KLAC

Process-control monopoly that gets paid more as nodes get harder:

- ~80%+ share at the leading-edge optical wafer inspection segment. As process nodes shrink (3nm → 2nm → A16), defect detection difficulty scales super-linearly with feature size - every node transition raises KLA's wallet share per wafer.

- High-NA EUV is structurally KLA-positive - the higher pattern density requires more sophisticated inspection + metrology to maintain yield.

- ~25%+ of revenue is service / spares with software-economics margins; recurring base compounds with installed shipments.

- AI capex cycle pulls leading-edge wafer demand → more inspection tools shipped + more service revenue accumulated.

Bear:

- Same TSMC + Samsung + Intel + SK Hynix + Micron customer concentration.

- Smaller TAM than $AMAT / $LRCX - when the cycle inflects up, KLA is more capacity-constrained on tool delivery, when it inflects down, smaller customer base = faster contraction.

- China legacy-node business has been partial offset historically; tighter export controls cap it.