The selloff in AI-linked chip stocks moved into regular U.S. trading Tuesday, with Micron, Nvidia, AMD and Broadcom falling after steep declines in Asia.
Adding to the debate was Hock Tan, CEO of semiconductor giant Broadcom. Speaking in Singapore Monday, he compared many AI investors to casino gamblers, according to Bloomberg. His remarks carry weight because Broadcom is earning billions from the technology those investors are betting on.
But his comments are only part of the story. Reported advances by a Chinese memory-chip competitor have also raised questions about the pricing power and future profits of established manufacturers.
The Selloff Continues After the Opening Bell
In the Associated Press’s early-session report, anchored to 9:35 a.m. Eastern, major U.S. chip stocks were lower:
| Company | Reported Decline |
|---|---|
| Micron Technology | 5.8% |
| Advanced Micro Devices | 3.3% |
| Nvidia | 1.3% |
| Broadcom | 1.3% |
At that point, the Nasdaq was down approximately 0.4%, while the Dow rose 0.6% and the S&P 500 slipped 0.1%. Strong earnings reports elsewhere helped support the broader market. These are early-session figures, not closing prices.
That divergence matters. The pressure was particularly pronounced in chip stocks, while other parts of Wall Street held up better. A portfolio’s exposure to the AI spending cycle could therefore matter more than the headline move in the overall market.
The U.S. declines followed a difficult session in Asia. Samsung Electronics fell 7.4%, SK Hynix dropped 10.1%, and Japan’s SoftBank Group declined 8.5%, according to Tuesday morning reporting.
What the “Casino” Comment Means
“A lot of investors have no clue what they are investing in,” Tan said, according to Bloomberg’s reporting. He argued that extreme expectations can leave investors behaving as though they are in a casino.
The context deserves care. MarketWatch reported that other accounts framed his remarks as criticism of short-term investor reactions, while emphasizing his confidence in continued AI demand. Goldman Sachs trader Rich Privorotsky also argued that the comments had been taken out of context. A full recording or transcript would help resolve those differences.
Tan’s warning should therefore be read as a challenge to investment assumptions. It does not establish that Broadcom expects AI spending to collapse.
The company’s financial results provide an important counterweight. In its September 2 earnings release, Broadcom reported $16.7 billion in quarterly AI semiconductor revenue, up 221% from a year earlier. It projected that figure would reach $21.7 billion in the following quarter. Broadcom also generated approximately $13.7 billion in quarterly free cash flow.
Those figures demonstrate substantial business activity. They leave shareholders with a separate question: how much future growth has already been built into the stock price?
China Creates a Different Kind of Risk
The selloff also followed reports that Chinese memory-chip maker ChangXin Memory Technologies, known as CXMT, had advanced a design capable of storing more data. Investors are concerned about the potential challenge to manufacturers including Samsung, SK Hynix and Micron.
A reported technical advance still has to become a commercially competitive product. Manufacturing at scale, controlling costs and winning customers are essential steps between a promising design and a meaningful change in market share.
If CXMT succeeds, the pressure could reach profits through pricing. More capable competitors can make it harder for existing suppliers to charge premium prices, even when overall demand is growing.
An expanding AI market can support higher sales while increased competition limits the profits companies earn on those sales. Technology adoption and investment returns can move in different directions.
Where This Leaves GMN Readers
For Individual Stockholders
The relevant question is what has changed in the company’s business. A falling share price deserves a different interpretation when customers are canceling orders than when investors are simply willing to pay less for the same expected earnings.
Watch revenue guidance, profit margins and cash generation together. Strong sales growth accompanied by weaker margins can indicate that competition or costs are absorbing more of the benefit. Growth supported by cash flow gives shareholders firmer evidence than distant projections alone.
For Fund Holders
Owning several funds can create more overlapping exposure than their names suggest. A portfolio containing technology, semiconductor and AI-themed funds may repeatedly hold the same companies or depend on the same spending cycle.
Review the largest holdings across the entire portfolio. The useful measure is how much of the portfolio relies on continued AI investment and favorable chip-industry economics.
A diversified fund spreads company-specific risk. It can still experience a substantial decline when an entire sector is being repriced.
This Week Brings a Test of Expectations
Two upcoming reports will give investors fresh evidence beyond conference remarks:
- ASML reports October 14. Its orders and outlook will help indicate how chip manufacturers view future equipment spending.
- TSMC reports October 15. Its revenue outlook, margins and customer commentary will help investors assess demand across the semiconductor industry.
Beyond those reports, watch whether memory-chip competition begins affecting prices and margins, and whether customers maintain their infrastructure spending plans. Those developments will carry more financial weight than a single provocative phrase.
Tan’s casino comparison is an attention-grabbing warning. The practical response is to examine what each company earns, how much cash it produces and what growth its valuation requires. AI can continue advancing rapidly while shareholders discover that they paid too much for the opportunity.
About Author
