Intel's 20% plunge: the market wants sales, not turnaround stories

What does it take to lose about a fifth of your market value in one evening? For Intel, it took an earnings beat. That is the uncomfortable lesson of Thursday's after-hours session, and it matters well beyond one chipmaker.

Intel reported third-quarter adjusted earnings of 52 cents per share against expectations of 40 cents, and revenue up 21% to $16.51 billion, essentially in line with forecasts. Its data center and AI unit grew 78% to $7.3 billion. Then came the guidance: fourth-quarter revenue of $15 billion to $16 billion versus the $17.1 billion analysts expected, and adjusted earnings of 25 cents versus 43 cents, according to TheStreet.

Reports of the drop vary with the time they were written, from about 15% to 23% in after-hours trading. We would not hang anything on the precise figure. The direction and the scale are clear enough.

The past quarter was never the point

A rise of roughly 150% this year, as SiliconANGLE and TheStreet describe it, means the market had already paid for the recovery. When a stock has priced in the future, the past quarter is almost irrelevant. What matters is whether the next one confirms the story, and this one did not.

Notice what the weak guidance was made of. The PC business is shrinking: client revenue fell 5% to $8.1 billion, and Bloomberg's report via Business Times says Intel expects the PC market to slide 10% in 2027 after a 12% fall in 2026. Memory costs are a main culprit, and ironically the AI boom that lifts Intel's server chips is also behind the memory squeeze. The same tide raises one division and drowns the other.

Then there were the supply problems. CEO Lip-Bu Tan said the data center unit would have done even better without raw-material shortages and a shutdown at a factory in Israel, which Intel estimates will cost about $600 million in revenue over two quarters. Demand in the data center, in other words, was not the issue. Delivery was. And that is precisely the distinction the market punishes, in our view: a turnaround story is a promise, and a missed shipment is a broken promise, however sympathetic the cause.

Cash is the quiet part of the story

The most telling detail may not be the revenue miss. Intel said capital spending will rise "up strongly" in 2027 from $20 billion this year, and that it will not generate positive cash flow that year. We think that combination is what turned disappointment into a rout. Investors can tolerate a soft quarter if the destination is near. They are far less patient when the road gets longer and more expensive at the same moment.

To be fair, the bull case has not vanished. Management called 2027 a "growth year," and the data center growth was the strongest in 15 years, per Proactive Investors. Tan also promised "a couple of customers ready to announce in the next two quarters" for the foundry business, which so far has one major client: Intel itself. That promise is now the next test. Notice, too, that it is again a promise.

Buyers of recovery stories should read guidance first

This is not advice to buy or sell anything. It is a way of reading these stories. First, when a stock has already doubled on a narrative, ask what the price assumes, not whether the company is improving. Improving and being worth the price are different claims. Second, watch guidance and cash flow more than the headline beat; they are where the narrative meets arithmetic. Third, be wary of stories that depend on outside forces, here memory prices and a factory outage, that management cannot switch off.

Our verdict: the market did not suddenly stop believing in Intel's long-term case. It stopped paying in advance for it. Intel may well deliver on data center and foundry. But anyone buying a recovery should assume the price will be tested on every forecast, and that being right eventually is cold comfort if you overpaid on the way.