They slashed the dividend to build factories for customers that haven't signed up yet, while gross margins collapsed to 38%.
Thesis: Intel is currently a foundry customer masquerading as a competitor. They are outsourcing their best products (Lunar Lake) to TSMC, paying someone else's margin, while burning capital to build capacity for a node (18A) that has no volume production. The 38% gross margin is not a cyclical low; it is the structural cost of uncompetitive yields and double-paying for capacity. Gaudi 3 competing on price ('two-thirds the cost') confirms they are a price taker in the only market that matters.
Verdict: SHORT — Conviction: HIGH
Catalyst: Panther Lake ramp in 2H 2025. This is the first real test of 'bringing tiles home' to 18A. If yields miss or volume delays occur, the 2026 margin recovery thesis evaporates.
Key Risk: Political intervention or a massive subsidy injection (CHIPS Act acceleration) could distort the bankruptcy risk pricing, squeezing shorts despite the fundamental deterioration.
The Tell: CFO Zinsner admitted the move to the Ireland fab caused a 'meaningful chunk' of the 400 basis point margin miss but defended it because 'it saves capital.' They are actively destroying unit economics to preserve liquidity. That is a distress signal, not an optimization.
Friction Level: HIGH_FRICTION — Bulls see a manufacturing turnaround de-risked by 18A milestones; Bears see structural margin collapse and a cash crunch disguised as discipline.
Report not found
The report data is no longer available. Please return to the archive.