The gatekeepers of capital used to block out two hours for a preliminary analysis of a pitch. Now they do it in 13 seconds.
A new peer-reviewed study has analysed five years of deal-flow from German deep-tech fund Freigeist . Researchers ran an AI agent against 61,814 startups to see if it could match human intuition.
The result feels more like a structural shift than a simple productivity gain.
The AI screened deals 537 times faster and at 0.1% of the cost. Crucially, the start ups it flagged were more likely to survive and raise follow on funding than the baseline.
This is the industrialisation of intuition.
Venture Capital has always been a blend of quantitative rigour and human gut feel. For decades, the first filter was often human. In some ways, it was artisanal.
Now, the first filter is becoming algorithmic, fed by data from PitchBook, Crunchbase and Dealroom.co.
What this suggests:
-
๐๐ก๐ ๐๐๐ฏ๐๐ฅ๐ฎ๐๐ญ๐ข๐จ๐ง ๐จ๐ ๐ญ๐ก๐ ๐ฐ๐๐ซ๐ฆ ๐ข๐ง๐ญ๐ซ๐จ: if an AI can scan 10,000 decks a night with zero fatigue, the advantage shifts from who you know to what data you hold.
-
๐๐ก๐ ๐ฌ๐ก๐ข๐๐ญ ๐๐ซ๐จ๐ฆ ๐ฌ๐๐ฆ๐ฉ๐ฅ๐ข๐ง๐ ๐ญ๐จ ๐ฆ๐๐ฉ๐ฉ๐ข๐ง๐ : we are moving from a world where investors spot check the market to one where software maps the entire universe of deals.
In an age where capital is screened by the same machines, the real edge may belong to the investor willing to read the pitch the model rejects.
