Why Intel's 28.6% Stock Plunge Is Suddenly Everywhere

Why it is suddenly everywhere

1. 01

   Intel’s official Q3 report showed revenue of $16.513 billion, up 21% year over year, and adjusted EPS of $0.52. LSEG’s consensus was $0.40, so the reported quarter itself beat expectations.

2. 02

   The same release guided Q4 revenue to $15–16 billion and adjusted EPS to $0.25. Analysts had expected roughly $17.1 billion and $0.43; Intel also forecast a non-GAAP gross margin of 38%.

3. 03

   Counterpoint found that global PC shipments fell 4% to 65 million in Q2 2026, ending five quarters of growth as higher memory costs worked their way into PC bills of materials.

4. 04

   Intel closed at $65.76 on October 30, down 28.58%. Norgate price data cited by Yahoo Finance show only July 5, 1974 was worse, making this Intel’s worst trading day since 1974 rather than conclusively its worst ever.

The apparent contradiction is the hook. Intel did not report a collapsing quarter: its data-center business grew rapidly, revenue rose, and adjusted profit exceeded the consensus. Yet nearly $140 billion of market value disappeared in one session. The answer sits in the gap between what the company had just delivered and what investors expected it to deliver next.

The timeline

1. 2021-03IDM 2.0 sets the course

   Intel says it will keep designing and manufacturing its own chips, build a major foundry business for outside customers, and use external foundries for some products. The strategy ties product recovery to expensive manufacturing expansion.

2. 2024-08-02An earlier confidence shock

   Intel shares fall roughly 26% after another earnings reset. The episode becomes the closest modern comparison until the October 2026 selloff goes further.

3. 2025-08-22The US government takes a stake

   An $8.9 billion agreement gives the government 433.3 million shares at $20.47 each, a 9.9% passive stake. Manufacturing capacity has by then become a matter of industrial policy as well as corporate strategy.

4. 2026-06-22Recovery expectations reach a record

   The shares hit a record after nearly quadrupling from the start of 2026. AI server demand, cost reductions and the foundry roadmap have all become part of the recovery story embedded in the price.

5. 2026 Q2The PC market turns down

   Counterpoint records a 4% shipment decline to 65 million units. Five quarters of growth end as the memory shortage raises costs for PC makers and weakens an already fragile replacement cycle.

6. 2026-10-29A beat meets a weak forecast

   Intel reports strong Q3 results but guides Q4 revenue and adjusted EPS well below consensus. Its client-computing group is expected to fall more than 20% sequentially, while gross margin drops to a forecast 38%.

7. 2026-10-30The reset reaches the share price

   The stock loses 28.58% in one day, second only to July 1974 in the Norgate dataset. Even after that fall, the shares remain up nearly 80% for 2026, showing how large the preceding rally had been.

One report, two very different stories

The quarter was strong; the next one looked much weaker

Financial reports contain two clocks. The income statement looks backward, while guidance sets the reference point for the next quarter. Intel’s backward-looking numbers were solid: Q3 revenue reached $16.513 billion, GAAP EPS was $0.60, and adjusted EPS was $0.52. Operating cash flow was approximately $5.1 billion.

The forward clock told another story. The midpoint of Intel’s Q4 revenue guidance was $15.5 billion, well below the roughly $17.1 billion consensus. Adjusted EPS guidance of $0.25 compared with $0.43 expected, and the non-GAAP gross-margin forecast fell from 44.1% in Q3 to 38%. For a stock priced around a rapid recovery, that difference mattered more than the Q3 beat.

The segment numbers explain why both things could be true. Data Center and AI revenue reached $7.318 billion, up 78% year over year. Client Computing Group revenue was $8.097 billion, down 5% from a year earlier and 9% sequentially. Intel then said client revenue would drop more than 20% sequentially in Q4. AI servers were booming inside the same company whose PC business was contracting.

AI helps Intel’s servers and hurts its PCs

That split is not as strange as it sounds. AI infrastructure needs huge amounts of server compute, which lifts demand for Intel’s data-center CPUs. It also needs enormous amounts of high-bandwidth memory. Memory makers allocate more wafer capacity to that profitable market, leaving conventional DRAM and NAND scarcer and more expensive.

The mechanism is the consumer side of the AI memory supercycle. A PC assembler cannot treat memory as an abstract commodities chart: it is a physical part of every machine. More expensive memory raises the bill of materials. If the assembler raises the retail price, demand can weaken; if it absorbs the cost, margins suffer. Either way, manufacturers become cautious about building inventory.

Counterpoint’s Q2 shipment decline supplied early evidence. Intel’s prepared remarks then put numbers on the downturn: management expected the 2026 PC unit market to fall about 12%, second-half units to drop more than 20%, and the 2027 market to decline roughly 10%. This is why an AI boom could strengthen one Intel segment while helping to depress another.

Foundry economics magnify the disappointment

Intel is also carrying a manufacturing strategy that a fabless chip designer does not. Under IDM 2.0, it spends heavily on fabs and process technology while trying to attract outside foundry customers. Those factories have large fixed costs. SemiAnalysis has repeatedly emphasized the importance of external wafer volume: more customers spread those costs across more production and help keep a process node economically useful after Intel’s own products move on.

The Q3 figures show how much work remains. Intel Foundry reported $6.697 billion of revenue, but most of that came from manufacturing for Intel’s own product groups. External foundry revenue was only $411 million, primarily from a non-recurring customer program, while intersegment eliminations were $6.291 billion. The headline foundry figure therefore cannot be read as $6.7 billion of outside customer demand. The segment still lost $1.853 billion at operating level, although that was an improvement from $2.321 billion a year earlier.

There was genuine manufacturing progress. Management said Intel 18A costs had declined about 40% since Q1 and Intel 3 costs about 15%. But gross capital expenditure remained around $20 billion for 2026, with heavier investment expected in 2027 as Intel 3, 18A and 14A overlap. Management said positive free cash flow in 2027 was unlikely. A production interruption at Fab 28 added a separate near-term drag: roughly $300 million of Q4 revenue and 200 basis points of gross margin, with a similar revenue effect possible in Q1.

None of this cancels the data-center growth or the cost improvements. It does explain why weak PC guidance landed so hard. The market was asked to reconcile a capital-intensive foundry build-out, a smaller PC market and lower near-term margins after the share price had already nearly quadrupled.

Key facts

October 30 close

$65.76, down 28.58%

Q3 revenue

$16.513B, up 21%

Q4 revenue guide

$15–16B vs ~$17.1B expected

Data Center and AI

$7.318B, up 78%

External foundry revenue

$411M in Q3

2026 PC unit outlook

Approximately 12% decline

What the sources are saying

Common questions

Why did Intel shares fall after an earnings beat?

The beat described Q3, but the guidance reset Q4. Revenue guidance of $15–16 billion was far below the roughly $17.1 billion consensus, adjusted EPS guidance was $0.25 versus $0.43 expected, and the client-computing group was forecast to fall more than 20% sequentially. The earlier share-price rally had left little room for that kind of miss.

Is Intel’s AI and data-center business shrinking?

No. Data Center and AI revenue was $7.318 billion in Q3, up 78% year over year. The contraction was in client computing, where revenue fell 5% year over year and was expected to drop sharply again in Q4. The same AI build-out can lift server demand while raising memory costs for PC makers.

Does Intel Foundry’s $6.7 billion revenue mean outside customers are buying that much?

No. Most foundry revenue is internal work for Intel’s product groups. External foundry revenue was $411 million in Q3, primarily from a non-recurring customer program, and intersegment eliminations were $6.291 billion. Internal and external revenue answer different questions about the business.

Was October 30 Intel’s worst trading day ever?

The defensible wording is worst since 1974. In Norgate data beginning in May 1973, only July 5, 1974 was worse, at nearly 30%. October 30, 2026 ranks second in that dataset, ahead of the roughly 26% fall on August 2, 2024.