brain/
← all entities
entityINTCstock-market

Intel

Notes

Intel

One-line summary: US-domiciled IDM and contract foundry; the leading US alternative to TSMC for advanced-node capacity, with four confirmed/preliminary anchor foundry customers (AWS, Microsoft, Apple-prelim, Terafab/Musk); Samsung is now a credible second alternative but Intel retains first-mover advantage and CHIPS Act structural tailwind.

What it is

Integrated device manufacturer with an external foundry arm (Intel Foundry Services / IFS). CEO Lip-Bu Tan (since early 2025), CFO David Zinsner. Roadmap: Intel 18A in volume production at Fab 52 as of late 2025; 18A-P variant attracting Apple evaluation; Intel 14A at the PDK stage with customer commitments expected by year-end 2026. The US government holds an ownership stake ($8.9B CHIPS grant + $11B loans, largest in the program) — Trump administration has been actively brokering deals.

Why it matters to stock-market

Intel is the primary alternative supplier in every "TSMC is capped" thesis chain. When AI demand has rationed TSMC's leading-edge capacity (see tsmc) and tariffs/policy push toward US production, Intel is the natural fallback — though Samsung's Taylor fab complicates the "only alternative" narrative. Each new anchor customer reduces single-customer risk and validates IFS — an explicit re-rating mechanism (see us-fab-capacity-bottleneck).

Why it matters to artificial-intelligence

Energy is the physical ceiling on GPU / data-center buildout, which is the physical ceiling on frontier-model availability. pat-gelsinger in 2026-07-15-podcast-all-in-podcast-former-intel-ceo-on-what-went-wrong-what-s-next: "nobody's going to build and buy GPUs and build data centers if they don't have energy." His stated objective is to drop cost- and energy-per-token by ~10,000× / "five orders of magnitude" so Jevons explodes access (note: 10,000× is four orders, not five — his own numbers disagree; recorded on ai-roi-reckoning, not reconciled). ASR: "2001, when I came back" = 2021. Do not treat Intel's foundry/CHIPS numbers as AI-thread claims — those stay in the stock-market lens.

Valuation (August 27, 2026)

Last: $92.09 at the Thu 27 Aug 2026 4:00 PM EDT close. From 2026-08-28-valuation-marks-intc-tsm-aug-27-close (stockanalysis INTC).

Gage's judgment (as of 2026-08-28)

Gage, 28 Aug 2026, via Leo. ~30% up / ~25% down over 18–24 months. Watch. Gage's judgment, not grain / not issuer fact. Last-price as-of stays the vault last: $92.09 (Thu 27 Aug 2026 4:00 PM EDT). NightDesk Friday morning $90.16 (NightDesk tape, not the last-price). External foundry ~$293M and ~$2.1B foundry op-loss at Q2 — already on this page from 2026-07-23-earnings-intc-q2-fy2026; Gage's watchpoints, not new grain. Graduation = a named outside 18A/14A order in a filing or call (already the open test / late tell on this page — not a reason to wait to start; do not invent a new customer). Not a buy/sell/size. No source URL — do not invent one.

Paul (via Leo, 28 Aug): a named outside 18A/14A order is now the late tell, not the buy trigger. Buy the bottleneck before that print; smaller size when early; re-evaluate at the checkpoint. Filed watch still stands.

Paul confirmed (28 Aug, via Leo): small paper INTC ~3% after Werner, after 9:35. After-Werner / after-9:35 is timing, not a vault last-price. Not issuer grain. Not Alfred's ticket. Not a live ticket.

Checkpoints (dated)

From 2026-08-28-intel-foundry-next-18a-14a-outside-order-checkpoints (28 Aug 2026). Still no named outside 18A/14A order. Q2 / VLSI / last-price $92.09 / Gage ~30/~25 Watch / NightDesk $90.16 on this page are not rewritten.

  1. October 2026 — 14A PDK 0.9 (and Tan's next update). Lip-Bu Tan: “PDK 0.9 is on track for October.” Dave Zinsner: Intel Foundry “met critical milestones towards delivering the 0.9 PDK for Intel 14A in October.” Tan: “I look forward to … providing another update in October.” Intel IR still lists Q2 2026 as the latest results. No official Q3 report-date notice. Do not invent a calendar day for Q3.
  2. H2 2026 — first half of the 14A customer-decision window. Already on this page (two unnamed prospects; decisions H2 2026 → H1 2027). The Q2 10-Q does not name those prospects and does not move the window.
  3. H1 2027 — remainder of the same decision window. Internal 14A risk production is H2 2027; HVM commit 2028 — execution, not the outside-order test.

IEDM 2026 (12–16 Dec, San Francisco) is not a checkpoint unless Intel ties it to a named customer process. Official program as of 2026-09-08 still lists only Keynote (3 Monday-morning presentations), Tutorials (6 × 80 min), and Short Courses (two full-day) landing stubs — no paper titles, no authors. From 2026-09-08-conference-iedm-2026-program.

Confirm: a named outside 18A or 14A order in an Intel filing or earnings call. Honest none-found since 23 Jul 2026. Socionext (18 Aug 2026) “will leverage Intel 18A-P process technology in the development of custom SoCs” is a design-house development note, not an Intel 8-K/call wafer order. Leave it off the graduation line. Do not mint a Socionext entity.

Break:

  • PDK 0.9 misses October, or Tan’s October update walks the on-track language.
  • H2 2026 closes with still-unnamed 14A prospects, or the window slips.
  • H1 2027 closes without a named outside 18A/14A order.
  • Q3 / October call still names no customer.

A Socionext-style ecosystem note, an internal 14A HVM commit, or another equity raise does not confirm or break the chain.

Old-world pricing: Yes. Vault last $92.09 (27 Aug 2026 close, already on this page). External foundry $293M vs Foundry $5.8B and ~$2.1B op-loss. 10-Q still “higher-cost” 18A wafers. August $20B offering is capital, not a foundry-order print.

Key facts

  • Ticker: INTC; ~240% YTD as of May 21, 2026 (ATH $129.44 on May 11)
  • INTC price trajectory: ATH $129.44 (May 11 closing); pulled back to ~$115 as of May 18 (~10% correction); 52-week high $132.75; market cap ~$582B; ~200% YTD. Stock trading 14–35% above BofA's raised PT of $96 throughout May — market pricing in more than BofA base case.
  • BofA analysis (post-Apple deal announcement): raised PT to $96 (from $56); models Apple deal as adding $10B/year in Intel Foundry revenue by 2030 — stock has been trading 14–35% above this target throughout May. Trefis (May 13) suggests "Intel Foundry $1 trillion upside" as a Street outlier thesis.
  • Q1 2026 earnings (April 23): Revenue $13.6B vs $12.32B consensus (beat by $1.4B — 6th consecutive quarterly beat); Non-GAAP gross margin 41% (+650bps vs guide). Foundry revenue $5.4B (+20% QoQ). AI revenue = 60% of total, growing 40% YoY. Advanced packaging backlog escalated from "hundreds of millions" → "billions of dollars per year". Core Ultra Series 3 and 18A-based processors achieved fastest new product ramp in 5 years. DCAI $5.1B (+22% YoY); ASIC revenue more than doubled YoY. Stock +24% post-earnings. Fab 34 (Irish facility): completed $7.7B buyout of 49% equity interest. Q2 guide: $13.8–14.8B. From 2026-04-23-earnings-intc-q1-2026. (Prior autoresearch coverage from 2026-05-21-autoresearch-apple-intel-deal-may-21-2026 corroborated by primary source.)
  • lip-bu-tan in 2026-04-23-earnings-intc-q1-2026: "Long-term agreements signed — including with Google, typically spanning 3-5 years." This is the primary-source confirmation that Google is a long-term anchor customer (3-5yr LTA) for server CPUs and ASICs — not just an evaluation.
  • Tesla: confirmed first external Intel 14A customer (Tweaktown/Alpha-sense). Reduces Apple-or-nothing concentration risk for 14A. From 2026-05-21-autoresearch-apple-intel-deal-may-21-2026.
  • Named anchor customers: Google (3-5yr LTA, server CPUs and ASICs — confirmed Q1 2026 primary source); AWS multi-billion-dollar 18A custom AI fabric chips; Microsoft Maia 2 on 18A/18A-P
  • Apple: M7 on Intel 18A-P (15-20M units/yr, production 2027, MacBook Air + iPad Pro); A21 iPhone chip on Intel 14A (~60M units/yr, mass production 2028, ~20% of total iPhone production); A21 represents ~80% of the deal's order value. From 2026-05-21-autoresearch-apple-intel-deal-may-21-2026. Apple described as "secondary foundry" with 2026 small-scale testing → 2027 ramp → 2028 continued growth → 2029 decline — supply diversification, not TSMC defection. TSMC/Bernstein: Apple Intel volumes "too small to matter" for TSMC (25% Apple revenue concentration at TSMC; 20% iPhone shift to 14A reduces TSMC's Apple share proportionally but TSMC fills from other AI/HPC customers).
  • Intel 18A HVM confirmed: January 30, 2026 — Fab 52, Chandler AZ, ~40K wafers/month. Process differentiators: RibbonFET (GAA transistor) + PowerVia (first backside power delivery in volume production). Internal yield gate passed (Panther Lake went to volume production at 18A). From 2026-05-21-autoresearch-apple-intel-deal-may-21-2026.
  • Terafab: Musk's SpaceX/Tesla/xAI consortium; Intel as foundry partner; $25B confirmed (24/7 Wall St., April 7, 2026) — earlier $55B/$119B figures may represent different phases
  • Tesla $3B Austin Intel-14A R&D fab (separate from Terafab): Tesla building a standalone $3B semiconductor R&D fab in Austin using Intel 14A, producing "a few thousand wafers per month" at R&D/pilot scale. Per Q1 Apr 22 2026 Tesla earnings. This is a distinct investment from Terafab ($25B JV) — Tesla's own fab for chip R&D, not the consortium-scale production facility.
  • 18A yields (June 3 pre-symposium claim, not in the official papers): A June 3 synthesis said VLSI 2026 Honolulu would confirm yield >60% improving at 7-8%/month. The symposium ran June 14–18. Official T1.2 abstract + June 16 newsroom publish no die-yield %. Treat >60% as a pre-symposium / other-channel claim, not a VLSI paper result. Flagged, not silently overwritten. Industry-standard still deferred to 2027. External foundry ~$174M of $5.4B (Q1). Per 2026-06-03-intel-18a-vlsi-2026-update-yield-data-customer-wins-mass-production vs 2026-08-21-autoresearch-intel-18a-vlsi-yield-followup.
  • 18A-P VLSI 2026 (what the papers actually published): Official T1.2 + Intel June 16 newsroom: 18A-P is design-compatible with 18A; >9% iso-power or >18% lower power at iso-performance; Power Boost; 20–40% thermal resistance; 10–30% via resistance; risk production. Karl invited talk: 11% routed-area / 10× droop / up to 6% freq or >15% dynamic power vs frontside. Shamanna ~30% frequency at ~0.5 V. Tan’s July 23 verbal (“around 5%” vs 18A) does not match the 9%/18% paper — do not average. The June 3 page’s 50% thermal conductivity is not the newsroom’s 20–40% thermal-resistance figure. TSMC T1.5 (same symposium): A16 vs N2P 8–10% speed or 15–20% power + 8–10% density; MP slated Q4’26 — same-family PPA, not an 18A-P vs A16 yield bake-off. Per 2026-08-21-autoresearch-intel-18a-vlsi-yield-followup.
  • 18A vs TSMC yield after VLSI + Q2: Q2 primaries (release, call, 10-Q) still give no die-yield %. Output ~25% above target and >50% QoQ; PTL SKU cost −50% YTD; yields “ahead of expectations.” The June 27 10-Q says Foundry profit is hurt by “higher-cost wafers manufactured on our Intel 18A process node.” KeyBanc (John Vinh, ~July 15) models 18A ~85% (from ~65%) vs TSMC N2 ~90% — same firm as the mid-2025 ~55% intel-18a-yield-vs-tsmc already distrusts, not a teardown. ASML July 15: High NA on a subset of Panther Lake 18A layers; “yields matched to the NXE platform” — tool-to-tool layer match, not 18A vs N2 die-yield. Per 2026-08-21-autoresearch-intel-18a-vlsi-yield-followup.
  • Apple confirmed (KeyBanc primary): 18A for low-end M-series (MacBooks, iPads) production targeted 2027; 14A under discussion for low-end A-series iPhone (2029). Not M7/high-end. 14A customer decisions expected H2 2026. Per 2026-06-03-intel-18a-vlsi-2026-update-yield-data-customer-wins-mass-production.
  • Intel 18A vs TSMC N2: 280 MTr/mm² (VLSI 2026) vs TSMC N2's 313 MTr/mm² — gap narrower than prior 24% differential. Per 2026-06-03-intel-18a-vlsi-2026-update-yield-data-customer-wins-mass-production.
  • Google EMIB win (new May 2026): Google selecting Intel EMIB advanced packaging for TPUv8e (H2 2027), directly substituting TSMC CoWoS; EMIB yields ~90% (TrendForce May 4); Meta also exploring Intel EMIB — packaging moat erosion at TSMC
  • High-NA EUV: Expanded from 1 to 2 units (only customer using High-NA at production scale in 2026, alongside Samsung 2 and SK Hynix 2); TSMC taking 0 High-NA in 2026
  • Equipment orders: Boosted 50%+ YoY in 2026 (TrendForce, April 20, 2026) — leading indicator of capacity commitment
  • 14A: May draw major customers (Google, Apple, AMD, Nvidia) by year-end 2026 per TrendForce
  • Advanced packaging: Targeting Microsoft, Tesla, Qualcomm, Nvidia as packaging customers (EMIB, Foveros)
  • Tan's stated threshold: "If Intel does not secure an anchor customer by late 2026, the likely outcome is a slower path forward" — makes Apple deal a make-or-break event
  • Nvidia clarification: The $5B Nvidia "commitment" is an equity stake + co-design arrangement (NOT a wafer production order); Nvidia reportedly tested 18A but stopped moving forward on foundry production

Strengths (from a thesis-input perspective)

  • Spare leading-edge capacity at the moment TSMC is sold out
  • US footprint matches tariff/CHIPS-era policy direction; largest CHIPS Act recipient
  • Equipment orders +50%+ YoY confirm genuine capex commitment
  • First-mover on High-NA EUV — Intel leads TSMC in High-NA experience at 2nm-class nodes
  • Political durability: White House active broker; government financial stake aligns incentives
  • 14A PDK approaching customer-commitment readiness by year-end 2026
  • Preliminary Apple agreement gives Intel first-mover advantage over Samsung in Apple evaluation

Weaknesses (from a thesis-input perspective)

  • Intel 18A is faster but less dense than TSMC N2 — a real penalty for Apple's SRAM-heavy SoC designs
  • Apple deal is preliminary — no orders, no products specified; "industry-standard" yields not until 2027
  • Samsung now a credible second alternative — Apple may dual-source instead of Intel-exclusive
  • Nvidia $5B is equity, not a production order — prior "Nvidia validates 18A" framing was overstated
  • Capex burden is enormous; equity dilution risk
  • Tan's self-identified anchor-customer threshold adds execution urgency

Legal overhang

  • VLSI v. Intel CAFC reversal (April 14, 2026): Federal Circuit reversed two of Intel's summary judgment wins in No. 24-1772: (1) extraterritoriality — parties had stipulated 70% of Intel's activities had US nexus, but CAFC found that stipulation does NOT constitute a noninfringement admission; (2) doctrine of equivalents — for claims 10, 11, 13, and 17 of U.S. Patent No. 8,566,836 ("Multi-core System on Chip"), CAFC found lower court wrongly imported prosecution disclaimer limitations. Dr. Sullivan's NPV/VPU damages theories stricken; Mr. Chandler's royalty opinions retained on remand.
  • June 2026 remand trial (Northern District of California): Prior VLSI verdicts against Intel exceed $3 billion in total, including a $2.18 billion jury verdict + $162 million in prejudgment interest. These set the damages ceiling expectation for the remand proceeding on the '836 patent.
  • Intel exposure: Headline risk if the June trial yields a multi-hundred-million verdict. Not existential given Intel's scale, but adds to legal overhang alongside ongoing PTAB proceedings.
  • From 2026-05-30-autoresearch-regulatory-antitrust-tech-biotech-utilities: "Intel (INTC) — VLSI remand trial June 2026; Federal Circuit reinstated infringement claims — Negative (damages exposure)"

Capital structure / financing (June 2026)

From 2026-06-11-feed-semianalysis-intel-should-raise-capital (SemiAnalysis — Doug O'Laughlin / Sravan Kundojjala / Dylan Patel):

  • SemiAnalysis argues Intel should issue ~$25B of equity ("a reverse buyback," ~4–5% dilution against a ~$498B market cap) into the hot equity window to fund the buildout, rather than carry more debt or sell more fab economics.
  • The Apollo Fab 34 unwind is the tell: Intel agreed (Mar 31, 2026) to buy back Apollo's 49% of Fab 34, closing Apr 8 for $14.2B (~$7.7B cash + $6.5B bridge), called "accretive." SemiAnalysis: "If buying a fab stake back is accretive, then selling fab economics to a partner was expensive money all along" — i.e. SCIPs (Apollo/Brookfield/Silver Lake-Altera) were costly capital, so equity is now "the cheapest capital in the building."
  • Balance sheet: ~$45.0B debt, ~$51.5B once the Apollo bridge is in; Mobileye and Altera (51% to Silver Lake at $8.75B EV) largely harvested — so equity is the remaining lever.
  • Sovereign anchor as a price floor: US government took up to 433M shares at $20.47 (9.9% stake), SoftBank at $23.00, Nvidia at $23.28 — all above water; a raise at today's price lifts book value per share and the 10% sovereign stake "holds the floor," letting a large offering clear cheaply.
  • The order book that funds it: Terafab (SpaceX/Tesla) is the 14A anchor (initial 100k WSPM target scaling toward 1M); plus Nvidia DGX Rubin NVL8 dual Xeon 6, a Google multiyear Xeon + custom-IPU deal, and SambaNova on inference. Full multi-phase Terafab project costed at up to $119B, with SpaceX putting up initial capital and Intel contributing meaningfully. New technically-literate board (Lip Bu Tan CEO, ex-Qualcomm chair, Eric Meurice ex-ASML, Steve Sanghi).
  • Read-through: a large INTC equity raise is dilutive near-term but de-risks the foundry buildout the us-fab-capacity-bottleneck / taiwan-chokepoint-to-allied-reshoring / tsmc-saturation-to-intel-anchor-stack theses depend on; the signed order book is what lets it price against demand, not a turnaround promise.

Open questions

Sources

Related

Referenced by
Mechanisms
Jan 2026 tariffs + Taiwan trade deal + CHIPS Act expansion → 5pp cost-incentive shift → US-domiciled fab beneficiary stackSamsung SF2P yields hit 70% + Taylor 90% ready + Apple visit → genuine third foundry leg → order flow splits, not concentratesTaiwan chokepoint → frontier-AI supply at risk → allied-reshoring imperativeTaiwan energy blockade → island browns out in <3 weeks → fabs power down → 90-day restart → global shock, no shot firedTSMC packaging scarcity caps customer growth → strands TSMC's own (larger) front-end wafer revenue → incumbent welcomes a second-source packager → Intel EMIB-T demand is validated by the monopolist itselfTSMC saturation → 4-year price hikes → marginal buyers shift → Intel anchor stack → re-rate
Concepts
Picks-and-shovels: leading-edge fab buildoutTSMC capacity shortfall and 4-year pricing powerUS fab capacity bottleneck → Intel as the alternative foundry
Questions
Does Samsung have a credible leading-edge foundry comeback that would weaken the Intel-as-alternative thesis?Has Intel 18A reached yield parity with TSMC, or are partnerships outpacing the technical reality?Hypothesis: Section 232 Phase 2 tariffs → US chip fab structural premium pricingWill the Apple-Intel deal convert to actual orders, on which products, on what timeline?
Entities
AppleApplied MaterialsASMLSamsung FoundryTerafabTeslaTSMC (Taiwan Semiconductor Manufacturing Co.)