There Are Two Seed Markets Now — and Founders Need to Pick One
The PitchBook-NVCA Venture Monitor for Q1 2026 has one of the more important quiet data points of the year buried in it. The seed round, as a category, has stopped being a category. It has become two.[1]
The headline number — a $3 million median seed — looks unremarkable.[1] What's interesting is what's underneath it. The distribution is no longer a single bell curve. It's bimodal.
Two peaks, one valley
The first peak sits where seeds have always sat: $2–4M, 18–24 months of runway, pre-product or just-launched, priced on team and thesis. Nothing new there.
The second peak — smaller, but thickening fast — now sits at $20M and up. These are "mega-seeds" that are functionally Series A rounds wearing a seed label. In my experience, they're deployed almost entirely into AI-native companies where the founding team has previously exited, comes from a top frontier lab, or is building infrastructure that needs compute commitments before line one of revenue.[2]
The valley in between — the old $7–12M "priced seed" — is where the awkward rounds live. Too big to be scrappy. Too small to fund the AI play.
Why this matters for founders
The two markets reward opposite founder behaviors.
The classic seed selects for capital efficiency, fast cycles, and revenue discipline. You preserve dilution for Series A and B. The mega-seed selects for founder pedigree, model access, and a willingness to burn $15M before product-market fit in exchange for a 24-month head start. Different game, different rules.
The practical implication is that the worst place to be in 2026 is the middle. A $9M seed at a $40M post is, structurally, harder to graduate than either a $3M / $15M post (clean, efficient, clear path to A) or a $25M / $150M post (treated as a Series A by the market on graduation). The middle is where down rounds incubate.
In sum: the right question for a founder raising now isn't how big the seed should be. It's which seed market they're in.
Why this matters for investors
The bimodal seed reframes portfolio construction, and I suspect a lot of funds haven't internalized this yet.
The classic-seed portfolio works on the old math: many small bets, power-law exits, low check sizes. The mega-seed portfolio is a Series-A portfolio with a different label — concentrated, conviction-driven, underwriting on team and model access rather than traction. Funds trying to do both at once are, in practice, doing neither well.
The takeaway
The "average seed round" was always a statistical fiction. In Q1 2026 it became an actively misleading one. There is no longer a default answer for what a seed looks like — and the founders and investors who keep behaving as if there is one are the ones most likely to end up in the valley.
Sources
- PitchBook-NVCA Venture Monitor, Q1 2026. $3M median US seed deal value. https://nvca.org/pitchbook-nvca-venture-monitor/
- PitchBook. "Mega-seed rounds make headlines—and potentially hurt startups" (May 2026). https://pitchbook.com/news/articles/mega-seed-rounds-make-headlines-and-potentially-hurt-startups