Charles Shi

Ghazi rejected the disruption thesis, arguing autonomous AI workflows still need Synopsys sign-off and ground-truth physics, driving more software demand.

CEO cited robust AI design starts, accelerating standard evolution (3-4 years to ~2 years), and customer demand for foundry optionality. CFO added title availability is Q4-weighted.

CEO clarified it's prioritization not capability gap. Right skills in place, realigning priority and resources for hyperscaler delivery schedules. Making great progress on roadmap.

CEO acknowledged industry monetization gap. Hardware growing at record pace, but EDA software monetization lagging. Joint ANSYS solutions and agentic AI workflow changes are the two monetization catalysts.

CEO outlined three monetization vectors: ANSYS joint solutions (1+1>2), agentic AI workflow transformation (new business models), and IP royalty model evolution. Each represents a step change opportunity.

CFO declined to break out ANSYS vs legacy backlog. Stated strength across both core and ANSYS businesses with confidence in long-term growth.

CEO acknowledged shift from NRE+use fee to potentially royalty-based models for subsystem/chiplet delivery. Active discussions with customers on new business models to maintain profitability.

CEO confirmed China will be below corporate average but avoided providing specific dollar or percentage targets. Noted headwinds getting stronger.

CEO reaffirmed 12% DA CAGR target from Investor Day. Cited semiconductor R&D investment growing from 6% to 9% of sales and system company growth as supports.

CFO provided direct answer. Backlog declined from $8.1B at Q4 to $7.7B at Q1.

CFO cited $100M mechanical headwind from 10 fewer days plus IP and hardware timing being second-half weighted. CEO added the company has proven it delivers on commitments.

CEO acknowledged China pool of customers is shrinking due to restrictions and slowing economy. Wrapped up FY24 China at corporate average but avoided specific FY25 commitment.

CFO explained packaging is shorter lead-time business with momentum from multiple customers for additional capacity that was not visible entering the calendar year, now in the upper 50% growth range.

CEO said industry moves to highest capability tool for debug then shifts to cost-of-ownership play. X-ray adoption is mainly in memory at $75-100 million market, with KLA holding about 60% share.

Rick noted KLA holds many pieces across reticle qualification, requalification, verification, and print check, with a record reticle year reflecting customers buying with the future in mind regardless of pellicle strategy.

Rick explained DRAM inspection needs have reversed from a long holiday, with HBM design rules creating higher sensitivity requirements than logic on some layers. Bren quantified EUV adding about 1 point of intensity and HBM another point.

Bren maintained full-year China at 30% ± a few points, with second half higher than first half. Overall China business now expected down 10-15% vs prior 15-20% estimate.

Rick explained the decoupling as driven by design proliferation at leading edge increasing process control needs without proportional litho investment, plus HBM and advanced packaging drivers.

Bren guided China revenue percentage to drop to high 20s or 30% at most in March, down from 35% in December, with rev-rec timing a variable.

Bren cited potential license approvals, derisked customer opportunities, and possible China upside as swing factors but maintained guidance around current levels.

Higgins confirmed 3nm demand remains strong and stronger than anticipated six months ago, with 2nm the big driver into next year.

Brice clarified: share held in accessible accounts, but macro share lost because restriction scope more than doubled. Gary noted DRAM (high AMAT share) restriction was a major impact, as was NAND (low AMAT share) strength benefiting peers.

Brice emphasized DRAM strength is sustainable, not a blip. HBM growing 30-40% with about 15% of DRAM capacity allocated to HBM creating significant demand pull. Should be either a record or second-best year for DRAM.

Gary acknowledged the competitive landscape but emphasized ICAPS innovation pipeline including new TAM-expanding products and cost-competitive products, strong position at 28nm where investment is increasing.

Brice confirmed $2.5B in 2024 gate-all-around revenue with expectation to double in 2025, implying ~$7B cumulative and roughly 100K wafer starts of capacity, with ramp continuing beyond 2025.

Jason answered directly that increased license growth more than offsets royalty softness and that 20% plus royalty growth for FY28-29 remains intact.

Jason confirmed FY26 tracking to 22% at midpoint, exceeding the 'at least 20%' target. FY27 20% growth rate called 'very reasonable.' FY28 deferred entirely, citing possible new offerings whose revenue impact is still being modeled internally.

Jason stated unequivocally no revenue impact from Qualcomm litigation, explaining they had always forecasted as if they would not prevail.

Jason confirmed v9 remained flat at 25% of total royalties, then explained the absolute dollar growth was triple-digit despite the flat percentage.

Devgan rejected the disintermediation scenario and argued commoditization lands at the LLM layer, not the EDA engines.

Devgan emphasized 10,000 R&D staff with 1,000+ PhDs. Said agentic tools create new TAM in RTL generation and verification while driving more base tool consumption.

John Wall attributed low recurring mix to China pause plus hardware strength. Annual target still 80-20. Hardware demand continues to amaze with production as limiting factor.

John Wall maintained double-digit core EDA growth target. Acknowledged China and hardware transition year impacted 2024. Anirudh said product portfolio is better positioned than ever.

C.C. Wei explained CPUs are excluded from the AI accelerator definition because TSMC cannot identify their end use, but may consider including them later.

C.C. Wei confirmed TSMC is working on the customer's next-generation LPU and is confident in its technology position, but gave no specifics on winning back the current product.

C.C. Wei defined fairness as equal subsidies for all, and said the ~1,000-engineer Arizona R&D center will initially support the manufacturing cluster while doing some exploratory work.