Intel Q2 2026 Earnings: AI Beat Soars, But Foundry Risk Looms
Intel shares surged as much as 13% in after-hours trading to roughly $111 on Thursday after the chipmaker posted second-quarter revenue of $16.1 billion — well ahead of the $14.4 billion analysts expected — and non-GAAP earnings per share of $0.42, roughly double consensus estimates near $0.21. CEO Lip-Bu Tan called it Intel's strongest revenue growth in more than fifteen years, and the company followed with third-quarter guidance of $15.8 billion to $16.8 billion in revenue, also above expectations.
The headline GAAP figure told a different story: a $11 billion net loss, or $(2.16) per share, driven almost entirely by a non-cash mark-to-market charge tied to Intel's CHIPS Act escrowed shares held by the U.S. government. Operationally, the quarter was unambiguously strong.
The Engine Behind the Beat
Data Center and AI revenue jumped 59% year-over-year to $6.3 billion, with segment operating income nearly quadrupling to $2.47 billion from $633 million a year earlier. Management attributed the outperformance to higher factory yields, faster cycle times, and what Tan described as "unprecedented demand for compute." In response, Intel raised full-year 2026 capital spending above $20 billion and signaled 2027 capex will run "significantly" higher still, concentrated in U.S. fabs.
A Tale of Two Businesses
The results remain structurally bifurcated. Intel's product businesses — CCPG and DCAI combined — generated $4.8 billion in operating profit. Intel Foundry, by contrast, lost $2.1 billion, though that loss narrowed meaningfully from $3.17 billion a year ago. Foundry revenue rose 31% to $5.8 billion, but roughly $5.5 billion in intersegment eliminations indicates that growth remains largely a function of internal Intel product volume rather than external customer wins — Fortinet was the quarter's sole newly named outside partner. Gross margin improved to 40.4% GAAP (41.8% non-GAAP), but still trails TSMC's 67.7%, underscoring that process maturity and utilization — not just node availability — remain the industry's real moat.
The Balance Sheet Behind the Rally
The capex ramp is not free. Adjusted free cash flow was negative $8.4 billion for the quarter, and weighted-average shares outstanding rose roughly 17% year-over-year as government equity claims expanded. Intel ended the quarter with approximately $29.7 billion in cash and short-term investments against roughly $50.5 billion in debt — a position that leaves limited room for another major execution misstep on 18A or its successors.
Where the Market Gets Ahead of Itself
Here is the crux for anyone allocating capital today. Heading into the print, Intel traded near 8.5 times trailing sales and roughly 110 times forward earnings, already up approximately 170% year-to-date. At the after-hours price near $111, implied equity value approaches $565 billion — a figure consistent not merely with cyclical recovery, but with successful 18A execution and eventual foundry profitability that has not yet been demonstrated.
That is the mispricing worth naming plainly: the market is correctly rewarding Intel's survival and its product-segment recovery, but it is prematurely capitalizing a TSMC-like foundry outcome that remains unproven. The narrowing Foundry loss lowers turnaround risk; it does not yet establish external customer economics or pricing power. The asymmetry has consequently shifted — further estimate beats offer only incremental upside, while any disappointment on yields, customer commitments, or capex discipline invites sharp multiple compression.
The Practical Takeaway
For allocators, the more efficient way to express conviction in Intel's turnaround may be through the equipment and advanced-packaging suppliers profiting regardless of which foundry ultimately wins, rather than the equity itself — with an Intel upgrade reserved until the company shows two consecutive quarters of positive normalized free cash flow, sub-$1 billion quarterly Foundry losses, and disclosed external backlog exceeding $1 billion per quarter. For executives evaluating capacity, the message is measured engagement: reserve 18A allocation for 2027-2028, but keep it to a modest share of critical volume until Intel proves delivery reliability across full customer ramps. Intel has re-established relevance. It has not yet finished earning its valuation.
not investment advice