VC balances AI potential with avoiding its own extinction
Add Axios as your preferred source to
see more of our stories on Google.

Illustration: Aïda Amer/Axios
Venture capitalists are in an odd spot: Financially motivated to pump billions of dollars into AI startups that could hasten their own extinction.
- But not everyone is fretting propagation of the sector.
Driving the news: Marc Andreessen said on a recent podcast that he could envision venture capital being "one of the last remaining fields that people are actually doing."
- Andreessen's argument was that venture is more art than science, as evidenced by low batting averages of even the most successful investors.
- Sure, it comes off as self-serving. Like a frog croaking about its swimming ability as the water heats up. But that doesn't necessarily make him wrong.
The big picture: What we know for sure is that some VC firm functionality is already being automated, particularly in the back office, and investors themselves are increasingly using AI for due diligence.
- There's also less need to hire junior employees in what has historically been an apprenticeship business — raising pipeline questions about the next generation of leaders, although neither Andreessen nor Ben Horowitz traveled that path.
Zoom in: Venture capital is a two-part job. The picking and the post-pick management.
- Horowitz acknowledges on the same podcast that AI might eventually be better at picking, but argues that "who gets to pick" is just as important — and that access is tied to a VC's relationships and reputation.
- Moreover, they both argue that post-pick management is as much about a founder's psychology as the startup's income statement or product roadmap. In other words, it still needs a little of that human touch.
Yes, but: AI technologies, including those funded Andreessen Horowitz, seem designed to replace as much of that humanity as possible. If not explicitly, than certainly implicitly.
- Just yesterday, Mark Zuckerberg talked about how AI could fill the void for those who wish they had more friends. That's a far cry from automating rote tasks or making a better analytical mousetrap.
Between the lines: It feels that if many of these venture-backed startups reach their full potential, it would include both arts and sciences — or at least enough of it that institutional investors will no longer see the value of paying 2-and-20.
- After all, who needs human relationships when AI agents can just connect with other AI agents?
The bottom line: This still is mostly theoretical, and venture capital's future isn't the most pressing labor, economic, or social concern posed by AI.
- But venture capitalists are among the very few who are directly responsible for AI acceleration, no matter how it plays out.
