Nvidia Falls 3.2% as Chip Selloff Deepens in 2026

Nvidia (NVDA) fell 3.2% to $221.30 and led a global chip selloff after hot CPI, as CEO Jensen Huang said it would cut off buyers developing unsafe AI.

Lead

What Happened to Nvidia and Chip Stocks?

Nvidia opened from a prior close of $228.52 and touched an intraday low of $219.64 before finishing at $221.30, a loss of $7.22. Volume reached 116.5 million shares, below the 151.6 million average, and the company's market value stood near $5.33 trillion. Shares were down almost 2% in premarket trading on Thursday.

The losses were broad. The Philadelphia Semiconductor Index has fallen about 10% over six sessions, its longest losing streak since May 2025, and the sector is on pace to lose more than $1 trillion in market value.

• Micron (MU) fell 4.1% and ASML (ASML) fell 3.9%.
• South Korean memory maker SK Hynix dropped 6.1%.
• AMD (AMD) lost 5.7% and Broadcom (AVGO) lost 3.1%.
• Oracle (ORCL) slid 8.8%.
• CoreWeave (CRWV) fell more than 16% on debt concerns tied to AI infrastructure spending.
• Arm Holdings (ARM), Applied Materials (AMAT) and Intel (INTC) have lost roughly 20%, 15% and 11% in the recent slide.

The S&P 500 fell 1.11% to 7,718.94 and the Nasdaq Composite fell 1.88% to 26,692.87, its largest one-day decline since July 23. Information technology was the weakest of the S&P 500's 11 sectors, down 2.6%. The Dow Jones Industrial Average finished nearly flat, off 0.09% at 51,621.48.

Why Did Chip Stocks Sell Off?

Chip stocks sold off because September inflation raised the odds that borrowing costs will rise further. Consumer prices increased 0.6% in the month and 3.7% from a year earlier, against forecasts of 3.6%. Core CPI, which excludes food and energy, rose 0.3% on the month and 2.6% annually, above estimates of 0.2% and 2.5%. Gasoline climbed 9.1% and accounted for half the monthly increase.

Market-implied odds of a rate hike at the Fed's Oct. 27-28 meeting rose to roughly 37% from 22% a day earlier. Higher yields reduce the present value of long-dated earnings, which weighs most on ai stocks priced for years of rapid growth. The selloff also reflects unease about the capital needed to build AI data centers, with the debt-funded model behind some of those projects drawing particular scrutiny.

What Did Jensen Huang Say About Unsafe AI?

Huang said Nvidia would cut off a customer building unsafe AI, but only after talking to them first. Asked in a Bloomberg interview in London whether he would stop supplying an irresponsible developer, he replied: "It's a hypothetical, but yes." He added that Nvidia does not run its customers' businesses and that there are limits to the judgment it can apply to how its technology is used.

Huang framed safety concerns as growing pains in a fast-expanding industry and said the technology can be controlled and deployed safely. He argued for faster rollout of monitoring and containment tools, including using AI to supervise other AI systems. He also called for applying existing product liability and cybersecurity law before writing AI-specific rules. That stance contrasts with calls from some AI lab leaders to slow development.

In London, Nvidia announced $250 million in computing resources for UK research on AI safety and security, spread over five years at roughly $50 million annually. The resources will be delivered through UK-based Nvidia Cloud Partners rather than as cash. Recipients include Oxford, University College London and Imperial College London, with work covering agent security, open-model security and robotics safety.

Huang also dismissed concerns about an AI bubble. He said demand for AI infrastructure will outstrip the supply of components for years, a view at odds with the market's reaction the same day.

Strategic Context

Nvidia supplies the accelerators behind most large-scale AI training and deployment, which gives its statements about customer conduct unusual weight. A supplier's willingness to withhold chips is a form of private governance in a market where export controls and safety rules remain unsettled. Huang's conditional wording leaves the threshold for action undefined and puts no new restriction on current customers.

The company's financial position is not the source of pressure. Its board has approved a $150 billion stock buyback, and the stock carries a consensus Buy rating with an average price target of $321.89. The pressure comes from the rate and funding backdrop for the whole sector.

What Comes Next for AI Stocks?

The next catalysts are TSMC's quarterly results on Thursday, expected to show profit growth of about 60%, and the Fed's Oct. 27-28 decision. Strong results from the world's largest contract chipmaker would test whether demand for AI hardware can offset rate concerns. Another hot inflation print or a rate increase would extend pressure on valuations, while softer data would ease the yield backdrop that triggered the decline.

Outlook

Nvidia's 3.2% drop on Oct. 14 was part of a rate-driven rout that has erased about $1 trillion from chip stocks, not a company-specific setback. Huang's comments on cutting off unsafe AI buyers set out a conditional supplier stance while leaving the business case for AI infrastructure intact in his view. Near-term direction depends on TSMC's results and the Fed meeting later this month.