Intel reported third-quarter revenue of $16.5 billion on 29 October, up 21% year on year, with a GAAP gross margin of 43.0% and GAAP earnings of $0.60 per share.
The segments moved in different directions. Data Center and AI grew 78% to $7.3 billion and Intel Foundry 58% to $6.7 billion, while the client computing business fell 5% to $8.1 billion. The foundry still ran a segment operating loss of $1.85 billion in the quarter. Guidance for the fourth quarter is $15.0–16.0 billion at a GAAP gross margin of 37.1%, which implies the margin pressure is not over.
A foundry losing almost two billion dollars a quarter while growing by more than half is exactly what the early years of a foundry look like. Capacity is paid for before it is sold, and external customers arrive only after a process has been proven by somebody else’s product. The company said it released the 0.9 version of the Intel 14A process design kit in October, which is the step that lets external design teams start real work on the node.
For anyone outside the company, the question a PDK release answers is narrow but important: whether a second leading-edge foundry will exist by the end of the decade. On current numbers, that still depends on customers who have not yet committed publicly.
The client decline deserves a note of its own. A 5% fall in personal-computer silicon while the rest of the industry grows at double digits is partly memory prices feeding through to retail prices, and partly budgets moving from refreshing laptops to buying accelerators. Both pressures are outside Intel’s control.
Source: Intel