Tom O'Malley

Luther defended progress with 530bp of improvement over 8 quarters, highlighting cost reductions, yield improvements, and pricing optimization. Emphasized they are early stages of the margin expansion opportunity.

Anderson acknowledged internal component constraints were resolved, enabling faster production ramp across 2 sites, but did not specify the exact component.

Jim expects sustained InP supply-demand imbalance through at least 2027, possibly longer if scale-up customer forecasts materialize. Driving aggressive capacity expansion goals beyond initial doubling target.

Jim confirmed vast majority of 4x book-to-bill from 800G and 1.6T transceivers. 800G stronger than expected, 1.6T accelerating. OCS and CPO also contributing. Communications/DCI bookings also very strong.

Jim noted minimal gross margin impact this quarter since 6-inch just started. Benefits begin next calendar year, becoming more meaningful each sequential quarter. Sherri added 6-inch is less than half the cost of 3-inch.

Jim provided clear segment breakdown: datacenter accelerating from 4% to ~10% sequential in Q2, comms up single digits sequentially, industrial stable to slightly up.

Sherri identified mix as primary potential headwind for Q4 gross margin, but noted cost reduction benefits seen across all segments and pricing optimization in lasers and datacom.

Jim clarified SiC devices/modules was pre-revenue so no forecast impact; continued investment in differentiated substrate and EPI production with healthy customer demand.

Jim provided clear segment-level guidance: communications up sequentially offset by industrial-related declines netting to flat midpoint.

Jim sees share gain opportunity driven by technology roadmap and supply chain resilience, noted geographic diversity increasingly valued by hyperscaler customers.

Jim confirmed additional non-strategic actions underway beyond battery platform and Newton Aycliffe, to be shared at appropriate time and Investor Day.

Jim declined to specify exact 1.6T timing beyond calendar 2025 ramp, emphasized multi-technology approach as competitive strength across EML, VCSEL, and silicon photonics.

Framed peer deals as validation and described integration progress, but did not size the TAM the analyst asked about.

Said dispersion between low and high customers is unchanged from prior quarters but confirmed inventory digestion is over.

Rafael confirmed channel went to 11 weeks in Q1 as planned, Q2 guide holds flat at 11 weeks long-term target.

Bill confirmed data center products are favorable to corporate gross margins, consistent with focus on value extraction across all applications.

Rafael confirmed shift from tight control to long-term 11-week target in 2026, reflecting improving demand. Finished Q4 at about 10 weeks.

Rafael attributed C&I sequential Q1 strength to digital networking normalization and secure card growth that should benefit throughout 2026.

Rafael broke out 60% core industrial / 40% consumer IoT. Consumer IoT benefiting from smart glasses and wearables demand. Core industrial broad-based improvements in energy storage and building automation.

Rafael highlighted Q3 auto revenue only 3% below Q4 2023 peak. Accelerated growth drivers (SDV, radar, connectivity) continue to drive performance, reflecting secular shift to software-defined vehicles.

Kurt confirmed 11 weeks as the stated long-term target based on bottom-up assessment, same as pre-COVID. It was part of the November Investor Day model.

Kurt refused segment-level Q4 guidance, limiting commentary to total company flat to slightly up seasonal norm.

Kurt confirmed inventory digestion continues at Western Tier 1s with individual customer plans, unable to predict timing of completion.

Bill said they're balancing internal inventory dollars and utilization with foundry orders. Second half revenue growth would trigger increased utilization.

Irving emphasized predictable pricing philosophy over opportunistic pricing. Noted LTAs extend into CY2029 with base volume pricing and incremental upside on above-base volumes.

Irving declined to share deal terms (confidential) but highlighted benefits: less real estate in the drive improving manufacturability and reliability, plus reduced energy requirements vs conventional laser diodes. More details planned for Innovation Day.

Irving noted NVIDIA's KV cache offload initiative accelerates inference, which increases data velocity and volume, benefiting HDD demand. WDC working on innovations to improve drive bandwidth and throughput, details to be shared at Innovation Day.

Irving distinguished firm POs (5 hyperscalers) from LTAs. Calendar 2027 agreement has significant commercial protections. A structurally different environment from DRAM/NAND pandemic era.

Kris noted HDD share of AI CapEx trending from low single digits to 4-5%. Not a direct correlation but share is increasing as data storage needs grow.

Irving clarified ASP per drive up, ASP per terabyte down low single digits. Stable pricing environment; the per-TB decline is mix-driven from higher capacity drives.

Irving cited 190 exabytes achieved via technology rather than capacity investment. High yields at scale differentiate WDC. No incremental capacity investment needed.

Irving cited structural industry changes removing excess capacity. LTAs moved from 3-6 months to 9-12 months, giving better visibility and supply planning capability.

Don noted continued ASP progression driven by cloud mix shift. Price per unit expected stable to up as mix continues moving toward higher-capacity data center drives.

Wissam declined to comment on capacity ceiling, deflecting to supply-demand balance discipline and CapEx prudence.

Wissam confirmed no change to disciplined CapEx approach. Next catalyst will be ramping the next NAND node; more detail at investor day.

Asked to quantify the enterprise SSD margin tailwind; management called it accretive and among the best-priced Flash markets but did not quantify, partial.

Bill reassured on supply readiness (18-24 months of preparation, inventory build) and demand generation, but gave no hard numbers on the ramp.

CFO confirmed the entire optical portfolio (3 legs, each >$100M) is in the FY27 guide; about half of the absolute-dollar growth is optical.

Bill framed ZF like AECs, moving straight from samples to customer qualification, and confirmed the ramp pulls into Q1 fiscal '27 with units already shipped.

Dan gave the top-4 mix (42%, 24%, 16%, 11%) and noted a single hyperscaler's ramp is never linear, with 12-month visibility helping manage the give-and-take.

Bill said the ALC TAM is driven by both quantity and ASPs, with scale-up going row scale potentially up to 10x the number of scale-out connections.

Asked about competitiveness at the 1.6T node and the three-nanometer DSP platform; Brennan said Credo moved to three nanometer for power efficiency and expects the 1.6T transceiver market to take time to develop.

Asked whether PCIe retimer demand is concentrated in two customers or proliferating; Brennan said it will be broader and expects PCIe to be the protocol most hyperscalers pursue near term.

Brennan said scale-up starts with PCIe Gen5/Gen6, and at the physical layer Credo's SerDes can span Ethernet SUE, UALink and NVLink Fusion; open standards help the market.

Bill said PCIe switching is a natural future progression from retimers but the near-term focus is retimers and PCIe AECs.

Bill said Credo is agnostic between Ethernet and PCIe; the trend toward faster lane speeds favors Credo regardless of which architecture wins.

Bill said growth is driven across all product lines, with AECs the main FY26 driver by market size and optical potentially the faster-growing line.

Rick noted any incremental NAND demand showing up today would translate to 2027 deliveries given industry supply constraints, and corrected the questioner by saying the other company is a peer, not a competitor.

Bren cited positive momentum across packaging, memory, reticle inspection, and e-beam, with the largest businesses like broadband plasma growing faster than the market, influencing overall share numbers.

Rick broadly agreed but added packaging brings it closer to $10B per $100B, noting KLA's participation is above average intensity given process control relevance across all three components.

Rick confirmed both new customers entering leading edge and expanded opportunities with existing ones, driven by proliferation of design starts.

Rick explained customers are shifting WFE mix toward process control as design complexity, large die, packaging, and HBM all increase inspection needs at the expense of other equipment categories.

Bren confirmed DRAM, specifically HBM, as the biggest driver with NAND limited to technology investments and upgrades.

Bren described drilling through all business including service contracts and factory imports. Rick credited the logistics team that managed through COVID for the detailed analysis.

Rick described NAND improvement as broad-based but from very low levels, expecting some WFE growth there but not significant in 2025.

Rick noted no short-term plan changes but growing demand momentum for AI infrastructure, with several critical products still supply constrained.

Higgins expects some incremental NAND investment in 2025 off a very low base, driven by improving utilization rather than a broad recovery.

Su answered directly, framing both server and data center AI as significant 2027 drivers, with server up over 80% in the second half.

Lisa detailed Q2 double-digit sequential growth in both server and DC AI, Helios ramp starting Q3 with significant Q4 ramp. Acknowledged supply tightness but expressed confidence in meeting growth targets.

Lisa Su said AMD plans HBM and wafer supply over multi-year time frames with multi-year agreements already extending beyond the current ramp, positioning AMD to grow substantially in 2026 regardless of market tightness.

Lisa Su said the entire ecosystem has to plan together, AMD feels very good about its supply chain preparedness, and expects constraints to open up as everyone works through it.

Lisa Su said early MI450 customers are focused on rack-scale Helios solutions, with phenomenal OCP reception and customer engineering teams actively engaging; other form factors remain available.

Lisa Su said MI308 inventory is work-in-process not finished goods, so ramp will take a couple of quarters once licenses are granted; overall position is better than 90 days ago.

Lisa Su denied meaningful pull-forward, citing strong sell-through and ASP-driven share gains, and said Client growth in the second half will be a little bit less than seasonal but not flat to the first half.

Lisa Su said AMD has all the building blocks plus ecosystem partnerships, with interoperability on networking being a customer priority across reference and custom configurations.

Lisa Su emphasized revenue-share gains in high-end notebook, commercial, and desktop; said internal planning assumes sub-seasonal second half given the strong first-half start.