The AI Funding Slowdown Nobody's Talking About
AI deal volume dropped 23% in Q1 while check sizes hit record highs. What the divergence signals for founders and investors.
The Number
$4.7 billion — that's how much AI startups raised in February 2026 alone. But here's the catch: deal volume dropped 23% quarter-over-quarter. Fewer companies are raising, but the ones that do are raising massive rounds. The median Series A for AI companies hit $28M, nearly double the cross-sector median of $15M.
What this signals: capital is concentrating in fewer, larger bets. If you're an AI founder without clear product-market fit and revenue traction, the window is narrowing. The spray-and-pray era of AI funding is over.
The Breakdown
Why VCs Are Pulling Back on AI Seed Deals
The AI seed market is experiencing something unusual: abundant capital but declining deployment. Three forces are driving this:
1. Model layer commoditization. When GPT-4, Claude, and Gemini are all good enough, the moat moves from 'we have a better model' to 'we have proprietary data and distribution.' Most seed-stage AI companies haven't figured this out yet.
2. Margin compression. AI-native companies are burning through inference costs that look nothing like traditional SaaS unit economics. A chatbot doing $2M ARR might have 40% gross margins — not the 80%+ VCs expect.
3. Portfolio overlap. Every fund already has 3-5 AI bets from 2023-24. Adding more without conviction creates portfolio construction problems, especially for smaller funds.
The winners in this market: vertical AI applications with proprietary datasets and clear revenue — think AI for insurance underwriting, legal document review, or clinical trial matching. Horizontal 'AI for everything' plays are getting passed on.
Deal Anatomy
Inside Cognition's $250M Series C
Cognition — the company behind Devin, the AI software engineer — closed a $250M Series C at a $4B valuation. Here's what makes this deal interesting from a VC lens:
The round was led by a single investor (Founders Fund) with minimal syndication — a sign of high conviction and competitive dynamics. At $4B on roughly $15M in ARR, this prices at ~267x revenue. That only makes sense if you believe AI coding agents will capture a meaningful share of the $500B+ global software engineering market.
For founders: this deal structure (single lead, minimal dilution from syndication, massive valuation premium) only works when you have extreme leverage. Cognition had 5+ term sheets. Most founders won't have that luxury — which is why understanding competitive dynamics matters more than any valuation formula.
Tool of the Week
Fund Return Model Calculator
If you're an investor (or a founder trying to understand how your VC thinks), our Fund Return Model lets you simulate portfolio outcomes. Plug in your fund size, number of investments, and expected return distribution — then see how the math works out across different scenarios. It's the fastest way to internalize why VCs need 100x outcomes, not 3x.
The Edge
Three signals worth watching this week:
1. Stripe is quietly rolling out embedded lending to its platform customers — a move that could disrupt fintech VCs' portfolio thesis around vertical lending startups.
2. YC's latest batch is 40% international founders, up from 25% two years ago. The talent arbitrage is real, and funds without global sourcing are missing deals.
3. Secondary market pricing for late-stage tech companies is diverging sharply from public market comps. Private valuations are lagging reality by 6-9 months — creating opportunity for patient capital.