Broadcom CEO Hock Tan says investors in artificial intelligence should remember a key lesson from Warren Buffett: know what a business does before betting on it. Tan told Bloomberg in Singapore that many investors have “no clue what they are investing in” and hold “extreme expectations,” only to find the experience “almost like they are in a casino.”
Tan’s comments put a practical question beside the excitement around AI: can companies show that the technology is useful enough to justify its cost? He is bullish on AI as Broadcom’s chief executive, but said long-term industry growth could face a shortage of capital. He also said profitability will depend on businesses proving AI’s utility, and that some may be “wasting money.”
The Key Lesson in Tan’s AI Warning
Buffett’s investing advice appears in his 1996 letter to Berkshire Hathaway investors. He wrote that investors need the ability to evaluate selected businesses within their “circle of competence.” The circle’s size, he said, matters less than knowing where its boundaries are.
Tan’s concern is that investors may be crossing those boundaries when they buy into AI without understanding what a company does or how it might earn from the technology. He described some expectations as extreme, with investors discovering only later that their decision resembled a casino bet.
That warning does not amount to a call to avoid AI altogether. Tan runs Broadcom and is bullish on the sector. His stated concern is whether companies can demonstrate practical utility and profitability over time, while the industry continues to require capital.
Buffett’s own experience also shows the limits of expertise. The source reports that he regretted not investing in Google earlier because the company was outside his circle of competence. He bought late, after recognizing the opportunity.
AI may make that judgment harder for investors, the source argues, because the technology can be difficult to understand. At the same time, large bets by Meta, Microsoft and Alphabet can encourage individual investors to assume those companies know more about the technology and its consequences.
Why the Key Lesson Is About Business Value
The issue is not simply whether AI can produce impressive results. Tan’s comments point to the cost behind those results and whether a business can make useful applications worth that expense. The source compares the appeal to a Ferrari: performance can be impressive while remaining expensive, and that cost may not make sense for every business case.
For investors, the distinction is between joining an AI trade because it is popular and evaluating what a company is building. Buffett’s advice, as quoted in his letter, focuses on understanding selected businesses and recognizing the limits of that understanding.
Tan’s warning leaves the investment decision open while emphasizing the test he says matters: companies need to prove AI’s usefulness to their businesses. That evidence, alongside a clear understanding of what an investor owns, is the concrete measure behind the lesson.
Readers following executive decisions in technology and finance can also see how leaders’ choices shape their industries. Finance leaders recognized in 2026 offers related coverage of financial services leadership, while Broadcom CEOs on AI growth and capital examines a separate report on financing constraints. The site has also covered technology leaders’ AI terminology debate.
Tan’s comments therefore leave investors with a focused question: can the companies they back explain how AI helps customers and supports a profitable business?