Supply-Constrained Trough Cancelled: Revenue Crushes by 11%, Q2 Guide 9% Above Street, Foundry Attach Shows Up at Google and Nvidia
Initial Read: Blowout across every headline — revenue $13.6B vs. Street $12.3B, non-GAAP EPS $0.29 vs. $0.01, Q2 guide $14.3B midpoint vs. Street $13.1B, DCAI +22% Y/Y. The Q4 sell-off thesis (supply-constrained Q1 trough) is dead on arrival. Expect 15-25% gap up; AH tape is already +20%.
Key Takeaways
- Revenue of $13.6B beat consensus by ~$1.3B (+11%) and came in above the high end of management's own prior guide — Q1 was supposed to be the "supply-constrained trough" that drove the -17% Q4 reaction in January. Trough cancelled.
- DCAI +22% Y/Y to $5.1B and Intel Foundry +16% Y/Y to $5.4B — DCAI is the cleanest read we've had in years that Xeon is attaching to AI infrastructure spend, and Foundry growth with Google IPU co-dev and Nvidia Xeon 6 / DGX Rubin selection starts to answer the "who will actually use 18A externally" question.
- Non-GAAP gross margin 41.0% vs. 39.2% Y/Y (+180 bps) on higher volumes and mix; non-GAAP operating income $1.67B vs. $690M (+142%). The operating leverage inflection is real — this is the first quarter since the foundry turn where the non-GAAP P&L looks like a functioning semiconductor company rather than a restructuring story.
- Q2 2026 guide is the tell: revenue $13.8–14.8B (midpoint ~$14.3B) vs. Street $13.1B — 9% above consensus at the midpoint and a sequential acceleration, not a flattening. Non-GAAP EPS $0.20 vs. Street $0.09. Full-year non-GAAP OpEx reaffirmed at ~$16.5B.
- GAAP loss of $(3.7B) / $(0.73) EPS looks ugly but is entirely explained by a $4.1B restructuring charge (primarily Mobileye goodwill impairment) plus a $1.1B non-cash mark-to-market on the US government Escrowed Shares. Underlying operations are profitable.
- Rating: Maintaining Outperform. Our Q3 2025 upgrade thesis — that the foundry turn would produce real leading-edge attach before the Street capitulated — just got its single strongest data point. The -17% Q4 sell-off on the "supply-constrained Q1" guide was a gift, and this print validates maintaining (not upgrading further, only because we were already there).
Results vs. Consensus
| Metric | Actual (Q1 2026) | Consensus | Beat/Miss | Magnitude |
|---|---|---|---|---|
| Revenue | $13.6B | $12.26B | Beat | +10.9% |
| Non-GAAP EPS | $0.29 | $0.01 | Beat | +$0.28 |
| GAAP EPS | $(0.73) | n/a (one-time charges) | n/m | Impairment-driven |
| Non-GAAP Gross Margin | 41.0% | ~39% (implied) | Beat | +~200 bps |
| Non-GAAP Operating Income | $1.67B | ~$0.3B (implied from EPS) | Beat | >5x |
| Adjusted FCF | $(2.0B) | n/a | In line | Capex-heavy |
Quality of Beat
- Revenue: Beat is broad-based, not driven by a single segment. CCG roughly in line at +1% Y/Y ($7.7B), but DCAI (+22%) and Foundry (+16%) both materially exceeded the Y/Y-flat-to-down framing INTC used on the Q4 call. This is organic — Altera is out of the base as of Sept 2025, so the Y/Y comp is apples-to-apples on CCG/DCAI/Foundry.
- Gross margin: +180 bps of non-GAAP GM expansion Y/Y with a richer DCAI mix and — critically — an improving Foundry absorption as wafer volumes ramp. This is sustainable, not a one-time mix or pricing event. The 41.0% NG GM level is the highest INTC has printed since late 2023.
- EPS: Non-GAAP EPS of $0.29 is an operating beat, not a tax or share-count beat. Non-GAAP net income of $1.49B on non-GAAP operating income of $1.67B implies ~11% tax rate (consistent with Q2 guide) and no below-the-line help. Share count is flat. This is the real thing.
- GAAP loss: The $(3.73B) GAAP net loss is a non-issue for the investment case — it's $4.07B of restructuring/impairment (Mobileye goodwill write-down) plus $1.09B of non-cash MTM on the Escrowed Shares granted to the US government in Q3 2025. Strip both and GAAP earnings would be meaningfully positive. Mobileye write-down is a cleanup of a 2017 acquisition whose carrying value had long been indefensible given MBLY's standalone equity value.
Segment Performance
| Segment | Q1 2026 Revenue | Y/Y Growth | Our Assessment |
|---|---|---|---|
| Client Computing Group (CCG) | $7.7B | +1% | In line. Not the growth story — this is the cash-cow that funds the foundry. Panther Lake / Core Series 3 on 18A is now shipping, which matters more for FY27 ASP/margin than for Q1 volume. |
| Data Center and AI (DCAI) | $5.1B | +22% | The headline surprise. Xeon 6 attaching to Nvidia DGX Rubin NVL8 and the Google co-dev IPU program are both disclosed here — DCAI is no longer a share-loss story. This is the single most important segment print in 8 quarters. |
| Intel Foundry | $5.4B | +16% | Foundry revenue is still dominated by internal transfers, but the +16% growth with Fab 34 (Ireland) and 18A HVM ramping is the signal that the capital plan is producing volume. Google IPU co-dev is the first third-party advanced-node design engagement disclosed at this scale. |
| All Other (Altera-deconsol, Mobileye) | $0.6B | -33% | Optically ugly but fully explained by Altera deconsolidation (Sept 2025) and Mobileye weakness. Mobileye's goodwill impairment this quarter is the cleanup. Not thesis-relevant. |
CCG — The Cash Cow
CCG at +1% Y/Y is on-plan. The important news inside CCG is the commercial launch of Intel Core Series 3 (Panther Lake-class) on the Intel 18A process node — this is the first mainstream consumer product shipping on Intel's first HVM leading-edge node since the foundry-business turn. Volume contribution is modest in Q1, but the ASP/margin setup for 2H 2026 and FY 2027 is materially better than pre-18A.
DCAI — Xeon Attaches to AI Capex
DCAI revenue of $5.1B (+22% Y/Y) is the data point that reframes the thesis. The Q4 2025 concern was that hyperscalers were shifting datacenter capex from CPU-heavy general compute to GPU-heavy AI training, with Intel caught in the draft. This print — combined with Xeon 6 being selected as the host CPU for Nvidia's DGX Rubin NVL8 — shows the opposite dynamic: AI capex is pulling CPU demand because every GPU system still needs a capable host, and because inference/agentic workloads run on CPU. The CEO's "intelligence closer to the end user" framing in the release is a direct rebuttal to the "AI kills CPU" short thesis.
Intel Foundry — Attach Is Materializing
Foundry revenue of $5.4B (+16% Y/Y) is still primarily internal wafer transfers at standard cost, but the external-customer data points in this release are the most substantive in the foundry's history. The Google multi-year collaboration for custom IPUs is an advanced-node third-party design engagement at a hyperscale customer — the exact profile that the 14A business case requires. The Fab 34 Ireland 49% repurchase from Apollo also signals capital confidence. We still don't have a named 14A anchor committed, which was the Q4 gating item, but the 18A ecosystem is visibly filling in.
Notable Items in the Release
Mobileye Goodwill Impairment ($4.1B Restructuring Charge)
The entire GAAP net loss is a $4.07B restructuring and other charges line, primarily attributable to a goodwill impairment at the Mobileye reporting unit. Intel acquired Mobileye in 2017 for $15.3B, and while Intel retains majority ownership of the separately-listed entity, the carrying value on Intel's consolidated balance sheet had long exceeded MBLY's public market value. This is a non-cash accounting cleanup, not a new operating deterioration.
Assessment: Housekeeping, not news. The Street should look through it cleanly — the Q4 2025 Outperform thesis did not rely on Mobileye value, and removing the optical drag of an inflated carrying value actually improves future GAAP optics. Non-cash.
Escrowed Shares Mark-to-Market ($1.1B Non-Cash Loss)
The $1.09B below-the-line loss is tied to the US government equity arrangement from Q3 2025 (10% stake via Escrowed Shares). As INTC stock rallied post-Q3 2025 through Q1 2026, the liability associated with those shares marked up and flowed through the P&L. This is mechanically identical to how any equity-linked instrument would behave on a rising stock — paradoxically, the better the stock does, the larger the GAAP drag.
Assessment: Non-cash, non-operating, non-thesis-relevant. But worth flagging for PMs who see the GAAP line — this charge will continue to be noisy quarter-to-quarter and should be adjusted out in any model.
Google Multi-Year Collaboration on Custom IPUs
The release discloses a multi-year collaboration with Google covering custom infrastructure processing units (IPUs) and Xeon 6 deployment in Google datacenters. This is the first hyperscaler-scale advanced-node design engagement Intel has disclosed. No revenue figure attached.
Assessment: Strategically significant. A hyperscaler willing to co-develop silicon at Intel Foundry is the validation the foundry needs for the 14A business case. Watch the call for any hint on whether the IPU work is on 18A or 14A, and on the revenue ramp timing.
Nvidia Xeon 6 Attach on DGX Rubin NVL8
Nvidia selected Xeon 6 as the host CPU for its DGX Rubin NVL8 systems. This is a direct rebuttal of the "AMD EPYC is taking the AI host CPU socket" narrative that has hung over DCAI for two years.
Assessment: Modest revenue but large signaling value. DGX systems are the reference architecture that enterprises and second-tier clouds copy — an Xeon design win here propagates.
Fab 34 Ireland — 49% Minority Repurchase
Intel repurchased the 49% minority equity interest in Fab 34 (Ireland) from Apollo. The original 2022 deal brought in ~$11B of Apollo financing; the repurchase retires that and consolidates ownership. No terms disclosed in the release.
Assessment: Capital-structure-neutral to slightly dilutive on near-term FCF, but a statement of confidence that Intel now has the capital stack to own its advanced-node capacity outright. Inverse of the Q3 2025 moves (bringing in US government, Nvidia, SoftBank) — here Intel is buying out a dilutive financial partner.
Guidance & Outlook
| Metric | Q2 2026 Guide | Street (pre-print) | Delta |
|---|---|---|---|
| Revenue | $13.8B – $14.8B (mid $14.3B) | ~$13.07B | +9% vs. Street at midpoint |
| Non-GAAP EPS | $0.20 | ~$0.09 | +$0.11 (>2x) |
| GAAP EPS | $0.08 | n/a | — |
| Non-GAAP Gross Margin | 39.0% | ~38% (implied) | ~+100 bps |
| FY26 Non-GAAP OpEx | ~$16.5B (reaffirmed) | ~$16.5B | In line |
Our Interpretation
The Q2 guide is the single most important piece of this release. For the last three quarters, the frustrating pattern at Intel has been beat-and-lower — strong prints followed by guides that implied the next quarter would be weaker than the Street expected. Q4 2025 was the extreme version: the "supply-constrained Q1" framing drove the -17% reaction even as Q4 numbers themselves were fine. This Q2 guide breaks that pattern decisively. Revenue is guided to accelerate sequentially from $13.6B to a $14.3B midpoint (+5% Q/Q) and is 9% above what the Street was expecting.
Implied 2H ramp: The release reaffirmed full-year 2026 non-GAAP OpEx at ~$16.5B but did not provide a full-year revenue guide (INTC's convention remains one-quarter-forward). Taking H1 at $13.6B + $14.3B = $27.9B implies H2 needs to roughly match H1 to hit a $55–56B FY26 revenue number, which is well above the ~$52B Street range that was in place pre-print. Every quarter in this sequence now has upside revision risk.
Gross margin trajectory: Q2 NG GM guide of 39.0% is ~200 bps below the Q1 actual of 41.0%. Management will likely explain this as 18A ramp costs (Core Series 3 mix-in) and Fab 34 absorption — not a structural mix shift. If the call confirms that framing, 41% Q1 is sustainable as the H2 exit rate.
Guidance style: We read this guide as confident, not sandbagged. Intel does not have a tradition of massive in-quarter beats relative to the prior guide — they tend to guide close to their internal plan. A guide this far above the Street suggests backlog visibility management has not had in 2+ years.
Questions for the Call
- 14A external anchor customer: Q4 2025 explicitly flagged that 14A is gated on a committed external anchor. The Google IPU co-dev is a big step — is it on 14A or 18A? Bullish: Google IPU is on 14A and represents a committed volume agreement. Bearish: IPU is on 18A and 14A remains uncommitted, in which case the 14A capex decision is still an overhang.
- DCAI +22% — how much is AI host attach vs. enterprise refresh vs. one-time catchup? Bullish: >50% of Y/Y growth came from AI-host attach at hyperscalers (Nvidia DGX references, Google) — implies structural reacceleration. Bearish: most of the growth came from Granite Rapids catch-up vs. a weak Q1 2025 comp, with limited AI attach so far.
- Q2 sequential acceleration drivers: What's the split between CCG (Panther Lake ramp), DCAI (Xeon 6 ramp), and Foundry (Fab 34 absorption)? Bullish: broad-based with all three growing sequentially. Bearish: concentrated in one segment, implying less durability.
- Mobileye goodwill — is this the last of it? Bullish: this impairment takes Mobileye's carrying value down to public-market fair value; no further write-downs contemplated. Bearish: management hedges on future impairments, signaling Mobileye operating performance continues to deteriorate.
- Escrowed Shares MTM — how should investors model this forward? Bullish: management commits to a clear non-GAAP adjustment convention so the Street can model operating results cleanly. Bearish: management treats it as ongoing reported P&L with no clean adjustment, leaving GAAP permanently noisy.
- Full-year 2026 revenue framing: Will management provide any FY26 revenue bookend, or continue one-quarter-forward only? Bullish: management offers directional FY framing given the visibility that produced the Q2 guide. Bearish: refuses, citing continued visibility limitations — would partially undermine the confidence implied by the Q2 guide.
Market Reaction
After-hours tape at time of writing is +20%. Given the scale of the beat against expectations (revenue +11%, non-GAAP EPS +>$0.28 absolute) and the magnitude of the Q2 guide raise (+9% above Street on revenue, >2x on EPS), a 15–25% gap is consistent with our read. We do not expect this to fade materially intraday tomorrow barring a negative call surprise — the numbers themselves are clean and the forward guide is the key driver, not a one-quarter beat.
Model Implications
| Item | Prior View (Q4 2025) | Post-Q1 2026 | Reason |
|---|---|---|---|
| FY26 Revenue | ~$52B (Street-consensus range) | $55–56B | H1 tracking $27.9B vs. prior implied $25.5B; Q2 guide midpoint alone +9% above Street. |
| FY26 Non-GAAP GM | ~38–39% | 39.5–40.5% | Q1 printed 41.0%; Q2 guide 39.0% (on 18A ramp). Full-year trending ~100 bps above prior. |
| FY26 Non-GAAP EPS | ~$0.40–0.60 | ~$1.00–1.20 | Q1 alone was $0.29; Q2 guided $0.20; H2 implied ~$0.50–0.70 given OpEx discipline and GM trajectory. |
| DCAI trajectory | Flat-to-low-single-digit growth | Mid-teens growth | +22% Q1 with AI host attach structurally supportive. |
| Foundry thesis | Gated on 14A anchor | Partially de-risked | Google co-dev materially lowers 14A business case risk even if not yet confirmed on 14A node. |
| GAAP noise | Substantial | Still substantial | Mobileye goodwill largely cleaned up but Escrowed Shares MTM will persist. |
Valuation: If FY26 non-GAAP EPS power is ~$1.10 (midpoint of our new range) and FY27 scales to $1.75–2.00 on Panther Lake ramp / Foundry external revenue building, a 25–30x forward multiple on a turnaround leader with explicit US industrial policy support gets to a valuation well above the current tape even after the +20% AH move. Our prior Q4 2025 fair-value framework already implied this — this print accelerates the timeline rather than changing the ceiling.
Thesis Scorecard
| Thesis Point | Status | Notes |
|---|---|---|
| Bull #1: 18A in HVM unlocks margin and product-competitiveness inflection | Confirmed | Core Series 3 on 18A launched commercially in Q1; Q1 NG GM 41.0% is highest since late 2023. |
| Bull #2: Foundry will secure credible external customers on advanced nodes | Confirmed (partial) | Google multi-year IPU co-dev + Nvidia Xeon 6 DGX attach are the strongest external data points yet. 14A anchor still not explicitly named but the ecosystem is filling in. |
| Bull #3: US industrial policy (10% gov't stake, Nvidia/SoftBank) provides a capital backstop | Neutral | No change in Q1; Escrowed Shares MTM is a P&L noise item only. No new funding announcements. |
| Bull #4: DCAI is not being disintermediated by GPUs — AI capex pulls CPU attach | Confirmed | +22% Y/Y with explicit Nvidia DGX Rubin Xeon 6 attach disclosed. Highest-conviction datapoint in 8 quarters. |
| Bear #1: Supply-constrained Q1 2026 implies a deeper trough than consensus expects | Rejected | Q1 revenue +11% above Street, guidance accelerates sequentially. The bear thesis that drove the -17% Q4 reaction is simply wrong. |
| Bear #2: 14A capex decision without a committed anchor is a capital-allocation risk | Neutral | Risk lower given Google IPU deal, but not eliminated until 14A node specifically gets a named anchor. |
| Bear #3: Mobileye / goodwill is a hidden overhang | Rejected (largely) | $4.1B impairment recognized this quarter cleans it up. Non-cash. Carrying value now aligned with market. |
Overall: Thesis materially strengthened. The Q3 2025 upgrade from Hold to Outperform was premised on the foundry turn producing leading-edge product and external attach faster than the Street expected; Q1 2026 is the cleanest single data point supporting that call since the upgrade.
Preliminary Action: Hold existing positions through the call; add on any intraday weakness tomorrow. Do not chase the +20% AH move blindly — size discipline matters. If the call delivers a confirmed 14A anchor customer, this becomes a run-it-to-the-top position.