Silicon Valley Bank’s (SVB) 30th State of the Markets Report, released at the turn of the year, confirms what the data has been quietly whispering for nearly two years: we are not currently in a broad-based tech recovery.
What we’re witnessing instead is a winner-take-most land grab executed at the category level.
The headlines say “VC is back.” The reality is more precise – and far more brutal. Venture capital is back for a handful of category-defining platforms, while the long tail of startups is being left behind.
When nearly two-thirds of venture capital flows into $500M+ mega-deals, the signal is clear. Strip out OpenAI-scale bets and venture activity now below $100M is essentially flat versus pre-pandemic levels. Investors aren’t spreading risk – they’re concentrating belief.
From a category lens, this implies something uncomfortable: many investors believe most meaningful tech categories are already “spoken for.” That’s a problem – not just for founders, but for the long-term health of innovation itself. New categories need to be created. The open question is whether the conditions to create them currently exist.
Right Now, Category Creation Has Fundamentally Changed
Between 2018 and 2020, it was still possible to build a credible challenger inside an emerging SaaS uber-category. In 2025, unless your company is AI-native, platform-shaped, and strategically central, you’re mostly fighting for leftovers.
This shift should radically change how founders think about category creation in the short term.
It also raises an uncomfortable thought: do startups even need VCs for growth to exit anymore? With the rise of creative acquihires – licensing deals, talent grabs, and regulatory-skirting partnerships – growing a small, strategically relevant company and aggressively courting Big Tech may now be a more rational outcome than building a standalone category challenger.
AI Investment: Capital Dominance Without Operating Discipline
AI absorbed 58% of all VC dollars last year. Yet AI companies are less efficient by almost every operating metric: Lower revenue per employee; worse margins; higher burn.
This isn’t accidental. It’s what happens when investors fund AI not as a productivity layer, but as a speculative category land grab.
We’ve seen this movie before. Not just during the dot-com era, but during infrastructure cycles like the move to cloud (2008–2012), when investors weren’t optimising for efficiency – they were underwriting future monopolies alongside existing power players.
This is where founders currently misread the narrative. Yes, AI tooling is cheap to build and fast to deploy. But winning a foundational AI category – models, data layers, agent platforms – requires aggressive burn and patience. The result is a distorted market: a few massively funded winners and a long tail of AI apps permanently stuck at feature status.
The Maturity and Category Compression Problem
One of the harshest signals in the SVB report is the change in graduation rates – it now takes 10 years on average to go from Seed to Series D.
That’s not just a capital problem – it’s a category maturity problem.
Fewer companies scale because fewer categories are expanding fast enough to support multiple winners. As recently as 2020, early category momentum could carry a Series A. In 2025, investors want proof that the category itself is compounding, not just your product.
Bridging rounds – once a red flag – are now normal, even among top-quartile companies. The implicit message is clear: founders should expect to be “early” in their category for much longer than before.
The “Dead in the Water” Category Trap
SVB highlights a sobering statistic: around 20% of VC-backed companies are neither growing nor profitable.
Viewed through a category lens, this is rarely a team failure. It’s a category misfire. The usual culprits are familiar: Over-segmented mature SaaS markets; AI wrappers with no durable differentiation and “Markets” that turned out just to be features inside larger platforms
Once a category stops expanding, capital exits first. And, crucially, today’s market is no longer cyclical enough to save companies that just hang on. If your category isn’t structurally attractive now, there is no rebound coming.
Rising Revenue Bars Signal Category Fear
The revenue benchmarks tell the same story. A top-quartile Series C company now needs $45M in revenue, up 65% in just two years.
This isn’t confidence – it’s fear.
Investors are pricing-in category uncertainty. Optionality has vanished. The new message is: “Show us you’ve already won a meaningful share of a real market.” It raises an awkward question: if VCs won’t underwrite category risk anymore, what role are they actually playing?
The old playbook – raising on vision and a compelling category narrative – has been replaced by revenue density as proof of legitimacy.
AI Valuations: A Company Bubble or a Category Bubble?
SVB doesn’t mince words: AI valuations are almost certainly in a bubble. No kidding.
But this isn’t random speculation. It’s a new kind of bubble – a category compression bubble. Capital is being forced into fewer perceived “core” AI categories and a small set of incumbents to cement dominance. The top five AI unicorns are now worth more than every dot-com IPO combined.
Most AI startups, implicitly, aren’t being valued as businesses – they’re being valued as options. And options, by definition, either expire or get traded.
Immigration and the Category Pipeline Problem
Nearly 60% of the top 100 US unicorns were founded by immigrants. The same pattern holds in the UK and Europe.
From a category creation perspective, these founders represent the future supply of high-conviction, nothing-to-lose, free thinking outsiders – the very force that made Silicon Valley and other tech clusters work in the first place. If immigration slows, the damage won’t be immediate. It shows up five to ten years later as fewer new categories emerge and innovation becomes incremental.
That’s how ecosystems stagnate, how power consolidates and, ultimately how whole economies and societies decline.
IPOs, LP Pressure, and Category Cleanup
The IPO window may be reopening, but average growth at IPO is now 9%, down from 28%. Public markets aren’t funding category innovation anymore; they’re rewarding mature incumbents.
At the same time, LP liquidity pressure is reshaping venture behaviour. When LPs need cash back, VCs push for earlier exits, secondaries, and faster consolidation. That favours acquirable companies inside existing categories – not long-horizon category creation.
Yes, M&A activity is up, but it’s not growth M&A. Only 7% of deals return 3x+ capital. This is a category cleanup: platforms absorbing talent, IP, or customers to reinforce dominance. Windsurf-style deals – buying leadership but not the company – are the clearest warning sign of all.
The message is brutal: our category mattered; your company didn’t.
What’s the Lesson Here?
This is venture capital – but not as we’ve known it.
Capital is flowing, but only to categories perceived as massive, defensible, and monopolistic. Everything else is being starved, consolidated, or quietly shut down.
For founders and operators, the real question is no longer “Is VC back?”
It’s this: Is my category actually worth winning – or am I just building a faster route to someone else’s exit?
Whatever the answer, one thing is certain: it’s still all to play for – but, for now, at least, the rules have changed.