Management explicitly admitted 2025 HBM revenue will be flat, yet guided total AI revenue to double. The math only works if the tech is undeniable.
Thesis: Veeco is executing a high-stakes pivot: trading high-margin but terminal China mature-node revenue for lower-margin but secular AI advanced packaging revenue. The market hates the gross margin step-down (43% -> 42%), but misses the signal intensity of the NSA 'straight sale'—a customer skipping evaluation to buy immediately. This confirms the tech is a bottleneck solution. They aren't hoping for AI wins; they are scaling them. The gap between the China roll-off and AI ramp is the risk, but the valuation already prices in the fear.
Verdict: LONG — Conviction: MEDIUM
Catalyst: Conversion of the two ongoing NSA evaluations into volume orders in 2H 2025, validating the $30-40M per application unit economics.
Key Risk: The China revenue cliff (39% of Q4 sales) accelerates faster than the AI evaluation pipeline converts, creating a revenue air pocket in mid-2025.
The Tell: When asked about HBM, the CEO didn't spin. He stated: 'I would say our HBM revenue is nearly flat and steady with this one customer.' He refused to pull forward 2026 hope into 2025 guidance, lending massive credibility to the 20% AI revenue target he *did* commit to.
Friction Level: MODERATE_FRICTION — Bears see margin compression (43% to 42%) as deterioration; Bulls see it as the cost of swapping toxic China revenue for sticky AI leverage.
Report not found
The report data is no longer available. Please return to the archive.