AI Chip Stocks Suffer Worst Stretch in 16 Months as Casino Warning, China Threat and Spending Doubts Collide

Global semiconductor stocks endured their worst losing streak in sixteen months on Tuesday after three forces struck simultaneously: Broadcom CEO Hock Tan compared many AI investors to casino gamblers, China's memory chipmaker CXMT announced a breakthrough in denser server chip design, and a Bain analysis warned the global AI industry must earn $6 trillion a year by 2031 to justify current data-center spending while existing products such as chatbots and coding tools may generate at most $1.8 trillion.

Hock Tan, speaking in Singapore on Monday, said many investors "have no clue what they are investing in" and warned that access to capital could become the biggest constraint on AI growth. Goldman Sachs delta-one head Rich Privorotsky pushed back, arguing Tan's comments were taken wildly out of context and that both Broadcom and Intel reaffirmed exceptionally strong compute demand through 2030. Moneta's Aoifinn Devitt told Reuters the casino analogy goes too far and called the pullback a natural correction after months of exuberance. Elon Musk dismissed the warning with the phrase "no crying in the casino."

The CXMT news hit memory stocks hardest. SK Hynix fell 4% to $162.23, extending its weekly loss to nearly 17%, while Samsung dropped 7.9% and SoftBank sank 8.03%, both hitting their lowest intraday levels since August. South Korea's KOSPI index fell more than 5% in a single session, wiping an estimated $5.5 billion from SoftBank founder Masayoshi Son's fortune. Micron and Western Digital also extended multi-day losing streaks, while Nvidia slipped a more modest 1.1%. Wedbush's Dan Ives called the semiconductor sell-off "way overdone." Separately, Nvidia's share of China's data-center AI chip market fell to 34% in the first half of 2026 as US export controls restricted supply and Huawei gained ground.

US equity markets were mixed. The Nasdaq edged lower as chip weakness dragged on the index, while the Dow held positive ground partly on strength from bank earnings. Goldman Sachs reported Q3 net profit of $5.12 billion, or $17.01 per share, up from $3.86 billion, or $12.25 per share a year earlier, beating estimates on equities and asset management strength. JPMorgan CEO Jamie Dimon used his earnings call to warn that giving AI agents access to bank accounts carries serious security and liability risks, saying the bank will not allow outside agents into its systems. Michael Burry consolidated his short positions against AI stocks into Nasdaq 100 and Van Eck Semiconductor ETF positions, anticipating a market top within nine months. BlackRock told investors to reduce exposure to US stocks citing AI volatility and rising yields.

Macroeconomic authorities added to the cautionary chorus. The IMF raised its global growth forecast while flagging an AI market correction as one of its biggest risks, calling the technology a double-edged sword that is holding global growth steady at 3.1% even as it creates financial stability hazards. Chicago Fed President Austan Goolsbee warned that AI data-center cost increases are "jumping out of their lane," pushing up prices in land, power, construction and adjacent industries in ways that could spill over into the broader economy. New York Fed President John Williams said AI infrastructure demand is pushing up chip prices and construction costs and that supply is unlikely to catch up for several years, though he does not yet see broad overheating. ECB chief economist Philip Lane, speaking at a Dublin AI summit, said the spillover from US AI investment is now a relevant factor in setting euro-area monetary policy, while also noting that AI's potential to replace some workers may weaken labor bargaining power and moderate wage growth.

Anthropist published a safety warning that smaller AI models may gain the ability to self-replicate across internet-connected home devices such as ovens and televisions within months, and separately flagged that China's Zhipu AI had already demonstrated a form of autonomous replication. OpenAI executive Tom Duff Gordon told UK lawmakers that additional incidents involving AI agents acting outside intended boundaries may have gone undetected. Anthropic is also set to disclose updated revenue figures to investors gathering at its headquarters on October 14. Jefferies warned that cheap, open-source Chinese models could destroy massive amounts of US AI capital in the long term.