex-heavy companies. At least for the time being. This might change in the future as well.
