‘I’d Be Petrified’: Steve Eisman Says Cheap Chinese AI Models Could Wreck OpenAI and Anthropic’s Valuations
Steve Eisman has now delivered two warnings in two weeks: cheap Chinese open-weight AI models signal a price war, and OpenAI and Anthropic together underpin roughly 70% of hyperscaler AI revenue, making any pricing collapse a system-wide problem for Big…
“Big Short” investor Steve Eisman has leveled two pointed warnings at the artificial intelligence boom in the span of just a few weeks. On his Real Eisman Playbook podcast on July 29, he said that cheap Chinese open-weight models signal “price war” for the two biggest AI labs. Two weeks later, on CNBC’s “Fast Money” on August 13, he broadened the concern: OpenAI and Anthropic account for roughly 70% of AI-related revenue at Microsoft, Amazon, Alphabet (NASDAQ:GOOGL | GOOGL Price Prediction), and Oracle, and as much as 25% to 35% of those companies’ cloud revenue. His conclusion: “The futures of these massive companies, in a sense, are a bet that OpenAI, Anthropic are going to succeed.” The valuation-collapse framing in our headline is our inference layered on Eisman’s words, since a $1 trillion-plus IPO story assumes pricing power that a price war would steadily erode.
The Moonshot Threat: Kimi K3 and Open Weights
Eisman’s specific concern on the July broadcast was Moonshot AI’s Kimi K3, which charges $3 per million input tokens compared with $5 for OpenAI’s GPT-5.6 Sol and $10 for Anthropic’s Claude Fable 5. Pricing is only half the story. Moonshot released Kimi K3’s full model weights, so developers can run and customize it independently rather than staying locked to Moonshot’s platform. That portability undercuts the “stickiness” that closed-model economics depend on: if an enterprise buyer can host a comparable model on its own GPUs at a fraction of frontier API pricing, the switching cost that justifies premium subscriptions shrinks every quarter. Not everyone is equally alarmed. MIT professor Ethan Mollick has argued Kimi K3 is a strong but uneven model rather than another DeepSeek-scale breakthrough. Eisman made the price-war case while debating tech bulls Dan Ives and D.A. Davidson’s Gil Luria on AI moats.
The IPO Stakes
Anthropic filed confidentially with the SEC on June 1, 2026, with OpenAI following shortly after (around early June); both filings remain confidential rather than public S-1s. Anthropic’s most recent private-market valuation was set at $965 billion in its May 2026 Series H, but investors are now targeting a $2 trillion-plus valuation at the public debut, which would surpass SpaceX’s $1.77 trillion June 2026 listing as the largest IPO in history. Supporting that ambition, Anthropic’s annualized revenue run rate reached $65 billion by the end of July 2026. OpenAI has reportedly wavered toward a 2027 listing amid market volatility, with CEO Sam Altman said to have a “hard floor” of a $1 trillion listing price. As of Eisman’s July 29 broadcast, Polymarket traders priced Anthropic’s odds of going public by year-end at roughly 69%, versus just 19% for OpenAI. Public investors will price the moat directly, which makes Eisman’s price-war framing pointed rather than academic.
China’s Price War Is Already Underway: Baidu
On Bloomberg’s The Asia Trade on August 3, 2026, Bloomberg Intelligence analyst Robert Lee argued the commoditization Eisman fears is already playing out in China. “There’s a high level of commoditization in the AI sector. The sector is overpopulated, flooded with supply. At last count there were 988 large language models officially approved by China,” Lee said. He drew a parallel to solar’s collapse: an oversupplied market where price-cutting is the only lever left. DeepSeek cut API pricing by as much as 50%, and Baidu (NASDAQ:BIDU) slashed API pricing by roughly 99% as part of an industry-wide race to the bottom that began when ByteDance made the first dramatic cuts. Baidu’s own numbers show the structural shift beneath that price war: AI Cloud Infra revenue rose 79% year over year while Online Marketing Services fell 22%. Lee named Alibaba (NYSE:BABA), Tencent, and Huawei as the best-capitalized survivors of the shakeout. Alibaba backs that up with its Cloud Intelligence Group accelerating 40% externally and the Qwen open-source family surpassing 1 billion cumulative Hugging Face downloads, per its Q4 FY26 6-K filing.
The Bull Rebuttal: Alphabet and Real Revenue
Luria pushed back on Eisman by noting that OpenAI and Anthropic have a combined revenue run rate above $75 billion, “likely topping $100 billion” once Google’s Gemini and others are counted, up from essentially zero two years ago. Ives, who recently launched a firm called Yorkville Ives, framed the scare as one of several “gut check moments” expected each year during what he calls an eight-to-ten-year AI buildout now in year three. Alphabet offers the clearest data point that scale can absorb price pressure: Gemini now processes 22 billion API tokens per minute with 950 million monthly active users on the Gemini App, while Cloud revenue grew 82% year over year to $24.77 billion in Q2 2026. Even so, Eisman himself noted that capital-intensive AI spending has transformed once capital-light giants, citing Alphabet’s decision to raise its 2026 capex guidance to as much as $205 billion as evidence that the arms race carries real financial risk regardless of who wins the model layer.
The Tension for Public Investors
A price war is a boon for AI buyers and a direct threat to the trillion-dollar valuation story both labs need to sell at IPO. Eisman’s argument has grown sharper since the July broadcast: the concentration risk he flagged on CNBC, where two private startups underpin 70% of hyperscaler AI revenue, means any pricing deterioration at the model layer ripples quickly into the income statements of Microsoft, Amazon, Alphabet, and Oracle. Robert Lee’s China data suggests the compression pattern already rhymes with commodity cycles investors have seen in solar and semiconductors. Anthropic’s October listing window and any pricing responses from OpenAI or Alphabet’s Gemini will serve as the first real-world test of who is right.
Editor’s note: This update adds Eisman’s August 13 CNBC appearance and his claim that OpenAI and Anthropic account for roughly 70% of AI-related revenue at the major hyperscalers. It also updates Anthropic’s IPO valuation target from the $965 billion Series H price to the $2 trillion-plus figure investors are now seeking, and incorporates Anthropic’s $65 billion annualized revenue run rate as of July 2026.
Contact [email protected] for any questions or corrections.
AJ has spent the past 10 years writing about financial markets at The Motley Fool. His coverage centers on technology stocks and the broader macroeconomic trends, from interest rates to geopolitics, that shape where markets are headed next. AJ is drawn to the stories where big-picture economics and individual companies collide.
