Verdict
This is a bounded failure of the 2016 process-and-product-cadence underwriting thesis and of later manufacturing-capital economics. Intel subsequently disclosed a material process delay and FY2024 consolidated and Foundry stress; FY2025 Foundry reported revenue was overwhelmingly intersegment, and Intel warned that future leading-edge-node economics require significant external volume. The conclusion is deliberately narrower than “Intel failed”: FY2025 Intel Products was profitable, Intel reported an 18A product milestone, and selected raw reported consolidated measures improved from FY2024. The record does not establish insolvency, terminal failure, failure of every business, or the final outcome of Intel Foundry or 14A. [claim.intel.outcome.bounded-failure] [claim.intel.outcome.causal-boundary]
No actual investor action is observed. Frozen Part A was a simulation that selected defer and re-underwrite: initiate no new exposure, add no capital, and route any existing holding to an authorized human committee. No actual holding, mandate, order, execution, price path, or evidence that a real committee followed those gates is present. [judgment.intel.cutoff.defer-and-reunderwrite] [claim.intel.outcome.valuation-return-abstention]
The later roadmap and customer record
Apple announced on June 22, 2020 that it would transition the Mac to its own silicon, expected the first Apple-silicon Mac by year-end, and planned to complete the transition in about two years. That primary record establishes Apple's architecture strategy and announced timing. It does not disclose Intel-specific lost economics or prove that an Intel process delay caused Apple's decision. source · apple.outcome.mac-silicon-transition [evidence.intel.outcome.apple-transition-strategy] [evidence.intel.outcome.apple-transition-timeline] [claim.intel.outcome.apple-transition]
Intel's Q2 2020 Form 10-Q then said its 7nm CPU-product timing was delayed approximately six months from prior expectations and that 7nm yield was trending approximately twelve months behind the company's internal target. Intel also described greater use of external-foundry technology as a contingency. This is direct later evidence that one important process-cadence risk materialized; it is not a controlled bridge from the 2015-2016 10nm evidence to every subsequent node or financial outcome. source · intel.outcome.q2-2020-10q [evidence.intel.outcome.process-delay-2020] [claim.intel.outcome.2020-process-delay]
The process labels must stay distinct. The cutoff record concerned 10nm and Kaby Lake; the 2020 filing concerned 7nm; the FY2025 filing concerned 18A products and 14A risk. They are not one comparable yield, performance, or schedule series. [claim.intel.cutoff.process-delay] [claim.intel.outcome.18a-counterevidence] [claim.intel.outcome.14a-capital-risk]
FY2024: the stress bridge
Intel's FY2024 Form 10-K reported USD 53.101 billion of revenue, USD 17.345 billion of gross profit, a USD 11.678 billion operating loss, a USD 19.233 billion consolidated net loss, and a USD 18.756 billion net loss attributable to Intel. Deterministic arithmetic gives a 32.664168 percent gross margin and a negative 21.992053 percent operating margin. These are consolidated reported outcomes, not a valuation or causal decomposition. source · intel.outcome.fy2024-10k [evidence.intel.outcome.fy2024-consolidated-results] [claim.intel.outcome.fy2024-stress] [table.intel.outcome.consolidated-stress-endpoint] [table.intel.outcome.deterministic-endpoint-checks] [model-run.intel.outcome.endpoint-checks]
The original FY2024 filing's Intel Foundry presentation reported USD 17.543 billion of segment revenue, a USD 13.408 billion operating loss, and USD 385 million of external revenue. Intel stated that the segment included intersegment financial activity. The severity of the loss supports the manufacturing-economics concern, but segment totals cannot be treated as external sales or added to consolidated results. [evidence.intel.outcome.fy2024-foundry-results] [claim.intel.outcome.fy2024-foundry-stress]
The later FY2025 filing presented its FY2024 Foundry comparison differently: USD 17.317 billion of revenue, a USD 13.291 billion operating loss, and USD 159 million of external revenue. The archive preserves the original and later comparative facts side by side, with their different knowledge times. It does not silently overwrite one presentation or splice them into a trend. source · intel.outcome.fy2025-10k [evidence.intel.outcome.fy2025-foundry-results] [claim.intel.outcome.foundry-presentation-change] [table.intel.outcome.foundry-presentation-boundary]
FY2024 investing cash PP&E additions were USD 23.944 billion, compared with USD 8.288 billion of operating cash flow. The resulting negative USD 15.656 billion residual is only an operating-cash-flow-minus-investing-cash-PP&E scale check. The same record separately shows government-incentive proceeds, partner contributions, financing-activity PP&E payments, and noncash PP&E; those classifications must not be added into a synthetic capex number. [claim.intel.outcome.fy2024-capital-perimeters] [claim.intel.outcome.descriptive-cash-check] [table.intel.outcome.deterministic-endpoint-checks]
FY2025: improvement in raw reported results, not a normalized recovery claim
Intel's FY2025 Form 10-K reported USD 52.853 billion of revenue, USD 18.375 billion of gross profit, a USD 2.214 billion operating loss, USD 1.557 billion of pretax income, USD 26 million of consolidated net income, USD 9.697 billion of operating cash flow, and USD 14.646 billion of investing cash PP&E additions. Deterministic margins were 34.766238 percent gross and negative 4.188977 percent operating. source · intel.outcome.fy2025-10k [evidence.intel.outcome.fy2025-consolidated-results] [evidence.intel.outcome.fy2025-capital-cash-flows] [claim.intel.outcome.fy2025-consolidated] [table.intel.outcome.consolidated-stress-endpoint]
Against the selected FY2024 reported inputs, revenue changed negative 0.467035 percent, gross profit changed positive 5.938311 percent, the reported operating loss narrowed by USD 9.464 billion, and investing cash PP&E additions changed negative 38.832275 percent. Every one of these is explicitly a raw reported cross-period or cross-perimeter check. None is normalized for Altera deconsolidation, restructuring, impairments, organizational or segment-presentation changes, noncontrolling interests, or changes in required investment. [table.intel.outcome.year-over-year-checks] [model-run.intel.outcome.endpoint-checks] [claim.intel.outcome.altera-perimeter] [claim.intel.outcome.foundry-presentation-change] [claim.intel.outcome.causal-boundary]
Consolidated and Intel-attributable income also differ. FY2025 consolidated net income was USD 26 million, net income attributable to noncontrolling interests was USD 293 million, and net loss attributable to Intel was USD 267 million. Calling the consolidated result “Intel's profit” without the attribution bridge would erase an economically material perimeter. [claim.intel.outcome.income-perimeters] [table.intel.outcome.income-altera-perimeters]
Products and Foundry tell different stories
FY2025 Intel Products reported USD 49.147 billion of segment revenue and USD 12.739 billion of operating income, a deterministic operating margin of 25.920199 percent. Intel Foundry reported USD 17.826 billion of segment revenue and a USD 10.318 billion operating loss, a negative 57.881746 percent deterministic operating margin. Both segments include intersegment activity and do not sum to consolidated revenue or operating income. The coexistence of a profitable Products segment and a deeply loss-making Foundry segment is why the case rejects both an all-business failure label and an undifferentiated company-quality score. [evidence.intel.outcome.fy2025-intel-products] [evidence.intel.outcome.fy2025-foundry-results] [claim.intel.outcome.products-foundry-divergence] [table.intel.outcome.segment-economics]
Intel separately reported only USD 307 million of FY2025 Foundry external revenue. Subtracting that from USD 17.826 billion of total segment revenue gives a USD 17.519 billion intersegment residual; external revenue was 1.722204 percent and the residual 98.277796 percent of the segment total. This arithmetic does not validate transfer prices, eliminations, customer count, customer concentration, qualification, commitments, or external-customer economics. It supports a demand-perimeter warning, not a claim that there were “few customers.” [claim.intel.outcome.foundry-revenue-perimeter] [table.intel.outcome.foundry-revenue-perimeter] [model-run.intel.outcome.endpoint-checks]
The strongest technical counterevidence is also in the FY2025 filing: Intel reported releasing its first Core Ultra Series 3 processors manufactured on Intel 18A and described 18A features as industry firsts in high-volume semiconductor manufacturing. That is evidence of an issuer-reported product and manufacturing milestone. It is not independent yield, defect, reliability, customer-volume, or economic validation. [evidence.intel.outcome.fy2025-process-milestone] [claim.intel.outcome.18a-counterevidence]
The same filing disclosed a forward capital-risk boundary. Intel said leading-edge-node economics require volume beyond its own products and warned that it may pause or discontinue 14A and successor nodes if it cannot secure a significant external foundry customer. That is an issuer risk statement, not proof of final failure; it keeps the external-customer and capital-return gate open. [evidence.intel.outcome.fy2025-14a-capital-risk] [claim.intel.outcome.14a-capital-risk] [table.intel.outcome.underwriting-unknowns]
Altera, income, capital, and funding perimeters
Intel completed the sale of 51 percent of Altera on September 12, 2025, deconsolidated the business, retained a 49 percent equity-method interest, and recorded a USD 5.553 billion pretax gain that included approximately USD 2.1 billion from remeasuring the retained interest. The gain exceeded FY2025 consolidated pretax income of USD 1.557 billion by USD 3.996 billion. That subtraction is an algebraic scale check only—not a no-gain pro forma, an accounting adjustment, recurring earnings, or an acquisition-return calculation. Related costs, taxes, financing, operating-perimeter changes, and other effects are not reversed. [evidence.intel.outcome.fy2025-altera-deconsolidation] [claim.intel.outcome.altera-perimeter] [claim.intel.outcome.altera-gain-context] [table.intel.outcome.income-altera-perimeters]
FY2025 capital and funding disclosures must remain a set of distinct measures:
| Reported FY2025 measure | USD billions | Perimeter |
|---|---|---|
| Operating cash flow | 9.697 | Operating cash flow statement |
| Cash PP&E additions | 14.646 | Investing outflow |
| Capital-related government-incentive proceeds | 1.577 | Investing inflow |
| Net divestiture proceeds | 6.157 | Investing inflow |
| Partner contributions | 5.108 | Financing inflow |
| PP&E payments in financing activities | 3.026 | Financing outflow |
| Noncash PP&E acquisitions | 4.952 | Supplemental noncash disclosure |
| Equity-, warrant-, and escrow-share-related proceeds | 12.706 | Financing inflow and ownership perimeter |
These amounts are neither interchangeable nor safely additive. In particular, USD 14.646 billion, USD 3.026 billion, and USD 4.952 billion must not be summed into a synthetic “capex” figure: they differ by cash classification and timing, and the selected record lacks a complete nonoverlapping asset-and-payment-vintage reconciliation. [evidence.intel.outcome.fy2025-capital-cash-flows] [evidence.intel.outcome.fy2025-vendor-financing] [evidence.intel.outcome.fy2025-government-incentives] [evidence.intel.outcome.fy2025-equity-warrant-funding] [claim.intel.outcome.fy2025-capital-funding-perimeters] [table.intel.outcome.capital-funding-perimeters]
Operating cash flow less investing cash PP&E additions was negative USD 4.949 billion in FY2025, again only a descriptive scale check. Intel separately reported issuer-defined non-GAAP net capital expenditures of USD 11.204 billion and adjusted free cash flow of negative USD 1.612 billion. The issuer reconciliation is preserved as issuer-adjusted; it is not substituted for GAAP cash flows or for the archive's descriptive residual. [evidence.intel.outcome.fy2025-net-capex-reconciliation] [claim.intel.outcome.net-capex-nongaap] [claim.intel.outcome.descriptive-cash-check] [table.intel.outcome.issuer-adjusted-cash-measures]
Causal assessment
The primary hypothesis is cadence-capital interaction: Intel's integrated process-and-product model plausibly transmitted roadmap misses into product timing, fixed-cost absorption, and later manufacturing-capital economics. The 2020 delay, FY2024 stress, FY2025 Products/Foundry divergence, low external revenue relative to segment total, and 14A external-volume contingency fit that mechanism. Confidence is moderate because the record is observational and because Products profitability, the reported 18A milestone, and the improved raw consolidated endpoint are material counterevidence. [hypothesis.intel.cadence-capital-interaction]
The first rival gives more weight to customer architecture and competition. Apple's ecosystem strategy, competitive product cycles, and end-market conditions could explain a large unknown share of later customer and product outcomes independently of process execution. The record does not contain customer-level Intel economics or a controlled reason for Apple's choice. [hypothesis.intel.customer-architecture-and-competition] [assumption.intel.outcome.customer-economics-unobserved]
The second rival gives more weight to accounting, restructuring, and perimeter changes. FY2024 presentation changes, impairments and restructuring, Altera deconsolidation, noncontrolling interests, funding transactions, and issuer adjustments explain part of the reported cross-period movement. They do not prove that the underlying technical and capital concerns were immaterial. [hypothesis.intel.accounting-restructuring-and-perimeter] [assumption.intel.outcome.segment-presentations-distinct]
The selected evidence cannot assign causal percentages among these mechanisms. It also cannot establish that one 2015-2016 management or committee decision caused all later results. [claim.intel.outcome.causal-boundary]
What should guide an underwriting agent
Three rules graduate only as candidate rules from this case:
-
Require a versioned process-to-product evidence chain across independent technical validation, roadmap timing, qualification, customer deployment, segment economics, and funding before a human approves new capital. A node or “high-volume” label alone does not clear the chain. [rule.require-semiconductor-roadmap-evidence-chain]
-
Underwrite Products, Foundry, external-foundry demand, intersegment pricing, eliminations, and presentation vintages as separate ledgers. Do not infer an externally validated foundry business from total segment revenue dominated by intersegment activity. [rule.separate-products-foundry-and-intersegment-economics]
-
Reconcile investing cash PP&E, financing-activity PP&E, noncash PP&E, incentives, partner funding, equity-linked proceeds, divestitures, ownership, and issuer non-GAAP measures in one nonoverlapping sources-and-uses schedule. Until a human approves that schedule and its downside return gates, abstain from capital-return, valuation, and exposure changes. [rule.reconcile-semiconductor-capital-and-funding-perimeters]
The feasible counterfactual is the frozen Part A process: defer new exposure and require human re-underwriting as the six gates clear. Its financial effect is not estimated because actual trades, prices, capitalization, portfolio constraints, approved forecasts, and counterfactual cash flows are absent. [counterfactual.intel.follow-part-a-defer-and-reunderwrite]
Abstention and hindsight limits
The deterministic run is reproducible descriptive arithmetic, but it has no frozen model-registry validation receipt and therefore remains non-authoritative with validation_status: not_run. It performs no valuation and no accounting adjustment. [model-run.intel.outcome.endpoint-checks]
The outcome record ends with FY2025 on December 27, 2025 and selected evidence public by January 22, 2026, inside the declared February 3, 2026 outcome cutoff. It excludes later 14A customers, products, filings, financials, and Foundry outcomes. The evidence set also samples 2020, FY2024, and FY2025 rather than every intervening quarter, node, product generation, capital release, and customer decision.
No target price, price-attractiveness conclusion, position size, avoided loss, or total shareholder return is produced. Business quality, underwriting readiness, evidence confidence, and security attractiveness remain separate, and any accounting adjustment, investment action, ethics conclusion, publication, or external action requires explicit human approval. [claim.intel.outcome.valuation-return-abstention]