SANTA CLARA – Intel’s second quarter looked, for the first time in recent memory, like a company catching the wave rather than watching it pass. Revenue hit $16.1 billion, up 25 percent from the same period a year ago, with the data center and AI segment up 59 percent to $6.3 billion. Every number landed above analyst estimates.
The catch is in the bottom line. Intel recorded a GAAP loss of $2.16 per share for the quarter. The headline profit that Wall Street actually tracks, non-GAAP earnings, came in at $0.42 per share, and even there the company is being generous about what it excludes. Restructuring charges, share-based compensation, and amortization costs are not small.
Chief Executive Lip-Bu Tan, who took the job less than two years ago and has spent most of that time cutting costs and reorganizing the company’s structure, offered a forward-looking characterization of the results: “AI demand is accelerating; Intel captures growth across CPU lines, ASICs, packaging, and foundry operations.” That breadth of exposure, spanning client chips to contract manufacturing, is what Intel is betting will differentiate it from pure-play AI chip vendors that live or die by a single product line.
The Client Computing and Physical AI segment, which covers the processors that go into laptops and desktops, grew 13 percent to $8.9 billion. That number reflects a healthy PC replacement cycle as much as any AI transformation, though Intel has been marketing the AI PC angle aggressively and some demand is tied to on-device inference capabilities in its latest Core Ultra processors. Intel’s complete Q2 2026 earnings release on SEC EDGAR provides the full segment breakdown.
Intel Foundry, the contract manufacturing business that Tan has positioned as Intel’s long-term structural bet, grew 31 percent to $5.8 billion. This is the division with the most contested future. Intel is competing for advanced packaging and chip fabrication contracts against TSMC and Samsung, neither of which has had a rough quarter. Chief Financial Officer Dave Zinsner attributed the stronger performance to “stronger yields and faster cycle times,” which is precisely the operational language that investors needed to hear after several quarters of manufacturing setbacks.
Gross margin on a GAAP basis improved 12.9 percentage points to 40.4 percent, a real improvement, though still well below the margins that chip companies with lighter manufacturing footprints generate routinely. Non-GAAP gross margin reached 41.8 percent.
Operating cash flow for the first half of 2026 came in at $7.0 billion. That matters because Intel is running a capital-intensive foundry expansion simultaneously with a shareholder return program. The cash generation gives Tan’s team real flexibility, though not unlimited flexibility.
For the third quarter, Intel is guiding revenue of $15.8 billion to $16.8 billion, with non-GAAP gross margin projected at 42.0 percent and non-GAAP earnings per share of $0.38. On a GAAP basis, the company sees earnings of $0.31 per share, a rare instance where the reported number looks better than the adjusted figure. The guidance reflects some client-segment seasonality alongside continued strength in data center workloads.
Alphabet’s Q2 results, reported last week, showed Google Cloud growing at 35 percent and capital expenditure commitments that Intel’s foundry team would very much like to capture as manufacturing contracts. Whether those relationships develop, and on what timeline, is one of the most consequential open questions in the chip industry right now.
Supermicro’s $60 billion backlog provides a downstream proxy for the AI infrastructure buildout that Intel is attempting to serve from multiple positions simultaneously. The question for Intel is whether its pace of manufacturing improvement is fast enough to compete for that business before hyperscalers commit to alternative supply chains.
IBM’s Q2 results, which showed mainframe hardware falling 42 percent as AI budgets redirected enterprise IT spending, offered a cautionary note about the uneven nature of AI adoption across the market. Intel’s DCAI growth suggests the hyperscaler layer remains robust even where enterprise purchasing has stalled.
What the Q2 report does not resolve is Intel’s path to sustained GAAP profitability. The company remains in a costly transformation, carrying legacy manufacturing commitments and restructuring expenses that will shadow the income statement for several more quarters. Tan’s bet is that the AI infrastructure wave is large enough and durable enough to fund Intel’s reinvention while it is still happening.
That bet may yet pay off. The GAAP number, however, is a reminder that the costs of transformation are real, and $16.1 billion in revenue, while a significant improvement, is not the same thing as a solved problem.
