Intel Q2 2026 revenue jumps 25% to $16.1bn as DCAI operating income quadruples

Intel Corporation (NASDAQ: INTC) reported second-quarter 2026 revenue of $16.1 billion, up 25 per cent year-on-year and the company’s strongest quarterly top-line growth in more than fifteen years, alongside a non-GAAP earnings per share of $0.42 that comfortably cleared consensus of roughly $0.22. The result marks the clearest quantitative validation to date of the operational reset led by chief executive Lip-Bu Tan and chief financial officer Dave Zinsner, with the Data Center and AI Group operating income roughly quadrupling and Intel Foundry losses narrowing even as revenue in that segment grew 31 per cent. The company simultaneously raised full-year 2026 capital spending guidance from $18 billion to $20 billion, signalled a “meaningful” further increase in 2027, and reported a $12.5 billion non-cash mark-to-market charge tied to escrowed shares held for the United States Department of Commerce under the CHIPS Act Secure Enclave arrangement. The central tension is now sharp and specific. Intel has delivered the products-side proof of a genuine cyclical recovery, but the foundry monetisation catalyst that a large share of the equity thesis rests on, a named external anchor customer contributing to revenue on the balance sheet, still has not landed.

How did Intel’s Q2 2026 revenue mix break down between Intel Products and Intel Foundry?

Total Intel Products revenue reached $15.1 billion, up 28 per cent, with the newly renamed Client Computing and Physical AI Group at $8.9 billion (up 13 per cent) and the Data Center and AI Group at $6.3 billion (up 59 per cent). Intel Foundry contributed $5.8 billion in reported segment revenue, up 31 per cent, though intersegment eliminations of $5.5 billion mean the bulk of that foundry activity remains internal wafer supply to Intel Products rather than genuinely external third-party demand. The “All Other” category, which now excludes Altera following the September 2025 sale of 51 per cent of that programmable-logic business to Silver Lake, fell 33 per cent to $701 million as expected. Non-GAAP gross margin expanded to 41.8 per cent, up 12.1 percentage points year-on-year, and non-GAAP operating margin swung from a negative 3.9 per cent in the prior-year quarter to a positive 17.2 per cent. Dave Zinsner attributed the beat to volume upside from higher factory yields and improved cycle times, describing the demand backdrop for AI-driven compute as continuing to strengthen.

Why did the Data Center and AI Group operating income quadruple in a single quarter?

The Data Center and AI Group generated $2.474 billion in operating income against $633 million a year earlier, a jump that materially exceeds the growth in the segment’s revenue base and points to genuine operating leverage rather than a simple mix effect. Two dynamics matter here. First, the industry-wide server CPU shortage that Morgan Stanley analyst Joseph Moore flagged pre-print has allowed Intel to raise both output and prices in the double digits, precisely the environment in which a scale incumbent with existing wafer supply captures disproportionate margin. Second, agentic AI workloads are consuming meaningfully more general-purpose CPU capacity than earlier inference architectures, and Intel’s positioning around rack-scale infrastructure with SambaNova Systems and Foxconn, plus the Vector Core Compute disaggregated agentic cloud combining Xeon processors with SambaNova reconfigurable dataflow units and NVIDIA Blackwell GPUs, targets that shift directly. The launch of Xeon 6+, described by Intel as its first server-class product on Intel 18A engineered for sustained performance under real-world power constraints, is the flagship product datapoint of the quarter and gives the sales organisation a leading-edge SKU to defend against Advanced Micro Devices’ EPYC roadmap.

What does the Panther Lake ramp and Intel 18A-P risk production mean for the process node story?

Intel confirmed that a subset of Intel Core Ultra Series 3 processors, code-named Panther Lake, has now entered high-volume manufacturing on Intel 18A using ASML Holding’s High-NA EUV lithography, the first commercial deployment of that generation of extreme ultraviolet equipment by any logic manufacturer. Alongside that, Intel 18A-P entered risk production on the timeline previously communicated to customers, extending the platform with enhanced performance, power and thermal characteristics. In parallel, the company announced a €5 billion investment to expand manufacturing capacity for Intel Xeon 6 and next-generation Xeon processors on the older Intel 3 node, and expanded capacity at the Bowers campus to support Intel Mask Operations across current and future process technology development. Read together, these updates indicate that the 18A generation is now transitioning from proof-of-yield into volume economics, a critical shift because the fixed cost of leading-edge fabs only earns its return once wafers move through them at competitive utilisation. The Fortinet Security Processor 6 collaboration, disclosed as a strategic engagement to develop that networking silicon using Intel design, packaging and manufacturing capabilities, extends the purpose-built silicon franchise beyond internal networking and infrastructure processing units.

Why did Intel Foundry still post a $2.1 billion operating loss and what does the market want to see next?

Intel Foundry revenue of $5.765 billion was accompanied by a segment operating loss of $2.089 billion, an improvement from the $3.168 billion loss in the prior-year quarter but still the structural drag on group profitability. Most of that reported revenue reflects internal wafer supply to Intel Products, not third-party foundry customers. This is precisely the gap several sell-side analysts flagged in post-print notes, with Morgan Stanley reiterating that conviction on the foundry business remains “low” and multiple firms observing that the most important catalyst for the Intel turnaround, a signed external anchor customer whose revenue lands visibly on the balance sheet, has still not arrived. The Fortinet engagement and previously reported design agreements around Intel 14A intellectual property matter directionally, but until a hyperscaler, a fabless leader or a strategic partner commits volume that shows up as external Intel Foundry revenue, the segment’s path to sustained gross profitability rests on internal transfer economics and process ramp rather than genuine third-party demand. This is the single hardest-to-forge datapoint in the current Intel narrative, and its absence is why Bank of America’s Vivek Arya could double-upgrade to Buy in June with a $135 target while Morgan Stanley, JPMorgan and Rosenblatt Securities remained at Equal Weight, Underweight and Sell respectively even after raising near-term price targets on the Q2 beat.

What is the $12.5 billion non-cash charge on Escrowed Shares and why does it explain the $11 billion GAAP loss?

The GAAP net loss attributable to Intel of $11.033 billion, versus non-GAAP net income of $2.197 billion, is dominated by a $12.5 billion mark-to-market charge on escrowed shares held for the United States Department of Commerce under the CHIPS Act Secure Enclave agreement. Under that arrangement, Intel common stock is held in escrow and released to the federal government as the company performs on and receives cash proceeds tied to the underlying grant. Because the share price rose sharply during the quarter, the derivative liability associated with those escrowed shares expanded, producing a large non-cash accounting charge that flows through the interest and other, net line of the income statement. The economic substance is that a rising Intel share price makes the government’s contingent equity claim more valuable in dollar terms, which registers as a liability on Intel’s books. This dynamic is not a proxy for operating performance and management, correctly in the view of Business News Today, excludes it from the non-GAAP reconciliation. Investors evaluating the quarter’s true operational profitability should anchor on the $2.2 billion non-GAAP net income and $2.77 billion non-GAAP operating income rather than the headline GAAP loss.

How does the $20 billion 2026 capex raise interact with Intel’s adjusted free cash flow of negative $8.4 billion in the first half?

Intel raised full-year 2026 capital expenditure guidance from $18 billion to approximately $20 billion, with Dave Zinsner indicating on the call that 2027 capex is expected to increase “meaningfully” beyond that. In the first half of 2026, gross capital expenditure totalled $6.192 billion within the consolidated cash flow statement, partner contributions from Semiconductor Co-Investment Program vehicles added $4.082 billion, partner distributions absorbed $14.339 billion, and adjusted free cash flow, as Intel defines it, registered at negative $8.419 billion. Cash and short-term investments closed the quarter at $29.7 billion combined, down from $37.4 billion at year-end 2025, while gross debt rose to $48.5 billion from $44.1 billion and non-controlling interests grew to $15.6 billion. The company also stated that it could tap the capital markets for further funds. Read together, these numbers frame the funding question that continues to sit alongside the operational story. The equity market has rerated Intel on the assumption that AI CPU demand and 18A ramp will convert into cash generation over the next several quarters, but until adjusted free cash flow turns positive, incremental capacity commitments have to be financed through some combination of operating cash flow, SCIP partner contributions, incremental debt issuance, government support and, potentially, further equity or equity-linked issuance.

How should investors read the Q3 2026 guidance and the split reaction across the sell side?

For the third quarter, Intel guided to revenue of $15.8 billion to $16.8 billion, GAAP earnings per share of $0.31 and non-GAAP earnings per share of $0.38, with non-GAAP gross margin of approximately 42 per cent. The midpoint of the revenue range implies a modest sequential improvement and comfortably clears the pre-print consensus of roughly $15.1 billion. Sell-side reaction has been asymmetric. Stifel and Susquehanna raised targets while retaining constructive stances, KeyBanc Capital Markets had already lifted its target to $155 pre-print, UBS took its target from $83 to $121, and JPMorgan’s Harlan Sur pushed his target from $45 to $85 while keeping an Underweight rating. Rosenblatt Securities lifted its target to $80 while maintaining a Sell. Bank of America’s Vivek Arya was already at Buy with $135 following the June double-upgrade. Morgan Stanley moved to $84 from $75 at Equal Weight, citing “mixed at best” server share and low conviction on foundry. The spread between the highest and lowest published price targets has widened rather than compressed, which is unusual after a clean beat and typically indicates that the market is pricing a bifurcated outcome distribution rather than a consensus view. The share price closed the earnings day at $104.00, up 3.76 per cent, before trading in a wide intraday range on 24 July around $95 to $104, against a 52-week high of $140.94 recorded on 22 June 2026 and a year-to-date gain of roughly 170 per cent. The market response validates the beat without extrapolating it linearly, and the muted reaction relative to the size of the first-quarter surprise, which produced a 23.6 per cent day-of move, suggests that a meaningful portion of good news is already in the price.

What should investors track as Intel converts Q2 momentum into a sustained turnaround under Lip-Bu Tan?

• Intel Corporation reported Q2 2026 revenue of $16.1 billion, up 25 per cent year-on-year, the strongest quarterly revenue growth in more than fifteen years, with non-GAAP EPS of $0.42 and $7.0 billion of cash generated from operations
• Data Center and AI Group operating income rose to $2.474 billion from $633 million a year earlier, reflecting the industry-wide server CPU shortage, AI agentic workload growth and pricing traction; this is now the most important near-term earnings contributor to the group
• Intel Foundry revenue of $5.8 billion (up 31 per cent) still generated a $2.089 billion operating loss and remains dominated by internal wafer supply to Intel Products; the missing signed external anchor customer is the single most important unresolved catalyst for the equity thesis
• Panther Lake entered high-volume manufacturing on Intel 18A using ASML High-NA EUV, Xeon 6+ became the first server-class product on Intel 18A, and Intel 18A-P entered risk production on schedule, together marking a genuine transition from yield proof to volume economics
• The $11.0 billion GAAP net loss reflects a $12.5 billion non-cash mark-to-market charge on escrowed shares tied to the CHIPS Act Secure Enclave agreement with the United States Department of Commerce, not deterioration in the underlying operating business
• Full-year 2026 capex was raised to $20 billion from $18 billion, with 2027 capex guided “meaningfully” higher; adjusted free cash flow was negative $8.4 billion in the first half and management indicated potential capital markets access to fund the ramp
• Q3 2026 guidance of $15.8 billion to $16.8 billion in revenue, non-GAAP EPS of $0.38 and non-GAAP gross margin of 42 per cent comfortably beat pre-print consensus and implies sequential margin expansion
• The sell side has bifurcated post-print, with Bank of America at Buy $135 and KeyBanc Capital Markets at Overweight $155 anchoring the bull case and Rosenblatt Securities at Sell $80 plus JPMorgan at Underweight $85 anchoring the sceptics
• The next measurable proof points are a named external foundry customer with volume commitments visible on the balance sheet, sustained gross margin expansion into the 43 to 45 per cent range on Intel 18A mix shift, and a clear inflection in adjusted free cash flow that reduces the equity’s dependence on external capital
• The share price closed at $104.00 on earnings day, off the $140.94 52-week high recorded on 22 June 2026, with year-to-date gains of roughly 170 per cent already discounting a meaningful portion of the operational recovery Lip-Bu Tan and Dave Zinsner are now delivering