Broadcom chief executive Hock Tan said on Monday, 12 October 2026, that many artificial intelligence investors do not understand what they are betting on, Bloomberg reported from an event in Singapore. “A lot of investors have no clue what they are investing in,” he said, adding that they “set very extreme expectations only to find out that it’s almost like they are in a casino.”
As we read it, the money point in his remarks was about the business model of the AI labs. Tan said OpenAI reaching annualized revenue of $50 billion in September points to a viable business but offers little guarantee on long-term profitability, and he warned that insufficient capital could become the next constraint on the industry’s growth, according to Bloomberg.
Bloomberg said Tan, one of the chip industry’s most prominent figures, homed in on the rush to invest in large language models on the so-called technology frontier, a technology Bloomberg said many improperly understood.
“For now it’s still losing money because they’re spending a lot on training,” Tan said of OpenAI. “So frankly, it doesn’t tell anything at this point except that there is a lot of potential revenuegrowth out of providing frontier generative AI models.”
He also questioned whether businesses are getting value from AI. Bloomberg reported that he said model developers need to find customers that can generate returns from deploying the technology, rather than relying on users who are now “wasting money” on tokens as they experiment.
“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,” he said. “That’s what happening in enterprises today.”
Share moves the Associated Press reported after Hock Tan’s remarks: the Kospi down 7.4% during Tuesday’s session, SK Hynix down 8.8% and Samsung Electronics down 7.7%. The AP tied the Korean chip declines to reports about China’s CXMT.
Business Pill · COST PER TASK
A one-minute explainer of cost per task: why the price of finishing a job, not the price per token, is the number that matters. It teaches the general idea only and says nothing about any company in this story.
The key insight: As we read it, the line most useful to businesses is the arcade one: Tan located the return problem inside enterprises’ own spending, $10 of tokens gone in 10 minutes with nothing learned, not only in investors’ valuations.
What Hock Tan said in Singapore on 12 October 2026, as Bloomberg reported it. The labels paraphrase his remarks; quotations in the text are exact.
Capital as the Next Constraint
Tan warned that insufficient capital could become the next constraint on growth, Bloomberg reported. It set the remark beside Firmus Grid, the Nvidia-backed Australian data centre company that withdrew its planned $5.5 billion initial public offering last week, which Bloomberg called a sign that investors are growing more skeptical of AI infrastructure valuations based heavily on future revenue.
Before Tan took the stage, DBS Group Holdings chief executive Tan Su Shan said the market is repricing the risks on AI financing following heavy investment, according to the same Bloomberg report.
Intel chief executive Lip-Bu Tan, who joined the panel, said the rapid AI buildout means hardware will remain in short supply for the next two to three years, and that memory could be among the first segments to lose some steam, Bloomberg reported.
A pushback on the casino label. Not every investor accepts Tan’s framing. In a Reuters Market Talk video on 13 October, Aoifinn Devitt of Moneta told Reuters the casino analogy goes too far. In the AI trade, she said, the house doesn’t always win, and she sees the pullback as a natural correction after months of exuberance. That is her view, as Reuters summarised it.
The Chip Selloff on Tuesday
The Associated Press reported on 13 October that South Korea’s Kospi was down 7.4% at 5,914.35, its lowest level since late July, with Samsung Electronics down 7.7% and SK Hynix down 8.8%. The Korea Economic Daily said the fall below 6,000 triggered a sell-side trading curb for the first time in nearly two months.
The AP tied the Samsung and SK Hynix declines to reports that advances by China’s memory chipmaker CXMT could increasingly challenge Samsung, SK Hynix and Micron. A Bloomberg headline the same morning named both “Chinese Rival Worry” and “Broadcom Remarks”; we read that headline only, not the article body.
The AP also reported that Tokyo Electron fell 6.6%, SoftBank Group 8% and TSMC 1.6% in Asian trading, after Broadcom fell 1.9% and Micron 3.9% on Wall Street on Monday.
The Structural Read
Tan’s OpenAI figures cut both ways: Bloomberg reported that he called its $50 billion annualized revenue in September a sign of a viable businessmodel, while saying it is still losing money because it is spending a lot on training.
The capital warning came from a supplier. Broadcom designs some of the components integral to the global data center buildout, Bloomberg noted, and Tan said insufficient capital could become the next constraint on the industry’s growth.
As we read it, the panel put two chip chief executives side by side: Intel’s Lip-Bu Tan said hardware will stay in short supply for two to three years, while Broadcom’s Hock Tan said insufficient capital could become the next constraint.
Hock Tan, Broadcom CEO, as reported by Bloomberg, 12 October 2026
“A lot of investors have no clue what they are investing in”
Three Implications
AI INVESTORS Tan said many set very extreme expectations and that it is almost like they are in a casino, Bloomberg reported.
ENTERPRISES Tan said AI developers need customers that can generate returns, rather than users now wasting money on tokens as they experiment, according to Bloomberg.
AI LABS Tan said OpenAI’s $50 billion annualized revenue points to a viable model but that it is still losing money because of training spending, Bloomberg reported.
The framework puts it this way: “The new job of finance is to manage the ratio between the two — work produced per token burned. That ratio is the gross-margin equation of the agentic era.”
As we read it, Tan’s $10 of arcade tokens is that ratio at its worst: tokens burned, no work produced.
What Is Not Established
We found no video or transcript of Tan’s remarks; his words appear here as Bloomberg reported them, and Broadcom has made no separate statement that we found. The Kospi figures are as the AP reported them during Tuesday’s session; we could not read the Korea Exchange’s own data, which requires a member login, so the official close is not confirmed here.
How much of Tuesday’s selloff came from Tan’s remarks rather than the CXMT reports is not established by the reporting we read.
Business Engineer Framework
The CFO’s Guide to the Token Economy
A Business Engineer guide to the money side of AI: what tokens cost, what work they produce, and how finance teams should manage the ratio.
Hock Tan told an audience in Singapore that many AI investors are setting extreme expectations, that OpenAI’s $50 billion annualized revenue still comes with losses from training, and that insufficient capital could become the industry’s next constraint, according to Bloomberg. The next day the AP reported the Kospi down 7.4%, with Samsung and SK Hynix leading the fall.
A note on sourcing. This piece is based on Bloomberg’s report of 12 October 2026 on Hock Tan’s remarks at an event in Singapore, which we read in full through its Yahoo Finance syndication on 13 October 2026; we found no video or transcript, so his words appear as Bloomberg reported them. Market figures are from The Associated Press report of 13 October 2026, read in full, and the Korea Economic Daily’s headline and standfirst of the same day; the Korea Exchange’s own data requires a member login and was not read.
A Bloomberg headline naming Broadcom’s remarks among the causes was read as a headline only. Nothing here is a forecast, and nothing here is financial or investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
