Intel's Q2 looks absurd only until you separate the business from the accounting. The company had its strongest revenue growth in more than 15 years, but its CHIPS Act equity deal turned a rising share price into a huge GAAP loss.
The numbers still look wrong at first glance. Intel reported $16.1 billion in Q2 2026 revenue on July 23, up 25% from a year earlier, and its Data Center and AI business rose 59% to $6.3 billion. Intel's own release put non-GAAP earnings at $0.42 a share, against a Zacks consensus estimate of $0.21. That is a real beat. Then you hit the bottom line: an $11.0 billion GAAP net loss attributable to Intel.
That's the trap. If you read only the loss, you miss the quarter. If you read only the adjusted profit, you miss the cost of the deal Intel struck with Washington.
The charge sits in a very specific place. Intel's reconciliation shows a $12.529 billion Q2 loss from the mark-to-market of escrowed shares tied to its Warrant and Common Stock Agreement with the U.S. government. Its six-month cash flow table shows $13.619 billion in mark-to-market losses on those escrowed shares. The shares are connected to Intel's CHIPS Act Secure Enclave agreement with the Department of Commerce, and Intel says they are released as the company performs and receives cash proceeds under that agreement.
The original deal is not vague. In an August 2025 SEC filing, Intel said the U.S. government agreed to make $8.8698 billion of disbursements, including $5.695 billion in accelerated CHIPS Act funding and $3.1748 billion tied to Secure Enclave. Intel agreed to issue up to 433.323 million shares to the Department of Commerce, including 158.740 million shares held in escrow. Those escrowed shares are the accounting problem now sitting in the middle of Intel's income statement.
When Intel's stock rises, the fair value of that obligation rises too. The company can have a better operating quarter and still report a worse GAAP loss. That's not cosmetic. It changes how the quarter looks to anyone skimming the headline number.
The business finally showed up #
Look at the operating business. Intel Products revenue was $15.1 billion, up 28%. Data Center and AI produced $2.5 billion of operating profit. Intel Foundry revenue rose 31% to $5.8 billion, although the foundry unit still lost $2.1 billion. You don't have to pretend Intel is fixed to admit this is a different report from the dead-money quarters investors got used to seeing.
Lip-Bu Tan's prepared remarks made the point plainly: demand for Intel's products is still outpacing supply, and the company says it is seeing the strongest revenue growth in more than 15 years. Dave Zinsner, Intel's CFO, said Q2 operating cash flow was $7.0 billion and that Intel ended the quarter with about $30 billion in cash and short-term investments. Those are not vibes. They are the figures you should look at before deciding the loss tells the whole story.
There is still plenty to dislike. Intel Foundry remains deeply unprofitable, and the company is spending hard into a cycle where everyone is already nervous about AI infrastructure bills. Zinsner said Intel now expects more than $20 billion of 2026 capital expenditure, with 2027 capex significantly above 2026 levels. That is a big call. It only works if customer demand stays real and Intel's factory execution keeps improving.
The stock had already run too far #
The timing was awkward. Yahoo Finance reported that Intel shares popped nearly 13% after the earnings release before falling almost 8% on Friday as investors digested the higher capex plan. Benzinga, citing Bespoke Investment Group, had already flagged the larger rotation: Intel had climbed 257% in the first half of 2026, then dropped 28% in July before earnings. The stock wasn't reacting from a quiet place. It was coming off a sprint.
So you get the oddity. Intel delivered revenue growth investors had been waiting years to see, and the market still had reason to flinch. A $12.529 billion mark-to-market charge muddies the headline loss. A more than $20 billion capex plan raises the pressure on future cash flow. A foundry unit losing $2.1 billion still has to prove it can win outside customers at scale.
Frankly, this is why the quarter is useful. It doesn't give you a clean turnaround story or a clean disaster story. It gives you Intel as it actually is: a company with demand finally moving in its favor, a government equity deal distorting GAAP earnings, and a manufacturing plan expensive enough that investors have every right to keep asking when the spending turns into durable profit.
The next test is not whether Intel can explain away the accounting charge. It has done that. The test is whether the company can turn tight server CPU supply, 18A progress, and higher foundry volume into cash flow strong enough to justify the next round of spending.
Also read: Trump's tariff threat against foreign chipmakers is making the AI memory shortage worse • John Ternus takes the Apple CEO role in September with a hardware empire and a broken AI story to fix • The AI industry is repricing itself around intelligence per dollar and Amazon is showing how