Intel’s stock fell roughly 28% on Friday after a weak fourth-quarter forecast, and the drop alone wiped out about $12 billion of the value in the US government’s own stake in the company.
TL;DR: Intel had one of the worst trading days in its history after CEO Lip-Bu Tan warned that rising memory prices are already slowing PC sales. The federal government, which took a roughly 10% stake in Intel last year as part of a Trump administration deal, watched a chunk of that position evaporate in a single session. Both the government and Nvidia, which also holds Intel shares, are still sitting on gains for the year. This was simply the day the gains got a lot smaller.

A forecast that spooked the whole chip sector
Intel told investors to expect fourth-quarter revenue between $15 billion and $16 billion, well short of the roughly $17 billion Wall Street had modeled. Gross margin is expected to slide to 37% from 43%, and the company said it still doesn’t expect positive cash flow given how much it’s spending on new capacity. None of that is catastrophic on its own, but combined it was enough to send the stock into its steepest one-day decline in years, and it dragged other chipmakers down with it, including Nvidia, whose own shares slipped on the same news.
The real culprit is memory, not Intel’s chips
CEO Lip-Bu Tan was direct about where the pressure is coming from. On the earnings call he said customers are already signaling a PC slowdown tied to memory prices and availability, not demand for Intel’s processors specifically. The AI boom has been hoovering up memory chips for server racks and training clusters, which has pushed prices up across the board and made PCs, laptops, and other consumer devices more expensive to build. Tan added that system and endpoint prices have climbed meaningfully this year, squeezing IT budgets that were already tight.
“Our customers began signaling a PC slowdown based on the memory prices and availability,” Tan told analysts on the earnings call.
That’s an uncomfortable dynamic for Intel, since it means the AI industry’s appetite is making life harder for the part of the business that still pays most of the bills.
What the government and Nvidia actually hold
The federal government’s Intel stake dates back to a deal the Trump administration struck with the company roughly a year ago, converting around $11.1 billion in existing government funding into equity rather than letting it sit as grants. Nvidia separately bought a sizable chunk of Intel shares in 2025 as part of its own strategic positioning in the chip world. Both positions rode Intel’s stock up through the summer, and both are still net positive for the year even after Friday’s crash, since Intel remains up more than 70% year to date. But being up 70% instead of up 90% on a multi-billion dollar position is still a very real paper loss, and Friday’s session made that loss concrete in a way that’s hard to spin.
It’s an odd position for the US government to be in: a passive shareholder in a company whose fortunes now swing on headlines about memory chip shortages and AI infrastructure spending, two things no administration directly controls.
The one bright spot
Buried in an otherwise rough quarter, Intel’s data center business actually grew 78% year over year, which is the kind of number that would normally dominate the headline. The problem is that growth is capped by how much manufacturing capacity Intel can bring online. One new fab in Arizona is up and running, but the next one isn’t expected until 2028, and in the meantime Intel is still competing with TSMC for outside foundry customers who want their chips made somewhere other than Taiwan. Until that capacity catches up, Intel can’t fully cash in on the same AI data center boom that’s busy making its PC business more expensive to run.
A strange position for a government to be in
Equity stakes like this one are unusual for the US government to begin with, and Friday’s drop is a preview of what that arrangement actually looks like in practice. When a private shareholder loses $12 billion on paper, that’s a bad day for a fund or a pension. When it’s the federal government, the losses are attached to decisions made in the name of economic and national security, which puts a different kind of scrutiny on them. Supporters of the original stake will point out the position is still well in the black for the year. Critics will point out that taxpayers are now exposed to the same memory pricing swings and PC demand cycles as any Intel shareholder, with none of the upside control a private investor would normally expect in exchange for that risk.
Why this matters beyond one trading session
Intel has spent the past several years trying to rebuild itself around two bets: a turnaround in chip manufacturing and a bigger slice of the AI data center market. Friday’s selloff shows how exposed that story still is to forces outside Intel’s control, memory pricing chief among them. It’s a smaller-scale version of the same nervousness showing up across AI infrastructure spending this week, including Nvidia reportedly weighing a multi-billion-dollar backstop just to keep one data center project funded. It also puts a spotlight on what it means for a government to hold equity in a semiconductor company at all. When Intel’s stock drops 28% in a day, that’s not just a headline for traders, it’s a line item that shows up in how much value the public actually holds in a company the government decided was too strategically important to let fail on its own.
The bottom line
Intel’s bad day is really a memory problem wearing a PC-slowdown costume, and it’s one that traces straight back to how much the AI industry is paying for chips that have nothing to do with AI. The government’s stake and Nvidia’s position both took real hits, even if neither is underwater yet. Watch Intel’s fab ramp in Arizona and the 2028 timeline for the next one, because that’s the only lever Intel actually controls here, and it’s a slow one.



