Intel's 29% Crash Puts $12 Billion of Washington's Chip Stake in Play and Exposes the Flaw in CHIPS-for-Equity

Intel beat on every headline number and still lost roughly a third of its value in a day. Third-quarter revenue of $16.51 billion matched the $16.5 billion consensus and adjusted earnings of $0.52 beat the $0.40 estimate, with GAAP gross margin at 43.0% and non-GAAP gross margin at 44.1%. The damage came from the quarter that follows: revenue guided to $15 billion–$16 billion and adjusted earnings to $0.25, against a sell-side expectation of $17.1 billion and $0.43.

On Aug. 22, 2025, Intel agreed to sell the U.S. government 433.3 million primary shares at $20.47 — about $8.9 billion, or 9.9% of the company — funded by $5.7 billion of unpaid CHIPS Act grants plus $3.2 billion of Secure Enclave disbursements. Those shares closed Oct. 29 at $92.07, worth $39.9 billion; the Oct. 30 intraday low of $64.11 put them at $27.8 billion. The session wiped roughly $12 billion of paper value from Washington's position without leaving the government underwater against its $8.87 billion cost.

Intel did not simply sell shares to Washington. It escrowed them. On Aug. 27, 2025, Intel issued 159 million shares into escrow, to be released to the Department of Commerce at a fixed $20.00 a share as Secure Enclave money arrives, and it collected the full remaining $5.7 billion of CHIPS Act disbursements in exchange. Because the delivery price is fixed while the market price is not, that obligation is carried as a derivative liability marked through the income statement — and it has ballooned. Intel's own share price became a line item in its GAAP earnings.

The reason PC units collapsed in 2026 is not that people stopped wanting computers. Gartner estimates combined DRAM and SSD prices rose 130% by the end of 2026, lifting PC prices 17% and pushing memory to 23% of a notebook's bill of materials from 16% in 2025, and it expects the sub-$500 entry-level PC segment to disappear by 2028. Third-quarter PC shipments fell 22.4% year over year to 55.4 million units, with Gartner noting DRAM costs surged fivefold year over year in the second quarter as vendors passed component costs through to buyers. The margin in this cycle was captured upstream, and the PC industry is paying for it in units.

Look at the revenue lines and the PC industry looks like a boom; look at the unit lines and it is a bust. Dell Technologies grew revenue from $27.0 billion in its October 2025 quarter to $47.0 billion in its July 2026 quarter and widened gross margin from 17.8% to 20.9% between April and July 2026 — while Gartner counted its third-quarter PC shipments down 25.4%. HP Inc. revenue was roughly flat at $15.7 billion, but operating margin fell to 5.7% from 7.3% in its October 2025 quarter, and its PC shipments dropped 30.9%, the steepest decline among the three leaders. Lenovo held the No. 1 spot with 14.99 million units and 27.1% share, but shipped 22.4% fewer of them. These companies are selling fewer machines at higher prices, which flatters revenue and hollows out the unit economics underneath it.

The dispersion inside the table is the tell. Apple, the one top-six vendor with the least exposure to the commodity notebook market, cut its decline to 7.1% and gained 1.7 points of share; the two vendors with the heaviest mainstream-PC exposure lost a quarter to a third of their units in a single quarter. Consumer resistance to component-driven price inflation, not a collapse in compute demand, is what broke Intel's client guidance.

Intel's second problem is that the part of the story investors paid for is still almost entirely internal. Foundry segment revenue of $6.7 billion, up 58%, looks like traction until you read the footnote: external foundry revenue in the quarter was $411 million, and the segment lost $1.85 billion operating money. Meanwhile Intel sold $20 billion of stock at $95 a share in August 2026, lifting diluted weighted-average shares from 4.53 billion a year earlier to 5.31 billion. The foundry premium is running on $411 million of external revenue while the share count carries Washington's stake down to roughly 8.2%.

Near term, Intel has guided to $15 billion–$16 billion of revenue and 38.0% non-GAAP gross margin, will book Kiryat Gat charges for scrapped inventory and idle capacity, and has told investors it will burn cash this year. The escrow liability shrinks as the share price falls, so fourth-quarter GAAP results will likely carry a non-operating gain that flatters a quarter the business did not earn. Expect the sell-side to follow the price: Cantor already cut to $90, and the stock ended the session roughly 29% below even that.

Long term, the only thing that repairs the client business is a turn in the memory cycle, and both the memory makers and the forecasters see the shortage lasting. SK hynix expects the memory supply shortage to persist through 2030 and has locked long-term agreements with 14 major customers covering more than half of its 2027–2028 capacity. If DRAM pricing holds, Intel's unit problem persists no matter how aggressively it raises CPU prices; if it breaks, OEM revenue growth reverses into volume recovery and the margin slides back down the chain toward the CPU. The swing factor is a memory contract cycle Washington cannot control, which is precisely the risk a fixed-price equity stake does not hedge.

Strip out the noise and the Oct. 30 session repriced three things. It repriced the PC unit cycle, because Intel's guide said the recovery is at least six months away while memory prices are still rising. It repriced the political template, because a 10% federal stake in a company that can fall 30% in a day is not a subsidy but a liquid position with a political constituency. And it repriced the gap between the two halves of Intel's story: $7.3 billion of data-center revenue growing 78% against $8.1 billion of client revenue that shrank 5%. The market stopped paying for the policy and started paying for the client segment — and the client segment is priced by memory makers.