TechCrunch went through October's SEC filings and found a busy month for venture funds. The largest new AI vehicle is Diffusion ONE, the debut fund from former Benchmark partners Victor Lazarte and Kris Fredrickson, which raised $1.252 billion against a target of $800 million reported in May. It has already co-led Harvey's $550 million Series H with Lightspeed and joined Etched's $300 million Series C. Elsewhere, 8VC filed for a $400 million fund plus a small AI Fund XI that has raised $13.5 million, Benchmark filed four partner-only funds, and Industry Ventures raised at least $1.7 billion for secondaries.
A first-time fund clearing its target by more than 50 percent tells you limited partners are not spreading AI money evenly. They are backing a few managers with access to the most crowded rounds. Two of Diffusion's named deals are a legal AI company and a chip company, which matches what I see: capital is concentrating at the application layer where revenue is visible and at the hardware underneath it.
For smaller investors that cuts both ways. Late rounds in names like Harvey are priced for funds writing nine-figure checks, so there is little room for the rest of us there. The opening is earlier, with founders solving a narrow, paid problem before the large funds arrive. That is where I focus through ELP Ventures, and it is the same logic behind iQuash, which applies AI to one specific job: checking tickets and fines for defects.
The secondaries figure matters too. When a fund raises $1.7 billion just to buy existing stakes, it suggests a lot of early holders want liquidity before exits arrive. In private equity at EMC Acquisitions we price that patience into every deal, and I expect AI investors will have to do the same.