Terafab Names 14A on the Call: The Single Gating Item on the Foundry Thesis Just Got Delivered — Conviction Raised

Results vs. Consensus

Quality of the Beat

• Revenue: Broad-based beat, not concentrated. CCG +1% Y/Y is in line; DCAI +22% and Foundry +16% are the surprise. CFO attributed the beat to "higher volume, which included previously reserved inventory" plus favorable mix and pricing. Organic — Altera is out of the base as of September 2025. Zinsner noted the quarter benefited from supply catching up to demand that had been reserved/deferred, implying Q1 is both a demand and an execution beat.
• Gross margin: 41.0% NG GM, +180 bps Y/Y and ~650 bps above the Q1 guide. Zinsner credited (a) volume leverage including the reserved-inventory release, (b) mix, (c) "better yields on Intel 18A." The Q2 guide at 39.0% implies a ~200 bps step-down as Panther Lake volumes increase 6-7x (below-corporate-average margin) and memory input costs rise. The release/reserved-inventory dynamic is the modest one-time piece — the underlying trajectory is still up.
• EPS: Non-GAAP EPS $0.29 is clean — $1.49B net income on $1.67B operating income, ~11% tax rate consistent with Q2 guide. Share count flat. No below-the-line help. The beat is operating leverage, not financial engineering.
• GAAP loss: $(3.73B) net loss fully explained by $4.07B of restructuring/impairment (primarily Mobileye goodwill write-down — not new operating deterioration; MBLY itself reported +27% Y/Y Q1 revenue the same week) plus $1.09B non-cash MTM on the US government Escrowed Shares. Strip both and GAAP earnings are meaningfully positive. Zero Street analysts pressed on Mobileye during Q&A — already written off as non-operating.

Segment Performance

Client Computing Group (CCG) — The Cash Cow Is Fine

CCG at +1% Y/Y landed on-plan. Volume is steady; the news inside CCG is the full-volume ramp of Core Ultra Series 3 and Intel 18A-based Core Series 3. Tan called this "the fastest new product ramp in five years" — a claim worth taking seriously given the product-execution cadence under this leadership team. Zinsner separately flagged that Panther Lake volumes will increase 6-7x in Q2 as the ramp scales, and acknowledged that at early-ramp yield and cost these units run below the corporate-average margin. That is the direct mechanical explanation for the Q2 GM guide at 39.0% vs. the Q1 actual of 41.0%.

"Core Ultra Series 3 and Intel 18A–based Core Series 3 entered full-volume production — this is the fastest new product ramp for Intel in five years." — Lip-Bu Tan, CEO

Assessment: CCG is doing exactly what it needs to do — stay flat-to-slightly-up while 18A volumes ramp. The FY26 PC TAM will be down "low double-digit percent" per management, so flat-to-up CCG implies modest share gain plus ASP expansion from premium SKUs. CCG is not the story; it's the denominator.

Data Center and AI (DCAI) — Xeon Is Pulling With AI, Not Against It

DCAI revenue of $5.1B (+22% Y/Y) is the single most important segment print Intel has produced in eight quarters. The entire 2024-2025 bear thesis on DCAI was that hyperscalers and enterprises were shifting datacenter capex from CPU-heavy general compute to GPU-heavy AI training, with Intel caught in the draft. Q1 2026 shows the opposite dynamic operating at scale. Zinsner attributed growth primarily to units, not ASP, with core counts expanding significantly.

"The CPU is reinserting itself as the indispensable foundation of the AI era. This isn't just our wishful thinking, it's what we hear from our customers. Customers are deploying server CPUs alongside accelerators in a ratio that is moving back towards CPU." — Lip-Bu Tan, CEO

The Nvidia DGX Rubin NVL8 host-CPU attach is the proof point. Nvidia's DGX reference systems are what enterprises and second-tier clouds copy — a design win there propagates horizontally across the ecosystem. More directly, Xeon 6 is the named host CPU for Nvidia's next-generation reference architecture, which affirmatively closes the "AMD EPYC takes the AI host socket" narrative that had hung over DCAI for two years. Zinsner also pointed to the Granite Rapids → Diamond Rapids → Coral Rapids roadmap with simultaneous multithreading returning at Coral as a direct response to ARM (Graviton, Axion) per-socket TCO competition.

Assessment: This is a structural growth reacceleration, not a Q-over-Q bump. A volume-led +22% with rising core counts is the highest-quality composition — it implies socket share stability plus ASP uplift from core-count mix. The Nvidia Xeon 6 attach plus the Google multi-year agreement (see Key Topics) give DCAI multi-year visibility that it didn't have entering 2026.

Intel Foundry — Attach Is No Longer Hypothetical

Foundry revenue of $5.4B (+16% Y/Y) is still mostly internal-wafer transfers, but the external-customer disclosures in this release are the most substantive in the foundry's history. Tan claimed on the call that 14A maturity, yield, and performance are tracking ahead of where 18A was at the equivalent point in time. Coming from a CEO who has made under-promise/over-deliver a stated discipline, this is a signal worth taking at face value.

"Intel 14A maturity, yield, and performance are outpacing Intel 18A at a similar point in time. We have multiple external customers actively evaluating the technology, and we expect early design commitments to emerge beginning in 2026." — Lip-Bu Tan, CEO

Zinsner on the quarterly cadence: external customer commitment signals are expected "in the back half of this year and into early next year" before Intel finalizes 14A supply plans. That tells us the capex-decision window for FY27+ is still open and could step up if additional anchors land. On advanced packaging specifically, Zinsner noted demand has moved from "hundreds of millions to billions of dollars per year" — no forward revenue guide, but a clear signal that advanced packaging is becoming a material foundry revenue line in its own right.

Assessment: Foundry is transitioning from a cost-absorption story to a revenue story. The Fab 34 Ireland Apollo buyback disclosed in the release (consolidating ownership, ~$250M/quarter non-controlling-interest drag Q2-Q4 2026, ~$1.1B in 2027 and 2028) is a capital-structure vote of confidence. And the Terafab commitment (below) is the anchor the business case required.

Key Topics & Management Commentary

Overall Management Tone: Confident but disciplined. Tan repeatedly used the "under-promise, over-deliver" framing, declined to name customers the customers didn't want named, and stuck to the one-quarter-forward guidance convention even with visibility that clearly supports more. Zinsner was candid about demand exceeding supply ("starts with a 'b'") and refused to be drawn into precise yield numbers (called them proprietary). This is a call from a team that is both winning and careful not to get ahead of execution — a material tone shift from the Gelsinger-era "next quarter it turns" posture and even from the Tan-era Q4 2025 defensive framing.

Topic 1: Terafab Commits to 14A — The Anchor Thesis

The most important single development on the call — and, arguably, in Intel's entire post-2021 narrative arc — is that Tesla's Terafab project has been confirmed to use Intel 14A as its process technology. The engagement was initially announced in early April 2026 when Intel joined the Terafab consortium (alongside SpaceX, xAI, and Tesla). Musk has since publicly stated that Terafab will be built on Intel 14A, with SpaceX taking responsibility for high-volume manufacturing in what appears to be a licensing-plus-partnership structure.

"Elon and I share a strong conviction that global semiconductor supply is not keeping pace with the rapid acceleration in demand. Together we are looking for unconventional ways to improve manufacturing efficiency. I can think of no better partner than Elon Musk." — Lip-Bu Tan, CEO

Tan declined to spell out the commercial structure on the call — whether this is a traditional foundry offtake agreement, a technology license, or a co-investment with SpaceX carrying production capex. He referenced customer confidentiality preferences: "my style is under-promise, over-deliver, so we have no plan to announce the customer unless the customer wants to announce it, and we support that." Musk's separate public statements are what enabled the 14A node-level detail to surface.

Why this matters: the Q4 2025 recap explicitly called out "14A requires a significant external anchor customer" as the single dominant binary gating the Outperform thesis. Firm decisions were expected in H2 2026. Tonight's call delivers that commitment six months ahead of the expected decision window, and with a customer (Tesla + SpaceX + xAI) that brings (a) volume scale across three compute-intensive end markets, (b) a willingness to co-invest in manufacturing efficiency, and (c) a high-profile brand signal that attracts additional evaluators. Tan's statement that "multiple external customers" are evaluating 14A now has tangible credibility behind it because the first named anchor is on the board.

Assessment: This is the rating-change-worthy event. In a three-tier framework (Outperform / Hold / Underperform) we were already at the top, which is why we are framing this as maximum conviction rather than a notional upgrade. The relevant risk has shifted from "will 14A find a customer?" to "will 14A execute on schedule for the customers it has?" — a far healthier problem.

Topic 2: Google Long-Term Agreement + Xeon 6 Hyperscaler Traction

The press release disclosed the Google multi-year collaboration covering custom IPUs and Xeon 6 deployment. The call added texture: these are three-to-five-year volume-and-pricing agreements, several were signed in the quarter, and some remain confidential at customer request. This is a meaningful shift in how Intel does business with hyperscalers — volume-and-price commitments lock in DCAI revenue visibility in a way that quarter-by-quarter PO motion never did.

"Stay tuned. At the right time, we will announce other contracts." — Lip-Bu Tan, CEO

Combined with the Nvidia Xeon 6 / DGX Rubin NVL8 selection (also called out in prepared remarks), the emerging pattern is that the three most important silicon buyers in the world — Nvidia (in reference architectures), Google (in internal fleet), and Tesla/SpaceX (in Terafab) — are all making multi-year Intel commitments in the same quarter. This is not coincidence; it's the US industrial policy, capex scale, and 18A/14A readiness finally converging.

Assessment: The Google LTA is as important as the Terafab/14A item for forward revenue visibility. The stock is pricing this in; the question is whether additional similar contracts surface over the next two quarters. We expect at least one additional named hyperscaler custom ASIC win in H2 2026.

Topic 3: Supply Constraints Are Demand-Led

The Q4 2025 "supply-constrained Q1" framing produced a -17% stock reaction in January because the Street interpreted it as a demand warning in disguise. Tonight's call affirmatively closed that interpretation.

"I probably do not want to put a specific number on it. Let us just say it starts with a 'b.'" — David Zinsner, CFO, when asked how much demand Intel is missing due to supply constraints

This is the second tell. The first was the +11% revenue beat itself; the second is that management is signaling billions of dollars of unmet demand even after delivering above the high end of guidance. Zinsner on capex strategy: "We see a lot of demand, and we want to make sure we are catching up on the supply front." Tool spending is up ~25% YoY in FY26 while space/shell spending is "brought down pretty materially" — meaning Intel is accelerating productive capacity without building greenfield shells.

Assessment: Demand-led supply constraint is the single best operating condition a capital-intensive manufacturer can have. It creates pricing power, enables customer loyalty (because customers have to pre-commit), and makes incremental capex high-confidence. This is the Tan operating framework working — cost discipline plus strategic demand capture.

Topic 4: GAAP Noise — Mobileye and Escrowed Shares

The $(0.73) GAAP EPS print is fully explained by two below-operating-line items: a $4.07B restructuring charge (primarily Mobileye goodwill impairment) and a $1.09B non-cash MTM loss on the Escrowed Shares granted to the US government in Q3 2025. Neither was even questioned during analyst Q&A — the Street has correctly categorized both as non-operating.

On Mobileye specifically: the goodwill impairment reconciles Intel's carrying value to MBLY's public market capitalization, which had long exceeded the carrying value by a meaningful amount. Mobileye's own Q1 report (+27% Y/Y revenue, raised FY26 outlook) shows the underlying business is performing; the impairment is a purely accounting cleanup of a 2017 acquisition goodwill line. The Escrowed Shares MTM will continue to be noisy — counterintuitively, the better Intel stock performs, the larger the GAAP drag, because the liability associated with those shares marks up on a rising stock.

Assessment: Housekeeping. Strip both items and GAAP earnings are meaningfully positive. Models should adjust out both lines to get the underlying earnings picture. No thesis implication.

Topic 5: Capex and Capital Allocation Discipline

FY26 gross capex is guided roughly flat year-over-year, but the mix is shifting: tool spending up ~25%, shell spending down materially. This is the correct posture for a company with demand-led supply constraints — add productive tools in existing shells rather than building new fab capacity. Zinsner flagged that external customer signals from H2 2026 into early 2027 will drive the 14A supply plan finalization, creating optionality to step up capex if additional anchors land.

On the Fab 34 Ireland Apollo buyback (disclosed in the release, unpacked in the Q&A): the 49% minority repurchase comes with ~$250M/quarter of non-controlling-interest drag in Q2–Q4 2026 and ~$1.1B in each of 2027 and 2028 — a multi-year FCF headwind, but one that comes with full ownership of capacity and a statement of balance-sheet confidence.

Assessment: This is a capital-allocation posture consistent with a company that has (a) visibility into demand it wants to capture, (b) selectivity on where it deploys capital, and (c) confidence that its balance sheet (post the US government + SoftBank + Nvidia capital raises and $4.3B debt paydown in 2025) can carry through on its own.

Guidance & Outlook

Q2 guidance is the single most important line-item in this release. Revenue is guided to accelerate sequentially from $13.6B to a $14.3B midpoint (+5% Q/Q) — the exact opposite of the sequential flattening the Street was modeling. Non-GAAP EPS of $0.20 is more than double the pre-print consensus. Gross margin at 39.0% is down ~200 bps Q/Q on 18A ramp dilution (Panther Lake volumes 6-7x Q/Q) plus memory input cost pressure — both of which are structural (18A) or transitory (memory) rather than a mix deterioration.

Implied 2H ramp: H1 2026 is tracking $13.6B + $14.3B = $27.9B. For FY26 to hit the $55–56B range our model now supports, H2 needs to approximate H1. With server CPU units guided to "double-digit" growth for FY26 and FY27 momentum called out specifically by the CFO, that's a tractable framing. Management continues to refuse full-year revenue guidance despite obviously having the visibility — we read this as discipline, not concern. Every quarter in the sequence from here has upside revision risk to the prior-Street view.

Gross margin trajectory: 41.0% Q1 has a small one-time release component (previously reserved inventory). 39.0% Q2 guide is the nearer-term run-rate. H2 should trend back toward the low 40s as Panther Lake yields mature and Fab 34 absorption improves. The 2H exit rate at ~41% is supportable.

Guidance style: Confident, not sandbagged. Intel has not historically delivered massive in-quarter beats against its own prior guide — Q4 2025 and Q1 2026 are both printing well above the prior quarter's guide, which is a meaningful shift in how the guidance function is being calibrated under this CFO. Backlog visibility appears higher than it has been in 2+ years.

Analyst Q&A Highlights

Topic: 14A / Terafab Structure

• Timothy Arcuri, UBS: Probed the Terafab engagement structure — licensing, volume commitment, or co-investment. Tan declined specifics; emphasized a "multi-customer engagement" and a "very broad relationship."
  Assessment: The commercial structure is deliberately ambiguous. What matters for the thesis is that 14A now has a named end customer with capacity commitments; the financial engineering (capital-light license vs. offtake) is a forward-looking modeling question, not a thesis question.

Topic: DCAI Growth Composition

• Vivek Arya, Bank of America: Asked whether the +22% Y/Y was volume-driven or ASP-driven. Zinsner: units are primary driver, with core counts increasing significantly. Arya pressed on ARM (Graviton, Axion) competitive dynamics; Tan pointed to Granite Rapids → Diamond Rapids → Coral Rapids with SMT returning at Coral as the ARM-response roadmap.
  Assessment: Volume-led growth with rising core counts is the highest-quality growth composition. It implies socket share stability plus mix-driven ASP uplift. This is exactly the decomposition we would want to see to validate a structural DCAI reacceleration thesis.

Topic: Foundry External Customer Timing

• Ben Reitzes, Melius: Asked when Intel would finalize 14A supply plans based on customer signals. Zinsner: customer signals expected "in the back half of this year and into early next year" before finalizing supply plans.
  Assessment: The capex-decision window for FY27+ is still open. If additional anchors land in H2 2026, capex could step up materially. Conversely, if only Terafab ultimately commits, the scale of 14A build-out will be moderated — which is fine for margin and FCF, just smaller absolute revenue opportunity.

Topic: Gross Margin Sustainability

• Stacy Rasgon, Bernstein: Flagged that 18A mix dilution in Q2 offsets server GM tailwinds. Zinsner acknowledged Panther Lake is below corporate-average at early ramp; expects improvement as yields mature. Called out that Q1 benefited from release of previously reserved inventory — a one-time contribution.
  Assessment: Candid and sober. The 41.0% Q1 is not the run-rate; the 39% Q2 guide is nearer the truth, with trajectory back up in H2. The right way to model H2 exit rate is ~41%, not 39% nor 41%+.

Topic: Supply Constraints Quantification

• Multiple analysts (paraphrased): Pressed on the magnitude of demand Intel cannot serve. Zinsner's answer — "I probably do not want to put a specific number on it. Let us just say it starts with a 'b'" — is the most telling admission in the call. Billions of dollars of demand exist at current prices that Intel cannot currently supply.
  Assessment: This is the demand-led supply constraint evidence the Street needed. Q4 2025's "supply-constrained" framing is now definitively demand-strength, not demand-concealment.

Topic: Mobileye

• No analyst pressed on the $4.07B goodwill impairment. Zinsner's only reference was the "strong quarter" in All Other (up 9% Q/Q). Mobileye separately reported Q1 revenue +27% Y/Y and raised FY26 outlook.
  Assessment: The market has correctly categorized the impairment as non-operating and written it off. Zero analyst attention suggests zero residual overhang risk from this item.

What They're NOT Saying

1. No full-year 2026 revenue guidance: Despite visibility that plainly supports it (sixth consecutive beat, three-to-five-year LTAs signed in-quarter, Q2 guide +9% above Street), management held the line on one-quarter-forward convention. We read this as Tan's "under-promise, over-deliver" discipline, not concern — but investors modeling FY26/FY27 are forced to build up from Q2 guidance and segment commentary.
2. No Terafab commercial structure: Tan deflected on whether this is an offtake, license, or co-investment. The structure has major implications for capital efficiency — a licensing arrangement with SpaceX carrying production capex is meaningfully more valuable to Intel NPV than a traditional offtake.
3. No 18A yield numbers: Zinsner called them "proprietary." Directionally said Intel will hit its mid-year yield targets "likely met by mid-year." Lack of precision is notable but consistent with competitive discipline.
4. No revised FY26 OpEx dollar figure: Zinsner acknowledged OpEx will run above the ~$16.5B prior target due to inflation, variable comp, and targeted investments — but provided no replacement number. A mild soft signal that OpEx discipline is loosening at the margin.
5. No additional named custom ASIC wins: Beyond the Google IPU reference, no additional hyperscalers named. Tan's "stay tuned… we will announce other contracts" is a tease — but PMs should not model it until named.
6. No foundry breakeven year recommitment: Absent from the release and the call. This was not explicitly re-timed; investors looking for a firm "2027 foundry profitability" bookend did not get it tonight.
7. No update on US government Escrowed Shares mechanics or SoftBank funding progress: Neither topic was revisited. Treated as settled structural items for now.

Market Reaction

• Regular-session close (2026-04-23): $66.78, after a +3.5% intraday lift on a BNP Paribas upgrade (Underperform → Neutral; PT $34 → $60) that landed pre-print.
• After-hours peak: $79.74 (+19.4% vs. regular close). Initial AH move was +15–16% on the press release cross; additional +3–4% came as Tan's Terafab / 14A commentary surfaced on the call.
• Next-day (April 24, 2026) indicative close: ~$77.46 (+16% vs. prior day). The move held through the regular session with only a modest fade from the AH peak.
• Volume: Elevated; among the heaviest trading days in INTC's post-2020 history.
• YTD context: INTC entered 2026 at ~$37.50. After the Q4 2025 -17% reaction, stock bottomed near $31. The rally through April reflects (a) the BNP Paribas upgrade and earlier PT raises, (b) the March Terafab consortium announcement, (c) anticipation into Q1. Post-print, INTC is up roughly +107% YTD — a dot-com-era-high level of positive single-day post-earnings reaction.
• Analyst reactions within 48h:
  • BNP Paribas (pre-print, same day): Underperform → Neutral; PT $34 → $60; cited 14A data points and AI-agent demand lifting pricing power.
  • Several additional price-target raises and ratings actions are expected in the next 24–48 hours given the Q2 guide magnitude and Terafab confirmation.

Why the stock reacted the way it did: the move is a fundamental re-rating, not a squeeze, evidenced by the next-day durability. The Flashcap's +15–25% gap-up range was inside the realized move. The three drivers, in order of magnitude: (1) Terafab/14A commitment being surfaced on the call — the gating item delivered; (2) Q2 guide +9% above Street — the pattern break from Q4 2025; (3) DCAI +22% + Nvidia Xeon 6 host attach — the structural narrative shift. The base-case fundamental data points alone supported a +15% move; Terafab is what takes it to the high teens.

Street Perspective

Debate: Is the Terafab Commitment the 14A Anchor?

Bull view: Musk has publicly named 14A as the Terafab node; Tan confirmed the engagement on the call; the combination of SpaceX manufacturing ambition, xAI AI-infrastructure needs, and Tesla autonomy/robotics silicon demand provides multi-end-market volume at scale. This is the anchor.

Bear view: The commercial structure is still undefined; Terafab is a speculative project that has not broken ground; Musk's timeline predictions carry a well-documented discount factor; a non-traditional licensing structure could leave Intel with less revenue than a conventional foundry customer.

Our take: Even the bear case here is a win. The gating item was "will 14A find a named external anchor" — that binary is now resolved positive. The commercial-structure concerns are second-order modeling questions that affect the magnitude of the opportunity, not the existence of it. "Multiple customers evaluating" also means Terafab is not the only potential commitment. The bull view has the stronger argument.

Debate: Is 41% Gross Margin Sustainable?

Bull view: 18A yield improvement, mix shift toward server with rising core counts, and the release of the "previously reserved" inventory drove the print; underlying trajectory is structurally up. Q2 step-down to 39% is mechanical (Panther Lake dilution + memory costs), not thesis-relevant.

Bear view: 41.0% included a one-time release of reserved inventory; Panther Lake runs below corporate-average for several quarters; memory input costs are trending up; the run-rate is closer to 38–40% than 41%+.

Our take: Both sides are partially right. Run-rate near-term is 39–40%; H2 exit is back at ~41%; FY27 average above 41% is credible given Panther Lake maturity and foundry external mix. The structural direction is up, the near-term path is choppy. 41% is directionally where the business is heading; 39% is where the next quarter specifically is.

Debate: Is the Supply-Constrained Narrative Demand Strength or Execution Failure?

Bull view: Zinsner's "starts with a 'b'" admission of unmet demand plus the +22% DCAI print plus the three-to-five-year LTAs signed in-quarter demonstrate this is genuine demand pressure; Intel is prioritizing high-margin allocations and is rational not to chase every dollar.

Bear view: A mature semiconductor company should not be demand-constrained in a market this size; if execution were better Intel would be printing $15B+ revenue quarters; the constraint is self-inflicted and reveals capacity misallocation from prior years.

Our take: The bull case is the correct read. Demand-led supply constraints are the best operating condition a capital-intensive business can be in — they create pricing power, customer pre-commitment dynamics, and high-confidence incremental capex economics. The "execution failure" framing is backward-looking; the forward-looking posture is that Intel is rational to prioritize high-margin server allocations over lower-margin fill work.

Debate: Is the Stock Ahead of the Fundamentals at $77?

Bull view: FY26 non-GAAP EPS power now ~$1.00–1.20 vs. pre-print consensus ~$0.50; FY27 scales to $1.75–2.00 on Panther Lake full-year + 14A Terafab ramp + foundry external revenue scaling. At 25–30x forward EPS (appropriate for a turnaround leader with explicit US industrial policy tailwinds and named foundry anchor), fair value is above $77.

Bear view: Stock is up +107% YTD; +19% in a single AH session; the easy gap has been closed; FY27 earnings power is three years from realization and depends on execution assumptions (Terafab timing, 14A yield, foundry external mix) that can slip.

Our take: The bear view on entry-point discipline has merit — chasing +19% AH moves is rarely good process. But the underlying fair-value framework is intact. PMs sized in from Q3 2025 / Q4 2025 have meaningful embedded gains and should hold. PMs adding fresh should wait for pullback rather than chase, but the 12-month return outlook remains Outperform-consistent.

Model Implications

Valuation: With FY26 non-GAAP EPS power ~$1.10 and FY27 framework $1.75–2.00, a 25–30x forward multiple on the FY27 earnings base gets to a $44–60 fair value range on a prior-methodology basis — but the correct post-print lens includes (a) foundry external revenue optionality through FY28 as 14A ramps beyond Terafab, (b) multiple expansion credit for explicit US industrial policy support, and (c) the reduction in thesis risk from the Terafab anchor resolution. A 30–35x multiple on FY27 $2.00 EPS reaches the $60–70 range; the additional premium above $70 that the stock is showing at $77+ is pricing in FY28+ foundry external revenue scaling and the "multiple customers evaluating" optionality turning into more named wins. We think that's largely appropriate but would prefer not to chase the +19% AH move at current levels.

Thesis Scorecard Post-Earnings

Overall: Thesis materially strengthened. The dominant unresolved binary from the backfill summary — 14A anchor — is resolved. The second backfill binary (Q1 supply trough recovery) is resolved with a sequential acceleration guide. The third (custom ASIC customer disclosures) is partially resolved via the Google LTA. Only the foundry-breakeven-year recommitment remains explicitly open.

Action: Hold existing positions through the reaction. Do not chase the +19% AH move; wait for any pullback for incremental adds. For new coverage exposure, scale in rather than build full position at current levels. The 12-month outlook supports Outperform; the operating conditions (demand-led supply, LTA lock-ins, 14A anchor named, 18A in HVM) are the best Intel has experienced since the mid-2010s.