CMBU margins were flat at 83% despite high-teens pricing because HBM mix is eating the gains, and management called it progress.
Thesis: The near-term is locked: 75% of 2027 output committed, $32B in cash-backed deposits, pricing up across every segment, and Q1 GM guided as the floor. But the SCA architecture includes ceiling prices on three-quarters of committed revenue, which caps the upside the market is pricing in. Meanwhile the $50B+ CapEx wave creates capacity arriving 2028-2030 into whatever the AI cycle looks like then. You own a fortress with a ceiling built into its walls.
Verdict: HOLD — Conviction: MEDIUM
Catalyst: Fiscal Q1 gross margin confirmation above the guided floor, and SCA negotiations for 2028+ volumes pricing above current ceiling bands. Singapore HBM packaging facility initial output in early calendar 2027 validates the ramp timeline.
Key Risk: CMBU margins printed 83% flat despite high-teens pricing because HBM mix is diluting. If HBM margin dilution persists while conventional DRAM ASP gains decelerate, the margin expansion engine stalls before the CapEx wave arrives.
The Tell: When Krish Sankar asked directly about one large customer de-speccing HBM, Sanjay responded with generic language about 'optimizations' and 'latent nature of need' without a single denial, customer name, or quantification. The same passage has CMBU margins flat at 83% despite high-teens pricing, which management attributes to HBM mix. HBM is diluting margins while the narrative claims the opposite trajectory.
Friction Level: HIGH_FRICTION — The bull sees 26 SCAs, $32B in cash deposits, and 75% of 2027 committed as proof this cycle has structurally reset. The bear sees ceiling prices inside three-quarters of those SCAs, a $50B+ CapEx wave arriving 2028-2030, and 87% gross margins in memory that have never held. Same numbers, opposite reads on whether this is a plateau or a peak.
Report not found
The report data is no longer available. Please return to the archive.