They claim process leadership while buying their flagship compute tiles from TSMC.
Thesis: Intel is a construction project disguised as a chip company. The '5 nodes in 4 years' sprint is crushing cash flow ($2B depreciation increase + peak startup costs), while they are forced to outsource their best product (Panther Lake) to TSMC to remain competitive. The 'AI Everywhere' narrative is a defensive pivot because they missed the training cycle. Until 18A proves yield AND external customer volume, this is a capital trap.
Verdict: AVOID — Conviction: MEDIUM
Catalyst: Feb 21st IFS Direct Connect event – disclosure of 18A customer names beyond the current 'lifetime deal value' abstraction.
Key Risk: 18A yield or performance miss. If 18A isn't leadership, the IDM 2.0 thesis collapses into a sub-scale foundry with high fixed costs.
The Tell: Pat Gelsinger claiming 'Process Leadership' while confirming Panther Lake, the 2025 client flagship, uses external foundry for compute tiles. You don't outsource the crown jewels if your own vault is secure.
Friction Level: HIGH_FRICTION — The 60% flow-through rule vs. peak startup costs. Bull believes the operating leverage math; Bear sees the structural cost of '5 nodes in 4 years' plus TSMC outsourcing fees breaking the model.
Report not found
The report data is no longer available. Please return to the archive.