Nvidia is a victim of the computing market it created

Long the industry leader, Nvidia has had a bad couple of months. Bloomberg It has ugly detailsBut the result is that the company’s stock price is down 15% since its peak in May, even as projected revenue continues to grow. Compared to expected earnings, the company is now cheaper than the S&P average; Investors pay less per dollar of expected earnings for Nvidia than they would for a typical large U.S. company.

Money is still flowing into AI infrastructure stocks, but most of it is going to memory companies. During the same period, the value of Micron — one of the world’s largest manufacturers of dynamic random access memory (DRAM), the standard type of memory chip found in computers and servers — has nearly tripled, making memory the new bottleneck for data centers and the hot new AI business. The basic reason is simple: the GPU shortage that seemed alarming last year has eased a bit. Meanwhile, data centers need all the memory money can buy.

For anyone who appreciates Nvidia’s technological achievements, this may feel a bit underwhelming. There’s a lot of truly impressive technology behind Nvidia’s rise, both in developing CUDA, the widely adopted programming platform that made Nvidia GPUs the default engine for AI research, and in pushing the pace of GPU development to a speed few thought possible. Nvidia’s success is something you could write entire books about, and the GPUs themselves are among the most complex devices ever produced, on the brink of bleeding human capabilities.

For memory companies like Micron, the story is much simpler. They make high-bandwidth memory chips — specialized components designed to move data to and from processors as quickly as possible — which have been getting progressively better for 20 years. Without changing chips or companies much, the service they offer suddenly becomes very valuable – and since demand is growing faster than anyone’s ability to increase supply, they’ve been able to increase prices tenfold over the past year.

This, via Datatrack, is what the DRAM spot price — the price buyers pay for chips on the open market, rather than long-term contract prices — looks like since 2023:

You might think that there would have been amazing technical progress in the summer of 2025, but no, the industry as a whole has greatly underestimated the amount of memory you would need to build a data center.

In comparison, this (Heb Ornn Computing Market) is how the hourly spot price on the Nvidia H100 GPU has changed over the past year:

Just like Nvidia’s stock price, there is a peak in May (about $3.20 per hour) and then a steady decline. For better or worse, Nvidia’s value as a company is tied to the price of computing and that price is declining. Microns and their combination are tied to the price of DRAM, and that price continues to rise.

When I spoke with Ornn co-founder and CTO Wayne Nelms about the forces driving this disparity, he framed it as a simple issue of supply and demand. Google, Amazon, MicrosoftAnd even OpenAI They launched their own custom processors to reduce their dependence on Nvidia; Even if these chips aren’t as good as Nvidia’s latest model, they’re good enough to bring down the price of computing.

“More GPU drivers and accelerators are entering the market. Everyone wants to make their own silicon, but no one makes their own DRAM,” Nelms told me. “Until there is a major technological advance in HBM (high-bandwidth memory), a shift in supply and demand, or someone new (enters the memory market), I think things will continue more or less as we see today.”

It’s a frustrating situation for Nvidia, and largely a product of its success. Having proven how valuable computing is, the company finds itself at the center of a market where everyone wants to be – while simpler technologies and less interesting companies get rich on the margins.