AlphaMaven AlphaMaven

The head of one of the largest US endowments says OpenAI and Anthropic are in big trouble

Reported by Businessinsider
Sep 25, 2026· 2 min read
Save Post LinkedIn Email
Key Takeaways
  • • Scott Wilson, CIO of Washington University's endowment, argues that OpenAI and Anthropic have taken on liabilities of trillion-dollar-plus scale that are unlikely to be justified as Chinese open-weight models close the performance gap at much lower cost.
  • • Wilson made the remarks onstage at Rock Yard Roundup, a Fort Worth gathering of roughly 100 asset managers, VCs and tech founders, calling his view an "unpopular opinion."
  • • Vinod Khosla, an early OpenAI investor, told Business Insider he has never been more bullish on OpenAI and that closed-model companies will have lower cost-to-serve advantages.
  • • OpenRouter data cited in the article shows DeepSeek at 25.3% of text-model requests versus 18.6% for OpenAI and 2.9% for Anthropic, a limited snapshot.
  • • Anthropic is reported to be expected to go public next month, raising the stakes of the debate over whether frontier labs can sustain their spending.
Summary
Washington University in St. Louis CIO Scott Wilson, whose early SpaceX investment generated a multi-billion dollar windfall for the endowment, argues that OpenAI and Anthropic are in serious trouble, according to Business Insider. Wilson contends that both companies have committed to enormous spending just as lower-cost Chinese models close the performance gap with U.S. frontier labs. "These trillion-dollar-plus frontier companies are not worth the liabilities that they signed up for," Wilson said. "There's going to be a ton of free alternatives."

Wilson made the comments at Rock Yard Roundup, a Fort Worth gathering of roughly 100 asset managers, venture capitalists and tech founders, where he described the frontier AI race as a frenzy of runaway spending and valuations. He cited Chinese developers including DeepSeek, Alibaba's Qwen, Zhipu AI and Tencent, which have narrowed the gap on performance while offering major cost savings. Wilson said conversations with colleagues in China and with founders of Washington University portfolio companies, who are switching to cheaper models, deepened his skepticism.

Wilson's view contrasts with that of Vinod Khosla, a billionaire venture investor and early OpenAI backer. Khosla told Business Insider he strongly disagrees with Wilson and has never been more bullish on OpenAI. He argued that the real advantage lies in controlling infrastructure, including co-designed chips such as OpenAI's Jalapeño inference chip, which could reduce reliance on Nvidia hardware and third-party cloud services and lower the cost to serve customers relative to open-weight models run on others' clouds.

Wilson also noted that the University of Michigan's early $20 million OpenAI investment is now reportedly worth over $2 billion, and that he passed on an early OpenAI investment because he was not convinced the company would make money. Data from OpenRouter, which Business Insider characterizes as a limited snapshot, shows DeepSeek accounting for 25.3% of text-model requests compared with 18.6% for OpenAI and 2.9% for Anthropic.

Go deeper than the headline

Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.

Start your free 7-day trial

Sources

1 outlet

AlphaMaven summarised this story from the reporting below. Each link opens the original article.

Businessinsider The head of one of the largest US endowments says OpenAI and Anthropic are in big trouble