Nvidia Has Lost a Trillion Dollars and Is Now Cheaper Than Before the AI Boom

The company that defined the AI trade has given back about a trillion dollars in market value in under two months, and on earnings multiples Nvidia is now the cheapest it has been since before the boom began. The semiconductor selloff has deepened around it, with Intel down 21 percent over seven sessions and Micron falling as much as 13 percent in a single day.

The trigger was Meta. Its plan to sell spare AI computing capacity through a new unit called Meta Compute wiped roughly 200 billion dollars off semiconductor and cloud stocks, because a world where idle GPUs get rented out cheaply is a world that needs fewer new ones. The buildout has produced a second-hand market. That is what a cycle turning looks like.

The deeper shift is what Wall Street is willing to reward. For two years, announcing enormous AI capital spending was enough to lift a stock, and now the same announcements get punished until a company can show the revenue that justifies them. The question moved from how much are you spending to what are you earning. That is a much harder question.

The valuation argument cuts both ways. Nvidia trading at pre-boom multiples while its revenue is still growing is either a rare opportunity or a warning that the market no longer believes the earnings are durable, and history is unkind here, because semiconductor stocks always look cheapest at the moment earnings are about to roll over. Cheap on peak earnings is a trap. Cheap on trough earnings is a gift.

The supply of paper is its own problem. A huge SpaceX listing has already landed and Anthropic and OpenAI are expected to go public, so investors are being asked to absorb more AI equity at exactly the moment their appetite has thinned. Every new listing competes for the same money. That mathematics does not favour the incumbents.

So the biggest winner of the AI era is being repriced from a story stock back into a chip company. A trillion dollars gone, multiples back at 2023 levels, and a market that now wants proof instead of promises. The AI revolution is not over. Its shareholders are just being asked to pay less for it.