Vijay Rakesh

Bill discussed ALC and OmniConnect multibillion-dollar TAMs and guided to outsized growth again in FY28 and beyond, without specific numbers.

CEO said FY27 200-gig-per-lane revenue is relatively light as the industry hasn't fully transitioned, but Credo is ready for production across copper and optical.

Bill sees the 200 gig per lane market addressed by AECs plus laser-based modules and ALCs, with ASP uplift from 800 gig to 1.6T across the portfolio; CPO stays further out.

Dan reiterated the long-term 63% to 65% gross-margin model, said Credo is currently above it, and expects normalization toward roughly 65% over the medium term.

Bill said scale-up entry is via PCIe Gen 6 (Gen 5 interim), bringing retimers and AECs simultaneously, with products protocol-agnostic at the 200-gig-per-lane generation.

Asked what drove the fourth customer's accelerated ramp; Brennan attributed it to conservative guidance and a fast qualification of the first ramping program.

Asked what drives the Q2 gross-margin guide; Fleming said at current scale any fluctuations are largely product-mix driven, with no change to the long-term gross-margin model.

Brennan described a cost-effective 12nm family and a power-efficient 5nm family, staying agnostic between full DSP and LRO as sub-10-watt 800 gig modules gain priority.

Bill said both new hyperscaler customers have committed design engagements ramping in fiscal '26 (conservatively the second half), with high confidence but timing still firming up.

Bill said the large majority of AEC shipments are AI-related, and essentially all 100-gig-per-lane deployments are for AI applications.

Bill declined to give an exact optical DSP customer count but cited much broader module-level design activity and increasing end-customer qualifications.

Steve said the 14 design wins span scale-up and scale-out with a mix of hyperscaler and enterprise customers, but declined to give dollar content per rack.

Steve said A&D is strong and will grow steadily, but acceleration is gated by primes' production capacity and very long (up to nine-month) cycle times.

Brian declined to size growth but said agentic AI and inference workloads drive PCIe-based component demand across hyperscalers, OEMs, and enterprise.

Steve cited multiple tailwinds: largest US defense budget, Boeing MAX production ramp, NATO countries doubling/tripling defense budgets, and space exploration acceleration with SpaceX and others all driving strong A&D demand.

Steve declined to quantify utilization but said it is quite low with ~$50M+ gross margin upside as factories return to full production. Eric explained OpEx growing in dollars from bonuses returning to target levels and pent-up demand for raises after last year's pay cuts, but declining as percent of revenue.

Steve believes Microchip's higher U.S. manufacturing percentage should be advantageous but acknowledged the rules are not yet clear enough to definitively say.

Eric said write-downs will continue beyond September but decline each quarter from $90M to $77M and lower. Underutilization charges will decline modestly in September with a bigger step-down in December when wafer starts increase.

Steve bluntly stated current rules incentivize production outside both U.S. and China. Will evolve strategy as rules continue changing.

Steve projected inventory below 200 days by September. Production must start growing before hitting 130-150 day target to avoid capacity ramp issues.

Eric declined to provide mid-year end market breakout, noting it will come with fiscal year-end results in May. All end markets remain weak.

Eric confirmed significant reserve charges are already reflected in gross margin. Steve's plan to reduce inventory by ~$250M through FY26 end should reduce future charges.

Hassane split content roughly 50/50 high vs medium/low voltage at $115,000 per rack, with high voltage content up ~10x driving an outsized increase.

Trent confirmed no further exits beyond 2026. Utilization at 77%, ramped up quickly to meet strong demand. Will match utilization to demand signals for remainder of year.

Trent said auto roughly flat in Q2 shipping to natural demand, no replenishment yet. Industrial up mid-single-digit. Other (including AI DC) up mid-teens QoQ.

Trent said 40%+ gross margin is within sight. Lean inventory and right distribution levels mean utilization can be matched quickly to recovery. No need to burn through excess first.

El-Khoury said reaching 10-15% of revenue is a matter of when, not if. Products are available, market opportunity exists. Question is speed of design cycle and proliferation.

El-Khoury said no crystal ball on tariff outcome. Manufacturing footprint flexibility is the preparation strategy. Recognized by customers as advantage.

El-Khoury said peer's struggles preceded bankruptcy. Customer sourcing decisions already baked into baseline. ON wins on its own product merit, not competitor weakness.

El-Khoury said no direct tariff impact but too soon to assess indirect effects. No material pull-ins or pushouts observed from tariff activity.

El-Khoury said pricing pressure is industry-wide, not ON-specific or product-specific. Cost improvements expected to offset low single-digit pricing declines.

Trent said utilization will remain in mid-50% range until demand recovery. Will increase utilization only when demand signals improve.

El-Khoury declined specific SiC guidance, saying share is gained and designs are won but 2025 is purely demand-dependent on EV vehicle sales.

Rene confirmed 128 cores today and said more cores is the direction of travel, but declined road map specifics. Jason said silicon breaks out once it is at least 10% of revenue.

Jason guided license revenue growth in the 20% range for FY2027, above the long-term high single-digit to low double-digit floor. Attributed strength to the ongoing AI investment cycle.

Rene confirmed data center royalties doubled YoY and expects them to double again in FY2027. Emphasized IP and AGI CPU businesses are additive, not cannibalizing each other.

Jason declined to provide any specifics about potential products from the SoftBank partnership. No timeline or product confirmation despite Vijay framing the question around the significant $200M quarterly revenue.

Rene declined to confirm or deny a merchant CPU product. Said market share gains are somewhat independent of that strategy and expressed confidence in the technology roadmap.

Rene described Stargate's 10GW vision with huge compute potential. All Stargate opportunities currently use Arm as core CPU. Declined specifics on products or timelines.

Jason confirmed 13 CSS customers split roughly 6 client, 6 infrastructure, 1 auto after adding first automotive CSS license this quarter.

Rene emphasized Arm's energy efficiency advantage as the sole CPU provider in Stargate. Confirmed Grace Blackwell as initial chips but declined further technology details.

Rene deferred Cobalt specifics to Microsoft but highlighted strong partner deployment momentum and the Grace Blackwell tailwind for data center Arm adoption.

Jason expects Arm China to remain consistent at ~25% of revenue near-term but to decline to mid-teens over time.

Gave a specific current SAM percentage and corrected the analyst's back-of-envelope share math over CSBG inclusion.

Declined to quantify and said he was not prepared to do so, offering only the qualitative taller-stack argument.

Tim deferred to customers on high bandwidth flash timing but emphasized early engagement on next-gen technologies and growing NAND importance in the AI memory stack.

Doug confirmed content per wafer increases with HBM4 higher stack counts but declined to provide specific dollar-per-wafer-start numbers.

Tim said future nodes bring backside power, more advanced packaging, and CFET, all expanding Lam's SAM and share opportunity. Expects robust investment across all segments continuing into 2027.

Doug said 2026 growth will likely come from both memory and foundry/logic, with more granularity coming on the December call.

Doug said he hasn't yet tried to correlate the ITC increase with potential WFE tailwinds for 2026.

Doug declined to give a specific NAND WFE growth number but confirmed it would be up year-over-year.

Gianluca said incremental gross margin is well above 60% but declined to set a target; Dave declined to model an attach rate, calling it application-specific.

Mosley pointed to aerial density driving more exabyte output with minimal incremental cost as the key margin driver. Customers lean heavily into space and power efficiency per terabyte.

Romano confirmed March guidance includes HAMR ramp and margin improvement continues through calendar 2026. OpEx approaching 10% of revenue target and could improve further with revenue growth.

Romano confirmed they achieved the first milestone earlier than expected. Every quarter will differ but the Analyst Day model could be exceeded. Mosley noted they are studying AI content generation uptake.

Romano confirmed all factors (HAMR mix, pricing, cost per TB decline) contributing to continued sequential margin improvement. Mosley noted HAMR qualification progress going well with more customers ramping beyond the initial three.

Mosley declined specific HAMR mix guidance but confirmed acceleration. Romano said wait for Analyst Day for future profitability model details.

Mosley and Romano declined specific HAMR mix targets but confirmed they expect the majority of the ramp in second half of calendar 2025 after multiple qualifications complete.

Asked for a 2025 unit-shipment outlook and the idle-capacity ramp plan; management declined to size units, saying factories are full and additional exabytes will come from technology transitions to HAMR.