Nvidia drops 16% as DeepSeek triggers a broad AI stock selloff

The market's loudest message came from one number: Nvidia fell 16%. That drop helped drag U.S. stocks mostly lower on Monday, with the Nasdaq leading declines as makers of AI infrastructure got hit hard, many in double digits. The trigger was China's DeepSeek, which sparked a broad rout across the group and forced investors to confront a simple but uncomfortable question: what if the AI buildout everyone has been paying for looks less inevitable, or less profitable, than it did last week?

That is why this move matters beyond a bad day for chip stocks. When the biggest and most closely watched name in AI loses that much value in a single session, it is not just a reaction to one company or one product. It is the market repricing a whole narrative. AI infrastructure has been one of the clearest winners in the market over the past stretch, with money flowing into the companies that sell the chips, servers, networking gear, and other hardware needed to train and run large models. Monday's slump showed how quickly that trade can go from consensus to crowded exit.

DeepSeek's role in the selloff is what gave the move extra bite. The source does not spell out the company's product details, but it does make clear that the market treated the news as a direct shock to AI infrastructure demand expectations. In plain English, if investors start believing more capable or cheaper AI systems can be built with less of the expensive hardware that has powered the current boom, the spending math changes fast. That is bad news not just for Nvidia, but for the broader ecosystem that has traded on the assumption that hyperscale AI spending will keep rippling outward.

For executives, the first takeaway is not that AI is over. It is that the market is no longer willing to price the theme as if demand will rise in a straight line forever. The AI stack has become a live test of capital intensity: how much money has to be spent on chips, data centers, and supporting infrastructure to produce the next jump in capability, and how durable those economics are if a competitor changes the cost curve. That is the kind of question boards and finance teams cannot leave to the hype cycle. Once investors start seeing an alternative path, even a partial one, the multiple compression can be brutal.

The sharpness of the move also matters because it tells you where pressure tends to show up first when sentiment turns. Infrastructure names are usually the front line of any AI boom, because they monetize the buildout directly. That makes them sensitive to every shift in spending expectations, every sign of competition, and every change in how investors think the value chain will be distributed. If a new entrant or a lower-cost approach challenges the premise that the largest gains will accrue to the hardware layer, then the market tends to punish those names before it revalues the rest of the ecosystem.

There is also a broader capital-markets lesson here. The Nasdaq led the declines, which is exactly what you would expect when a market is crowded into one high-growth narrative and then that narrative gets questioned. In those moments, investors often do not wait for perfect information. They sell first and sort out the implications later. That can create sharp, fast moves in the most expensive and most visible names, especially when a stock like Nvidia is already carrying huge expectations and a heavy role in benchmark indexes and portfolios. When one leader slips, the whole trade can feel less like an allocation and more like a trapdoor.

For peers in AI, the question now is whether their own capital plans, customer assumptions, and valuation stories still make sense if the market begins demanding proof rather than promise. That applies to chipmakers, server suppliers, networking firms, and any company whose pitch depends on AI adoption driving ever larger waves of infrastructure spend. It also applies to CEOs and boards outside tech who have been using AI as a shorthand for growth. Monday's rout was a reminder that the market loves a boom until it has to finance the bill. Once that happens, the scrutiny gets sharper, the margin for error gets thinner, and the winners are the companies that can show why their economics still work even when the crowd stops cheering.