Onto Innovation — AI supply-chain exposure

Onto Innovation makes the process-control and metrology equipment that checks whether chips are being built correctly — the inspection and measurement tools that catch defects before they compound into scrapped wafers. Its specialty is the hard part of the moment: 3D bump metrology and quality control for advanced packaging, where multiple dies, memory stacks and interposers are fused together. Its Dragonfly platform is qualified for HBM and 2.5D CoWoS inspection, placing it directly on the assembly line for AI accelerators, where packaging now drives the bulk of its revenue.

The hook is that packaging is where yield is now won or lost. As accelerators move from single dies to stacked, multi-die assemblies, a fault caught late destroys far more value, so buyers pay for metrology that can see inside the stack. Onto is levered to that shift toward inspection-heavy CoWoS and HBM work rather than to raw wafer volume. That exposure is narrower and more cyclical than a broad tool vendor's, tied to the specific advanced-packaging capacity being built for AI — which is why the model places it high but concentrated.