INTC · Q2 2025 Earnings
BearishIntel
Reported July 24, 2025
30-second summary
Intel printed $12.9B revenue (+0.2% YoY) — above the high end of its own guidance range — but lost $2.9B at the net line, with GAAP gross margin collapsing to 27.5% and operating margin at -24.7%. New CEO Lip-Bu Tan used the call to reset expectations on the foundry roadmap, stating "I do not subscribe to the belief that if you build it, they will come" and committing to deploy capex only against tangible volume commitments — a meaningful departure from the prior posture. On 14A specifically, Tan said the increased capital costs require both Intel products and a meaningful external customer to drive acceptable returns, and "I will only invest when I'm confident those returns exist." Q3 guidance of $12.6-13.6B revenue with breakeven non-GAAP EPS implies continued stagnation despite the Q2 topline beat, and the 75,000 core-headcount target signals deeper cuts than previously telegraphed.
Headline numbers
EPS
Q2 FY2025
$-0.10
Revenue
Q2 FY2025
$12.90B
+0.2% YoY
Gross margin
Q2 FY2025
27.5%
Operating margin
Q2 FY2025
-24.7%
Key financials
Q2 FY2025| Metric | Q2 FY2025 | YoY |
|---|---|---|
| Revenue | $12.90B | +0.2% |
| EPS | $-0.10 | — |
| Gross margin | 27.5% | — |
| Operating margin | -24.7% | — |
Guidance
Prior quarter data unavailable — comparison not possible.
Segment performance
Q2 FY2025| Segment | Q2 FY2025 | YoY |
|---|---|---|
| Client Computing Group (CCG) | $7.9B | -3.0% |
| Data Center and AI (DCAI) | $3.9B | +4.0% |
| Intel Foundry | $4.4B | +3.0% |
| All Other | $1.1B | +20.0% |
Capacity & utilization
Q2 FY2025| Segment | Q2 FY2025 |
|---|---|
| 2025 Gross Capital Expenditures Target | $18.0 billion |
| Core Workforce (end of 2025 target) | ~75,000 employees |
| Panther Lake Processor Timeline | First SKU shipping H2 2025, additional SKUs H1 2026 |
Profitability
Q2 FY2025| Segment | Q2 FY2025 |
|---|---|
| Non-GAAP Gross Margin | 29.7% |
| Non-GAAP Operating Margin | -3.9% |
| Operating Cash Flow (Q2) | $2.1 billion |
| 2025 Non-GAAP Operating Expenses Target | $17.0 billion |
| 2026 Non-GAAP Operating Expenses Target | $16.0 billion |
Management tone
The tone shift from the prior IDM 2.0 narrative to this quarter's posture is the loudest signal in the print. Three paragraphs:
From "regain process leadership" to "stabilize the business." Tan's framing — emphasizing "financial discipline," capital deployed "lockstep with tangible milestones and not before," and the explicit acknowledgment that "the capacity investment we made over the last several years were well ahead of demand and were unwise and excessive" — replaces the Gelsinger-era confident node-transition narrative with explicit triage language. The call's center of gravity is cost reduction, headcount cuts, and capex efficiency, not roadmap acceleration. Management now treats Intel as a turnaround, not a comeback.
From "build it and they will come" to volume-commitment-gated capex. Tan's verbatim "I do not subscribe to the belief that if you build it, they will come" is the most important sentence on the call — a direct repudiation of the prior capex posture. On 14A specifically, Tan said "I will make sure that until I see the internal customer, external customer volume commitment before I put CapEx into the operation," and framed 14A economics as requiring both Intel products and a meaningful external customer because the prior 18A-era math (returns achievable on Intel products alone) no longer holds at 14A capital intensity. Foundry binary risk is now explicit rather than hidden.
From "AI demand is a tailwind for Intel" to a deferred AI strategy. Zinsner pushed back on the share-loss narrative in Q&A, saying "we actually have held share relatively well, despite our position in the market in terms of performance" and crediting the x86 ecosystem. Our read is different: holding share with an admittedly uncompetitive portfolio is not a durable position, and Tan himself deferred full AI strategy disclosure to "the coming months" while citing inference and agentic AI as focus areas. For a company whose data center segment grew only 4% YoY, deferring AI strategy is a tell that there is no near-term differentiated answer.
Recurring themes management leaned on this quarter:
Risks management surfaced:
Q&A highlights
Ross Seymour · Deutsche Bank
How quickly can Intel fix the x86 business to enable foundry growth? What timeline should we expect for 18A and external customer ramps (2027-2028)?
Lipu stated 18A is foundation for next three generations of products with steady progress on yield and performance. Emphasized engagement with external ecosystem partners and government (secure/create programs). Trust-building through reliable delivery is prerequisite for external customer attraction. Internal product ramps will demonstrate capability before external customer acquisition.
Timothy O'Curry · UBS
14A hedging language suggests potential to walk away from development. How do you reconcile hedging on 14A with building external foundry business? Is this a pivot away from foundry?
Lipu clarified 14A development continues with laser focus on basic building blocks, PDKs, and foundation IP. Team applies lessons from 18A mistakes. Early customer engagement with clear milestones provides confidence. CapEx deployment contingent on customer volume commitments, performance, and yield validation. Not abandoning foundry but being disciplined with capital deployment.
Joseph Moore · Morgan Stanley
What is driving Intel 7 supply constraints through year-end? Will Intel 4 ramp alleviate this? Do you need additional wafer starts?
Dave attributed Intel 7 constraints to strong Raptor Lake demand at attractive price points. Mix shift expected toward Intel 4 with Meteor Lake and Granite Rapids ramps. Capacity expansion already underway for Intel 4.3. Arizona fab transitioning from construction to productive 18A ramp, reducing construction-in-progress from over $50B to mid-to-high $30s.
Stacey Roskin · Bernstein Research
If 18A supplies three generations (2026-2028), does 14A arrive 2029+? If 14A fails, does foundry strategy die? Can Intel run sustainable foundry on 18A internal volume alone?
Lipu confirmed 14A timing at 2028-2029, same as TSMC 14A. Will not deploy CapEx without clear customer milestones, yield validation, and volume commitments. 18A will reach peak volumes beginning next decade with good ROI on internal use. Multiple waves of external 18A adoption expected over time despite limited wave-one external uptake.
Benjamin Reitzes · Melius Research
What server CPU share trends expected for Q3? When will share losses dissipate? Can you clarify AI strategy focus and timing of full disclosure?
Dave indicated roughly flat sequential server/CPU guidance without segment detail. Noted Granite Rapids as better part; Diamond Rapids next-gen will improve competitive position. Share stabilization will take time despite strong x86 ecosystem. Lipu stated AI strategy focuses on inference and agentic AI, leveraging x86 CPU franchise with accelerators. Full AI strategy disclosure coming in future months. Adding software talent and incubating startup ideas to build system/software stack capabilities.
What to watch into next quarter
Q3 revenue vs. the $12.6-13.6B guide and whether DCAI growth re-accelerates above 4% YoY — anything sub-5% confirms competitive share dynamics are structural, not cyclical, despite management's "held share well" framing.
Non-GAAP gross margin progression toward the 36.0% Q3 guide vs. the 29.7% printed this quarter — note that Q2 non-GAAP GM ex ~800bps of impairment and one-time period charges would have been ~37.5% per Zinsner, so on a clean basis the 36.0% Q3 guide is actually a slight step-down driven by Lunar Lake mix and early Panther Lake ramp costs. Watch whether management can hold the line as Panther Lake yields mature.
Concrete 14A external customer commits disclosed by year-end — Tan tied 14A capex deployment to customer signups; absence of named commits by Q4 effectively kills the leading-edge foundry thesis.
Panther Lake first SKU shipment confirmation in H2 2025 — any slip from the H2 commit would replay the historical Intel execution credibility problem and undermine the new management reset.
Workforce reaching the ~75,000 core target and operating cash flow trajectory vs. the $2.1B printed this quarter — cash generation needs to be sustainable to fund the $18B 2025 capex without further balance sheet stress.
Sources
- Intel Q2 2025 Earnings Press Release, SEC Archives: https://www.sec.gov/Archives/edgar/data/50863/000005086325000107/q225_earningsrelease.htm
- Intel Q2 2025 Earnings Call Q&A commentary (analyst exchanges)
Get the next brief, free.
We publish analyst-grade earnings briefs the same day or morning after every call — headline numbers, segment KPIs, Q&A highlights, and tone analysis. Free during beta.
This is not investment advice.