Intel reported second-quarter revenue of $16.1 billion, up 25 percent, alongside an $11 billion attributable GAAP loss. It also reported $2.2 billion of non-GAAP income and $7 billion in operating cash. [1] Growth and red ink belong to the same quarter; neither headline explains the bridge between them.
The quarter tests the April 24 article that classified Intel's prior result as an operating beat. Q2 supplies stronger revenue and segment growth, but it reopens the questions that optimistic account left for later receipts: accounting reconciliation, durable demand, external foundry customers, capacity and return.
Data Center and AI revenue reached $6.3 billion, up 59 percent. Intel Foundry reported $5.8 billion, up 31 percent, while Intel Products reported $15.1 billion. [1] Those segment totals cannot be added to create a company total because they include transactions within Intel. The release records $5.5 billion of intersegment eliminations.
Foundry's $5.8 billion is therefore not external foundry revenue. [1] Internal sales can show factories serving Intel's product businesses, but they do not prove that outside customers have committed leading-edge volume. The accounting boundary determines whether a growth figure demonstrates organizational activity or external market conversion.
CNBC reported that Intel had 10 long-term server-CPU agreements and strong customer demand, while disclosing no major customer for its leading-edge external foundry. [2] CPU agreements and external wafer customers are different receipts. The former may support product demand; they do not fill the latter's empty column.
The GAAP and non-GAAP results require equally careful separation. The source stack attributes much of the gap to a mark-to-market item, but it does not authorize assigning the entire $11 billion loss to one adjustment without the exact reconciliation. [1] [2] Readers need the amount, instrument, tax treatment and other excluded items before deciding what is operating and what is accounting.
Cash from operations offers another completed measure, not a universal verdict. [1] Seven billion dollars of operating cash can coexist with investment needs, supply constraints and an accounting loss. Free cash flow, capital expenditure, equipment, clean-room space and substrates determine how much capacity can be built and at what cost.
Year-over-year growth also needs a stable company perimeter. Changes such as Altera's deconsolidation can alter comparison without changing demand in every remaining business. The quarter's figures are reportable as published, but durable analysis needs a bridge showing which operations remained comparable across both periods.
Future-node claims remain future. Progress on 14A, planned capacity and Q3 guidance are management plans or forecasts until yields, equipment, customer commitments, shipments, revenue and margin arrive. [1] The prior article's confidence about external validation cannot be carried forward as though a major leading-edge customer had now been disclosed.
No usable X post was verified for the quarter, so platform celebration or panic remains unobserved. CNBC's fastest-growth framing and an $11 billion-loss framing each isolate a real number. [2] The useful account holds both while asking which sales are internal, which customers are external and which adjustments reconcile the earnings measures.
Intel has delivered a stronger top line and a more complicated proof burden. [1] Durable turnaround requires repeat revenue, external foundry conversion, usable capacity, margins and return. Collapse would require more than one GAAP figure whose bridge is still being parsed. Q2 is a completed quarter, not a completed verdict.
-- THEO KAPLAN, San Francisco