Memory and Compute Growth

Q: C.J. Muse with Cantor Fitzgerald asked about the memory business trajectory into December and the rank order of growth for 2027.

A: Greg Smith (CEO) stated that memory will continue at a strong level in Q4 and expects significant year-over-year growth in 2027, with HBM being the primary driver followed by Flash and DRAM.

Q: Vivek Arya with Bank of America Securities asked about the impact of lower HBM stack heights on test complexity and opportunity.

A: Greg Smith (CEO) explained that the trend toward higher stack heights reversed in 2026, and lower stack heights help deconstrain shipments while increasing the need for testers because more stacks require more testing.

Q: Krish Sankar with TD Cowen asked about HBM destacking and competition from Korean suppliers.

A: Greg Smith (CEO) noted that destacking is not currently a factor and that Korean competitors primarily serve the wafer sort market rather than the performance test market.

Gross Margins and Product Test

Q: C.J. Muse with Cantor Fitzgerald asked about the Product Test book-to-bill of 1.4x and its impact on gross margins.

A: Michelle Turner (CFO) stated that Q3 gross margins were 59.1% and that Product Test is accretive to the portfolio, though 2027 margins face both volume headwinds and cost pressures.

Q: Vivek Arya with Bank of America Securities asked about the Q4 to Q1 transition and potential seasonality.

A: Michelle Turner (CFO) expects growth from Q4 to Q1 and throughout 2027, while Greg Smith (CEO) added that historical seasonal patterns are no longer relevant due to the shift toward data center build-out schedules.

Inventory and Capacity

Q: Timothy Arcuri with UBS asked about the $200 million inventory increase and if 2027 will be back-end loaded.

A: Greg Smith (CEO) attributed the inventory increase to growth preparation and longer lead times for components, noting that 2027 will likely be back-half loaded but strong overall.

Q: Timothy Arcuri with UBS asked about share gain pace and lead times relative to competitors.

A: Greg Smith (CEO) stated that capacity for the UltraFLEXplus platform is 4x higher than 18 months ago and that the company is using dual sourcing to maintain responsiveness.

Q: Krish Sankar with TD Cowen asked how much of the inventory build is for true demand versus component constraints.

A: Michelle Turner (CFO) estimated the split is about 50-50, noting that inventory levels were unusually low in Q1 and are currently normalizing.

Robotics and Physical AI

Q: Mehdi Hosseini with SIG asked what is needed for higher growth rates in industrial automation.

A: Greg Smith (CEO) explained that the company is building partnerships with labs and OEMs, but noted a gestation period, expecting robust transaction growth in 2027 and end-market growth in 2028.

Q: Jim Schneider with Goldman Sachs asked about visibility in the Robotics business compared to the broader company.

A: Greg Smith (CEO) noted that Robotics visibility is pipeline-based, though the largest customer revenue tripled from 2025 to 2026 and is expected to see robust growth next year.

CPO and Future Engagements

Q: Mehdi Hosseini with SIG asked for an update on demand vectors and the model update.

A: Greg Smith (CEO) stated that the company will update its model in January, noting that Semiconductor Test is growing faster than previous assumptions while mobile is underperforming.

Q: Jim Schneider with Goldman Sachs asked about near-term visibility and the Q1 outlook.

A: Michelle Turner (CFO) said the company has good line of sight for Q1 and expects progressive growth throughout the year.

Q: Joseph Cardoso with JPMorgan asked about the timing and TAM for CPO and NPO.

A: Greg Smith (CEO) expressed enthusiasm for CPO, noting a 2026 TAM of $100 million that could reach $300 million to $700 million by 2028, though test intensity remains a limiter.

Q: Joseph Cardoso with JPMorgan asked about potential engagements beyond the current 3 hyperscaler customers.

A: Greg Smith (CEO) said the company is approaching new customers with the same process used for the first 3, expecting 2027 to be a good year and 2028 to be even better.