Management is aggressively expanding capacity for a hyperscale ramp while simultaneously admitting they cannot see demand beyond the next quarter.
Thesis: MPS is being valued as a structural AI infrastructure winner (like NVDA) but is operating with the visibility of a cyclical analog vendor. The critical disconnect is 'short lead times' in a supposed super-cycle. True bottleneck owners have 52-week lead times and non-cancellable backlog. MPS has 'lean inventory' and orders that drop in quarter-to-quarter. They are building high fixed-cost capacity ($4B target) for demand that is admittedly 'lumpy' and 'dynamic'. When the cycle turns, they hold the bag on utilization.
Verdict: AVOID — Conviction: MEDIUM
Catalyst: Q4 Guidance. CFO Blegen explicitly flagged 'concern about Q4' due to 'atypical seasonality' and the lack of backlog coverage, despite the general bullish posturing.
Key Risk: ASIC volatility. The 'lumpy' nature of their new AI wins means a single customer pushout (typical in hyperscale) leaves their new capacity idle, hitting margins immediately.
The Tell: CFO Blegen's slip regarding Q4: 'I don't want to send a negative signal, it's just that that's the nature of the demand profile.' He explicitly linked 'short lead times' to 'concern about Q4', contradicting the CEO's 'I don't listen to any other calls' bravado. It reveals they are trading visibility for agility, which works until it doesn't.
Friction Level: HIGH_FRICTION — Lead times. Bulls view short lead times as 'responsiveness' and execution speed. Bears view them as a lack of pricing power and socket security in a supply-constrained market.
Report not found
The report data is no longer available. Please return to the archive.