HBM and chiplets just hit 60% of revenue, driving margins to 50.6%, yet management refuses to guide 2025. The mix shift is real, but the ceiling is visible.
Thesis: Camtek is the pure-play derivative on the HBM supply crunch. The thesis is simple: You buy the bottleneck. Revenue beat by 34% YoY, but the real story is the mix shift—HBM and chiplets exploded to 60% of sales (vs 30% expected). This mix shift is accretive, driving gross margins over 50%. The 'capacity ceiling' at $500M is a high-class problem when you're running at $400M. You ride the margin expansion until the order book softens.
Verdict: LONG — Conviction: HIGH
Catalyst: The transition to 12-Hi HBM stacks. While unit counts per GPU might drop, the complexity and yield challenges increase inspection intensity per wafer.
Key Risk: Customer concentration. Five players control the entire HBM market. If Samsung or SK Hynix yield improves faster than expected, inspection intensity drops.
The Tell: Ramy Langer correcting the analyst on 12-Hi scaling: 'Going to a stack of 12 they would put less HBMs around the chiplets... I'm not sure what eventually the total numbers of DRAMs that will be scanned.' He admitted volume doesn't scale linearly with stack height.
Friction Level: MODERATE_FRICTION — The durability of the HBM mix shift. Bulls see a structural new baseline. Bears see a cyclical peak concentrated in five customers who control the roadmap.
Report not found
The report data is no longer available. Please return to the archive.