Texas Instruments confirmed what the data already showed: the analog upcycle is dead. TI guided to the low end of capex and warned of multiple quarters of factory load cuts to drain inventory that kept growing through a supposed recovery. April's demand spike, which briefly looked like a cycle turn, was tariff-driven pull-forward tied to Liberation Day. Management said so plainly. The run rate of real demand is lower than it appeared.
The deeper problem is structural, not cyclical. Lagging-edge capacity, the older nodes serving auto and industrial markets, now competes directly with Chinese fabs operating on fully depreciated equipment. A new fab cannot profit selling chips at prices a fully depreciated Chinese fab can sustain. The current downturn has already run roughly twice the length of a typical cycle. Even a 5% supply increase during any future upturn would gut incumbents' pricing power. Global utilization remains lackluster. TI is right to cut inventory targets because they likely will not need as much capacity as their COVID-era capex plan assumed.
Read the full piece for the argument that lagging-edge capacity has entered a permanent glut regime, not a prolonged but recoverable downturn. The edge price war analysis from 2023 is the foundation, and this update closes the loop with hard earnings data from TI. If you cover semiconductors, auto supply chains, or industrial capex, the structural framing here will change how you model the next quote-unquote upcycle.
[READ ORIGINAL →]