Large First Rounds Draw Scrutiny
Yann LeCun raised $1 billion for a company that did not exist a week earlier. Project Prometheus launched with $6.2 billion. Unconventional AI secured $475 million two months after founding. These rounds have prompted questions about whether seed financing has changed permanently.
Bison Ventures examined every publicly available $100 million-plus first round over the last 15 years, compiling roughly 200 deals. Only 20% of those companies recorded exits. Of the exits, only a few produced 10x MOIC or better for the first-round investor. Approximately 1% of the full set generated returns that justify the venture asset class.
Biotech History and AI Parallels
Biotech has long featured large first rounds because Phase 1 trials cannot be funded with $3 million checks. The return profile for first-check investors has been mixed, with a handful of strong outcomes and a long tail of modest ones. The same pattern appears in AI data.
OpenAI and Anthropic are projected to double the number of outlier returns in the dataset upon exit. First-round investors in OpenAI are expected to realize 30-40x returns at projected IPO valuations. By comparison, Sequoia Capital and Kleiner Perkins each converted roughly $12.5 million invested in Google into around $4 billion, for returns north of 300x. First Round Capital converted a roughly $500,000 check in Uber into $2.5 billion, or nearly 5,000x.
Traditional Rounds Continue to Scale
The number of $50 million-plus seed rounds has grown sharply since 2018, yet traditionally sized first rounds are also increasing. Cursor raised less than $10 million in its first round. ElevenLabs raised $2 million. Legora raised $11 million. Sierra raised $25 million. Cohere raised $5 million. Each of these companies is now valued above $5 billion.
High entry valuations leave less room for upside to compound for early investors, regardless of sector. The dataset shows capital intensity has not produced a higher rate of venture-scale exits.
according to Crunchbase News. The companies frequently cited as current AI standouts began with modest first rounds, reinforcing the same distribution seen across prior cycles.
Entry Price Determines Scale of Returns
according to Crunchbase News. The difference between 30-40x outcomes and 300x or 5,000x outcomes has been entry price, not company quality alone. A few mega-seeded companies will still deliver 10x-plus MOICs, consistent with every prior technology wave.
according to Crunchbase News. Building a portfolio around the exceptions rather than the broader pattern has produced a long record of underperformance across 15 years of data.
