They are using 33% revenue exposure to China mature nodes to subsidize a 20% physics advantage in AI memory.
Thesis: Veeco is mispriced as a cyclical equipment dumping ground. The market misses the structural pivot: they are leveraging cash cow legacy business (China LSA) to buy their way into the AI stack bottleneck (IBD for low resistance). The 20% resistance improvement in DRAM bit-lines is a physical necessity, not a nice-to-have. While the Street frets over a 150bps gross margin guide-down, management is essentially expensing the launch cost of a monopoly-potential product line. You buy the margin compression now to own the HBM yield solution later.
Verdict: LONG — Conviction: MEDIUM
Catalyst: Conversion of the two IBD300 EVAL systems at DRAM customers into High Volume Manufacturing (HVM) orders, expected late 2024/2025.
Key Risk: China (33% of revenue) export controls tightening on mature node laser annealing before the AI/Logic revenue ramp can replace it.
The Tell: When pressed on HBM customer breadth, Miller admitted the 'recent wins... are really not related' to the new IBD EVALs. He explicitly decoupled the HBM momentum from the IBD product cycle, refusing to fake 'synergy' where it doesn't exist yet.
Friction Level: MODERATE_FRICTION — Bulls see structural share gains in HBM and Logic via new tools (IBD/NSA). Bears see a cyclical peak driven by unsustainable China demand (33% of rev) masking organic weakness.
Report not found
The report data is no longer available. Please return to the archive.