A non-AI startup raising a Series A in 2026 is being priced at a median of $55 million. An AI startup at the identical stage is being priced at a median of $300 million, a roughly 5x spread on Carta's Q1 2026 dataset, according to AngelInvestorsNetwork's analysis of that data. That gap did not exist at this size a few years ago, and it is the clearest sign yet that seed and Series A valuations have stopped moving together. They are splitting along a line drawn by one word: AI.
The money is concentrating, not spreading
US startups raised more than $412.7 billion in the first half of 2026, a total that already surpasses every previous full-year record, and megadeals of $100 million or more captured 87.5% of it, with AI accounting for 86% of all venture dollars deployed, per the PitchBook-NVCA Venture Monitor for Q2 2026. Three firms, Andreessen Horowitz, Thrive Capital, and Founders Fund, accounted for 48.1% of all venture capital raised over that same half, according to the same report. When close to half the market's dollars flow through three checkbooks writing large, AI-heavy rounds, the pricing signal at seed and Series A stops reflecting broad demand and starts reflecting whether a company can plausibly wear the AI label.
Pre-seed is compressing while Series A hits an all-time high
Follow the stages in order and the divergence gets sharper, not softer. Median pre-seed pre-money valuation was $7.7 million at the end of Q3 2025, down slightly from $8.0 million the prior quarter, per Zeni.ai's analysis of PitchBook-NVCA data. At seed, PitchBook-NVCA put the median pre-money valuation at $18.4 million in Q1 2026, with AI-labeled seed companies at $18.7 million against $18.0 million for non-AI companies, a gap of a few hundred thousand dollars, not yet a chasm, per the Q1 2026 Venture Monitor. Carta's own numbers show more separation building underneath that median: seed post-money valuation reached $24 million in Q4 2025, up from $18 million in Q4 2024, a 50% jump in two years that Flowjam's reporting on Carta data attributes largely to AI mega-rounds rather than gains spread across the whole seed population. Separately, seed-stage AI startups were pricing 42% above non-AI baseline valuations in early 2026, according to AgentMarketCap.
Then comes Series A, where the premium stops looking like a premium and starts looking like a different market. Median Series A pre-money valuation on Carta hit an all-time high of $49.3 million in Q3 2025, while median seed pre-money sat at $16 million that same quarter, already a three-fold jump between adjacent stages, per Carta's State of Private Markets Q3 2025. By Q1 2026, an AI foundational-model startup was raising its Series A at roughly $300 million median while a non-AI company at the same stage sat near $55 million. The blended Series A median investors quote in headlines hides two populations that no longer share a market.
The bar for actually clearing that price went up too
The premium is not free money. Carta's Q1 2026 data puts the Series A annual recurring revenue bar for AI startups at $3.5 million, nearly 3.5 times the roughly $1 million ARR bar startups needed three years earlier to clear the same round, according to Value Add VC's review of Carta's numbers. Investors are paying AI-scale prices, but they are also demanding AI-scale revenue proof before they write the check. A founder who raises seed at an AI-inflated valuation and shows up at Series A without real ARR is not competing for the $300 million bucket. They are competing in the $55 million bucket, against a valuation expectation their own seed round already set too high.
Roughly 3,000 US-based startups raised pre-seed funding in Q1 2026, totaling more than $2.3 billion, according to Carta's State of Pre-Seed report, meaning most companies are still starting in a market where the AI premium has not shown up yet.
That last figure matters more than the headline spread. The AI premium is real at seed and enormous at Series A, but it concentrates in a narrow slice of the thousands of companies raising pre-seed and seed each quarter. Most founders are still pricing in the $7.7 million to $18 million range no matter what label they use in the deck.
What to do with this before your next round
If you are building something genuinely AI-native, price your seed round with the Series A gap in mind. Investors already know the jump from an $18 million seed to a $300 million AI Series A requires $3.5 million in ARR, not a demo, so build your fundraising timeline backward from that revenue number rather than from whatever valuation your seed investors handed you. If you are not AI-native, do not chase AI-labeled comps from press coverage. Price your round against the $55 million non-AI Series A median and the $16 million to $18 million seed range, and raise on terms that survive a flat round if a $300 million comp was never yours to hit. This week, pull your last 90 days of ARR growth, plot it against the $3.5 million bar, and decide honestly which of these two markets you are actually in before you put a number on your next term sheet.