Brian Chin

New HDD customer complementing rather than fully replacing captive test. HDD revenue expected to double 2025 to 2026. In SoC, one big player has captive strategy expected to change over next few years as foundry broadens customer base. Memory captive strategy likely more persistent.

Four key differentiators: more resilient supply chain with shorter lead times, better reliability and uptime in production, next-gen instruments in beta with more power and memory, and superior mission-mode testing capability for defect detection.

View changed dramatically between January (optimistic) and March (pessimistic from uncertainty) to year-end. Compute and networking significantly stronger than expected. Mobile and auto/industrial weaker. Memory strengthened but less than compute.

Seasonal uplift expected in Q4 robotics but not a hockey stick. SLT Titan system designed with significant upgradability across AI accelerator generations within power envelope. SLT incumbency advantage similar to but slightly weaker than ATE.

Upside to VIP SAM from $400M and to total compute TAM, especially in merchant GPU space. Not providing formal TAM estimate currently.

Pull-forward hasn't had meaningful effect on Teradyne. Customers pulling forward end product demand were absorbing it with existing test capacity.

VIP TAM figures are ATE-only. SLT could add 10-30% of total TAM on top. Will provide updated SLT TAM estimate in January.

Onshoring changes ship-to location but not total test demand. Testers purchased at point of wafer volume need, unlike front-end fabs.

Provided full breakdowns for both 2024 ($4.6B) and 2025 ($4.9B) SOC TAM across Compute, Mobile, Auto/Industrial and service.

About 10% utilization increase year-on-year exiting 2024. Business mix already shifting from upgrades to system sales, past tipping point.

Unified marketing, sales and service across UR and MiR. 50 largest partners already sell both; single coverage improves efficiency and customer experience.

Brice confirmed approximately 100K wafer starts installed or to be installed by quarter end on gate-all-around nodes, with the build-out in the first 3 innings. Subsequent nodes will follow.

Brice noted datacenter growing 20% with AI components at 40%, leading-edge fabs at 100% utilization, and gate-all-around as a significant performance improvement supporting a strong node. Recent announcements from CSPs and foundries show continued energy.

Brice confirmed the $549M backlog reduction covers more than the fiscal year; the 12-month impact was about $380M, consistent with the $400M annual estimate.

Gary sees litho and materials engineering as complementary, with 4F-squared DRAM focused near-term and 3D DRAM further out, both increasing materials engineering intensity.

Brice noted China variability depends on ICAPS health, as DRAM/NAND is mostly served and leading-edge restricted; watching automotive, industrial, and analog signals.

Bren expressed confidence in packaging share gains across both logic and HBM memory, with positive trends in both segments expected to continue into 2026.

Bren acknowledged some space constraints could affect timing into second half 2026 and 2027 with new memory fabs coming online, but does not see it as a major gating factor currently.

Bren confirmed share gains primarily in CoWoS logic packaging with increasing HBM momentum, driven by inspection applications with high sampling rates rather than metrology.

Rick gave a brief answer indicating KLA has derisked spending from non-leading foundries in their 2025 plan, suggesting limited dependence on those customers.

Rick said it's too early to model 16A intensity but noted the fundamental shift toward larger, higher-value die with higher density creates growing opportunities for process control.

Doug revealed more than half of the $8B WFE per $100B data center spend comes from memory (enterprise SSDs and HBM), with GPUs and accelerators as the remainder.

Tim declined to specify the exact layer count but confirmed the Vantex win is multigenerational, covering current through next couple of technology nodes.

Doug highlighted the DQR production tool-of-record decision as a major milestone, though near-term dollars will be modest in 2025.

Sadana confirmed the structural production shift to leading-edge products is real, with leading-edge nodes very constrained and capacity dedicated to fast HBM ramp.

Sadana declined to speculate on competitor actions but emphasized Micron's own decisive supply reduction to align supply with demand.