Shipments rose 14.4% but utilization moved only 0.6 points. They added capacity just to keep up, and Q3 guide assumes 95% ceiling.
Thesis: SMIC is printing real numbers with real pricing power: 5.7% ASP lift, 520bp margin expansion, volume and price rising together. But 44% of revenue is consumer electronics, the AI story is peripheral companion silicon, and management explicitly attributed growth partly to shipment pull-in. Utilization is pinned near ceiling after adding capacity, leaving no headroom for the next margin step. The 90% China concentration and dependence on controlled equipment make the downside binary and unquantifiable.
Verdict: AVOID — Conviction: MEDIUM
Catalyst: Q3 gross margin at or above 28% top end would confirm ASP momentum is not pull-in distortion. Conversely, a Q4 shipment decline confirming pull-in normalization would validate the bear case.
Key Risk: Export control escalation. 90% of revenue sits in one jurisdiction while the capacity buildout depends on lithography and metrology tools subject to controls. Every new wafer start is a bet on continued equipment access.
The Tell: Management attributed shipment growth to 'surging demand for peripheral chips brought by AI as well as customers' shipment pull-in.' Volunteering pull-in alongside AI demand reveals they know part of this quarter is borrowed from future quarters. It is a hedge against Q4 disappointment built into the current narrative.
Friction Level: HIGH_FRICTION — Bull sees 90% China revenue as structural moat from order backflow and localization. Bear sees it as 90% concentration in a single export-control jurisdiction where every new wafer start depends on foreign equipment access. Same number, opposite thesis.
Report not found
The report data is no longer available. Please return to the archive.