Why I Am Upgrading FormFactor To A Strong Buy

FORM is one of the cleaner ways to express the AI memory bottleneck trade without paying the full multiple for GPU vendors. The important mechanism is mix, not just revenue: as HBM test complexity rises, probe-card content per device can expand faster than wafer starts, which supports ASPs and margin even if the broader semiconductor cycle cools. That makes the name a leveraged beneficiary of continued AI capex, with the best upside likely coming from customers that are still early in HBM ramp and therefore under-optimized on test spend.

The near-term risk is that this is a utilization and qualification story disguised as a secular one. If HBM yields improve or customers push for lower test cost, spending can flatten before unit demand does, which would hit FORM first through gross margin and operating leverage. The manufacturing ramp is supportive only if demand is synchronized; otherwise it can become a margin headwind for 2-4 quarters before the fixed-cost absorption shows up.

Contrarianly, the market may be over-weighting co-packaged optics as a 2028 driver when it is still more of a narrative enhancer than an earnings line item. At 40x forward earnings, the stock is already pricing in multi-year execution, so the main vulnerability is not a collapse in AI demand but any sign that HBM intensity normalizes or backlog converts slower than expected. The thesis is falsified by margin compression, a guide-down in advanced-node demand, or evidence that the Farmers Branch ramp is outpacing order growth.