“Half of these AI companies probably don't exist 10 years from now. And it's not clear how many have a defensible moat around their business. It's not clear to me what the ultimate profit pools look like. If you take the top five large language models and they all had the same user interface for 99% of tasks, you probably couldn't tell the difference.”
The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch · Startups & Venture · November 2026
Context
Harry Stebbings asked Wilson, who runs a university endowment, what he meant by calling some private-market pricing stupid. Wilson questions whether the model companies have durable profits, a view that bears on the AI IPO valuation debate.
Transcript
Could be. It would obviously be a lot of the same potential buyers, but I mean, I think that that's tough to kind of beta test.
You said about stupidity within private markets. What did you mean by stupidity? Like, what are you pointing to? Is it valuations? Is it deployment speed? Yeah, it's
just valuations. I'm sure that half of these kind of, you know, I guess super intelligence is the correct term now, but these AI Companies, like probably half of these companies don't exist 10 years from now. And it's not clear how many have like a defensible mode around their business. It's not clear to me what the ultimate profit pools look like for these companies. I think it's a super strange dichotomy. If you take the top five kind of large language models and they all had the same user interface for 99% of problems or tasks, like you probably couldn't tell the difference, right? You wouldn't know which model you're using. Like I can't describe another industry like that that has a big kind of deep profit pool.
Do you have much exposure to OpenAI or Anthropic, the model providers?
We have pretty decent exposure to Anthropic through a few of our partners.
With the greatest of respect, as you said, the ability and willingness to switch, the commoditization. Do you ascribe little value then to that model provider layer?
More from The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch
“The problem with artificial intelligence is that it can do everything. And you need a source of focus.”
sputnikintelligence.com
“If you want GDP growth, you need more production. You don't need more intelligence. You need more production. In order to have more production, you need better factories… better planes that are flying for longer with the same amount of oil.”
sputnikintelligence.com