Intel's Q2 2026 report is one of the stranger reads of this earnings season: a genuinely excellent operating quarter, wrapped around an accounting mechanism that makes the headline loss number get worse the better Intel's stock performs. Revenue hit $16.1 billion, up 25% year-over-year — Intel's fastest growth since 2011. Non-GAAP EPS of $0.42 beat the $0.21 consensus by roughly 2x. And the company reported an $11.0 billion GAAP net loss.
Both of those things are true, and neither one is the whole story.
The Headline Numbers
| Metric | Q2 2026 | YoY / vs. Estimate |
|---|---|---|
| Revenue | $16.1B | +25% YoY, fastest growth since 2011 |
| Non-GAAP EPS | $0.42 | vs. $0.21 consensus — ~2x beat |
| GAAP net loss | $(11.0)B | Driven almost entirely by one item (below) |
| Data Center & AI revenue | $6.3B | +59% YoY |
| Intel Foundry revenue | $5.8B | +31% YoY, still a $2.1B operating loss |
| Operating cash flow | $7.0B | — |
| Cash + short-term investments | ~$30B | — |
Intel Products revenue overall was $15.1 billion, up 28%, and Data Center & AI alone generated $2.5 billion of operating profit. This is not a company treading water — by every operating metric here, it's a genuine, demand-driven turnaround quarter.
The Accounting Story Behind the $11 Billion Loss
Here's the part that makes this quarter genuinely unusual, not just "big beat, big one-time charge" like several others this earnings season. The loss is driven by a $12.529 billion mark-to-market charge on shares held in escrow for the U.S. Department of Commerce — tied to Intel's CHIPS Act Secure Enclave agreement, under which Intel agreed to issue up to 433.3 million shares to the U.S. government (158.7 million of them held in escrow), released as Intel performs under the agreement and receives cash disbursements (the government's original 2025 commitment totaled roughly $8.87 billion, split between accelerated CHIPS Act funding and Secure Enclave-related disbursements).
The mechanism is the interesting part: as Intel's stock price rises, the fair value of that escrowed-share obligation rises too, and Intel has to book the increase as a loss. That means a stronger stock price literally produces a larger reported GAAP loss, independent of how the actual business performed that quarter. Intel could theoretically report a fantastic operating quarter, watch its stock rally on the news, and see its GAAP loss get bigger specifically because of that rally — a genuinely counterintuitive dynamic that has nothing to do with server chip demand or foundry execution.
Our take: this is a real accounting distortion worth understanding, not a red flag to dismiss with a shrug, and not a reason to ignore the headline either. Strip the mark-to-market charge out and Intel's quarter reads as a clear operating improvement — real profit in Data Center & AI, real revenue growth across Products, real cash generation ($7.0B in operating cash flow). But investors modeling Intel's GAAP profitability going forward need to understand this charge will keep appearing, in size proportional to how well the stock does, until the escrowed shares are fully released — a structural quirk unique to Intel among its semiconductor peers.
The Real Business, Underneath the Charge
CEO Lip-Bu Tan's framing was direct: demand for Intel's products is outpacing supply, and the company is seeing its strongest revenue growth in more than 15 years. Data Center & AI revenue of $6.3 billion (+59% YoY) with $2.5 billion of operating profit is the clearest evidence that Intel's server CPU business — long overshadowed by AMD and Nvidia in investor attention — is genuinely reaccelerating.
Intel Foundry is the part still not working. Revenue grew 31% to $5.8 billion, but the segment still lost $2.1 billion. This is the crux of Intel's multi-year turnaround bet: whether the foundry business can eventually win outside customers at scale (beyond manufacturing Intel's own chips) and turn that growing revenue line into an actual profit center. Nothing in this release resolves that question either way — the revenue growth is real, the loss is still real too.
Why the Stock Still Fell
Intel shares initially popped roughly 13% after the earnings release — the market's first read on the beat-and-reaccelerate story. By the end of that week, they'd fallen back nearly 8%, as investors digested CFO Dave Zinsner's capex guidance: more than $20 billion of 2026 capital expenditure, with 2027 capex guided "significantly above" 2026 levels. That's Intel joining the same AI-infrastructure spending wave that hit Alphabet, Meta, and Tesla this earnings season — a real, current cost being paid against a foundry turnaround that hasn't yet proven it can generate durable outside-customer profit.
The timing mattered here too. Intel's stock had already climbed a striking 257% in the first half of 2026, then fallen 28% in July heading into the print — this was already an extraordinarily volatile, overextended name before a single earnings number landed. A quarter that beat on revenue and non-GAAP EPS by wide margins, carried a headline loss that needed real explaining, and came with an aggressive new capex bill was never going to get a calm, one-directional reaction from a stock swinging that hard already.
What to Watch
- Whether the mark-to-market charge on escrowed CHIPS Act shares keeps recurring at similar scale — it's structurally tied to Intel's own stock price, so continued strength paradoxically means continued large GAAP losses until the shares fully release.
- Intel Foundry's path to profitability. Revenue growth (+31%) is real; the $2.1 billion quarterly loss is the actual test of whether this turnaround bet pays off.
- Whether Data Center & AI's 59% growth is sustainable or reflects a one-time supply-constrained catch-up quarter, as Lip-Bu Tan's "demand outpacing supply" framing suggests.
- The 2027 capex step-up, guided "significantly above" 2026's $20B+ — the same demand-versus-spending question facing every semiconductor and hyperscaler name this earnings season.
The Bottom Line
Intel's Q2 2026 is a genuinely good operating quarter buried under an unusual and genuinely confusing headline number. The $11 billion GAAP loss is real, but it's overwhelmingly an artifact of how Intel's CHIPS Act government-equity deal is structured, not a signal the core business is deteriorating — if anything, the operating numbers underneath it (25% revenue growth, a 2x non-GAAP EPS beat, 59% Data Center & AI growth) argue the opposite. The stock's round trip — up 13%, then down nearly 8% — reflects the market weighing a real turnaround story against a real, escalating capex bill, on top of a stock that had already moved an unusual amount before the print. Foundry profitability, not the accounting charge, remains the actual multi-year question investors should be tracking.
Intel Corporation (NASDAQ: INTC) reported Q2 2026 revenue of $16.1B (+25% YoY) and non-GAAP EPS of $0.42 (vs. $0.21 consensus). GAAP net loss was $11.0 billion, driven by a $12.5 billion mark-to-market charge on U.S. government-held escrowed shares. Shares rose as much as 13% after the report before falling back roughly 8% on raised capex guidance.