Broadcom CEO Says Many AI Investors Are "Like Gamblers" With No Clue What They're Backing

Broadcom CEO Hock Tan has warned that many investors pouring money into artificial intelligence are behaving like gamblers rather than informed backers, a striking comment from the head of a company whose own growth forecasts depend heavily on the AI buildout continuing.

What Tan Said

Speaking at the inaugural DBS Asia Future Forward Conference in Singapore on October 12, Tan said plainly: "A lot of investors have no clue what they are investing in." He described much of the current enthusiasm around frontier AI models as resembling casino behavior, extreme expectations set without a real understanding of what's being backed.

Tan extended the criticism beyond investors to businesses themselves, questioning whether companies are spending productively on AI. He compared enterprise experimentation with AI tokens to a child given arcade money: "When you are a kid you go to a video arcade, you play. Your parents give you $10 of tokens, you finish it in 10 minutes and you haven't learned anything. That's what's happening in enterprises today."

On OpenAI Specifically

Tan pointed to OpenAI's annualized revenue reaching roughly $50 billion as of September, the same corrected figure we covered recently after it rattled AI stocks, as evidence that a viable business model may exist. But he was direct about the catch: "For now it's still losing money because they're spending a lot on training."

He added that the revenue figure, on its own, "doesn't tell anything at this point except that there is a lot of potential revenue growth out of providing frontier generative AI models," stopping short of calling it proof of sustainable profitability.

The Part That Makes This Notable

Tan's comments carry unusual weight because of exactly who's making them. Broadcom's own AI revenue has roughly tripled in two years, and the company has guided investors to expect that business to double to $115 billion in fiscal 2027, then double again to $230 billion in fiscal 2028. Those forecasts rest substantially on custom AI chip deals with Google, OpenAI, Anthropic, and Meta.

More directly, Broadcom is reportedly arranging more than $145 billion in chip-collateralized financing for OpenAI and Anthropic, the same labs Tan described as money-losing and economically unproven. According to reporting on Anthropic's IPO prospectus, Broadcom has agreed to provide up to $42 billion in financing to Anthropic alone, tied to its five-year, $125 billion compute lease commitments.

In other words, the CEO most financially dependent on continued AI infrastructure spending is the one publicly cautioning that a lot of the money behind it doesn't fully understand what it's buying.

A Warning About Capital, Not Just Sentiment

Tan's most consequential point for the industry may not be the gambling comparison, but his warning that insufficient capital could become the next real constraint on AI's growth, a shift from the chip and power shortages that have dominated discussion of AI bottlenecks until now. DBS Group CEO Tan Su Shan, speaking just before him at the same event, made a related point: markets are now repricing the financing risk tied to AI after two years of heavy investment.

That repricing is already visible elsewhere. A company backed by Nvidia pulled its planned $5.5 billion IPO just last week, reportedly amid growing investor skepticism toward AI infrastructure valuations.

Why This Matters

This is one of the clearest statements yet from inside the AI supply chain itself, rather than from outside skeptics, acknowledging that AI investment has outpaced investors' understanding of what they're funding, even as the underlying business questions—can frontier labs become profitable, and how long will that take?—remain unresolved. It also puts a spotlight on just how tightly interconnected the chip, financing, and AI lab sides of this boom have become: the same company selling the chips is also arranging the loans to buy them, from customers its own CEO says the market doesn't fully understand.

Frequently Asked Questions

What exactly did the Broadcom CEO say? Hock Tan said many AI investors behave like gamblers, setting extreme expectations without understanding large language models, and that "a lot of investors have no clue what they are investing in."

Is OpenAI profitable? No. OpenAI's annualized revenue reached about $50 billion as of September 2026, but the company continues to lose money due to high training and compute costs, according to Tan.

Why is Broadcom's comment notable? Broadcom's own growth forecasts and stock price are heavily tied to AI chip demand, and the company is reportedly financing over $145 billion in chip purchases for OpenAI and Anthropic, the same labs Tan described as economically unproven.

What did Tan say about the next AI bottleneck? He warned that insufficient capital, rather than chip or power supply, could become the next major constraint on the AI industry's growth.

Bottom Line

When the head of the company most financially entangled with AI's biggest spenders says much of the money behind the boom doesn't understand what it's backing, it's a meaningful signal about where investor sentiment and underlying business fundamentals may be starting to diverge.

Sources: Bloomberg, Yahoo Finance, Startup Fortune, BigGo Finance, TFTC. Information is accurate as of October 12-13, 2026 and may be updated as the story develops.