CDNS earnings call summary

Prepared Highlights

• Q2 2025 total revenue of $1.275B, exceeding guidance; 20% YoY growth (CFO).
• Non-GAAP EPS of $1.65, up 29% YoY (CFO).
• Core EDA revenue grew 16% YoY (CEO).
• IP business delivered >25% YoY growth in Q2 (CEO).
• System design and analysis revenue grew 35% YoY (CEO).
• Hardware systems had best quarterly revenue ever in Q2 (CEO).
• Non-GAAP operating margin was 42.8% in Q2 (CFO).
• Operating cash flow of $378M in Q2 (CFO).
• Q2 bookings were “stronger than expected” (CEO).
• Settlement with DOJ/BIS for ~$141M payment in Q3, resolving investigations into ~$45M of China sales from 2015–2021 (CFO).
• Official FY2025 guidance: revenue $5.21B–$5.27B; non-GAAP EPS $6.85–$6.95; GAAP EPS $3.97–$4.07; non-GAAP op margin 43.5%–44.5%; GAAP op margin 28.5%–29.5%; operating cash flow $1.65B–$1.75B (CFO).
• Official Q3 2025 guidance: revenue $1.305B–$1.335B; non-GAAP EPS $1.75–$1.81; GAAP EPS $1.14–$1.20; non-GAAP op margin 45%–46%; GAAP op margin 32%–33% (CFO).
• FY2025 revenue growth raised to 13%; EPS growth raised to 16% (CEO).

Mgmt Quotes

• “Cadence delivered exceptional financial results for the second quarter of 2025, exceeding our Q2 revenue and EPS guidance driven by ongoing broad-based strength across our AI-driven product portfolio.” (CEO)
• “Demand for our technologies continues to grow, driven by customers embracing our products at scale, and we are raising our financial outlook for the year to 13% revenue growth and 16% EPS growth for 2025.” (CEO)
• “Strength in other regions more than offset the impact of the export restrictions on China outlined in the BIS letter dated May 23, which was later rescinded.” (CFO)
• “Robust design activity and customer demand, coupled with our strong execution, drove 20% revenue growth and 29% non-GAAP EPS growth year-over-year for Q2.” (CFO)
• “I'm pleased with our strong first half results and the robust pipeline for the second half of the year.” (CFO)

Q&A Batch (1-5 of 16)

Q1 — Joseph D. Vruwink

• Topic: 物理AI对Cadence订单与工具需求的影响
• Key points:
  • 物理AI（自动驾驶、机器人、无人机）的硅片设计更注重功耗与电池优化，与数据中心AI不同；物理AI需在数据中心训练模型，因此同时拉动边缘与数据中心需求。
  • 物理AI仍处于早期阶段，预计还需3-5年发展。
  • 客户环境比六个月前更好。
• Mgmt stance: 看好 – 认为物理AI潜在规模可能超过AI基础设施，且客户超前投入RD，Cadence受益于AI三层（硅、仿真、模型）的全面需求。

Q2 — Gianmarco Paolo Conti

• Topic: 上调增长展望的原因，以及中国收入下滑下的需求驱动力
• Key points:
  • Q2中国收入占比9%，较Q1的11%下降；但其他地区强劲需求完全抵消了中国疲软。
  • Q2末积压订单（backlog）强于预期；订单增长来自AI、HPC和系统设计工作负载。
  • 硬件（Palladium）录得史上最高单季收入；IP业务受益于四家先进制程代工厂（TSMC、三星、Intel、Rapidus）的需求。
• Mgmt stance: 看好 – 需求广泛，三大业务线均表现强劲，AI超级周期推动市场改善。

Q3 — Vivek Arya

• Topic: 中国收入近端与长期展望
• Key points:
  • Q2中国收入占比从11%降至9%，指引已考虑当前出口管制并加入审慎余量。
  • 此前预计2025年中国收入同比持平；管理层暗示可能略增，但指引保持保守。
  • 长期看，中国收入占比可能维持相似或略低，因世界其他地区（美、日、韩）投资增长更快。
• Mgmt stance: 中性偏乐观 – 对中国持“乐观且审慎”态度，认为中国会增长，但其他地区增速更快导致占比稳定或略降。

Q4 — Harlan L. Sur

• Topic: 3D-IC/先进封装（3.5D）对业务的贡献
• Key points:
  • 行业正从2.5D转向3.5D封装（芯片堆叠），HPC/AI设计进入下半年流片阶段；部分客户单个封装集成最多10颗芯片。
  • Cadence的Integrity 3D-IC平台（含Allegro、Virtuoso、Innovus、Clarity等）是TSMC主要客户的优选方案，并正向Samsung、Intel、Rapidus扩展。
  • 未单独披露Allegro收入，但它是SDA业务的重要部分，并带动分析工具等销售。
  • 摩尔定律至少持续至1纳米（约10年），而3D-IC提供正交的集成维度，TSMC等有激进扩产路线图。
• Mgmt stance: 看好 – 先进封装仅处于初期，Cadence拥有独特优势，且该趋势与摩尔定律并行，推动长期增长。

Q5 — Lee John Simpson

• Topic: Agentic AI系统的商业模式与货币化路径
• Key points:
  • Agentic AI工具（如Cerebrus AI Studio、RTL生成验证）已独立于基础工具销售，并获早期客户（Samsung、ST）。
  • Cerebrus经典AI已用于超过50%的设计；Cerebrus AI Studio实现从3-5百万实例到30-50百万实例的设计能力提升。
  • RTL生成与验证领域，LLM+推理模型用于代码生成是重要趋势。
• Mgmt stance: 看好 – Cadence历史经验表明，通过创新和提升客户生产力可自然获得回报；多种商业模式均可用于变现。

Q&A Batch (6–10 of 16)

Q6 — James Edward Schneider

• Topic: Core EDA growth drivers and H2 outlook
• Key points:
  • Core EDA “doing phenomenally well” with broadest portfolio (digital, Virtuoso, verification, hardware)
  • AI adoption driving both core products and AI agents; Q2 saw signs of this
  • Key customer wins: SK Hynix (AI/HBM), ADI; hardware strength broad-based
  • H2 looks “particularly strong” on both software and hardware side in core EDA
• Mgmt stance: Bullish — broad portfolio and customer wins support sustained growth

Q7 — Jason Vincent Celino

• Topic: China revenue outlook and impact of past restrictions
• Key points:
  • Previously expected China flat year-over-year; now expects “slightly up”
  • Restrictions came and went; impossible to estimate revenue in a “parallel universe” without them
  • Full-year guide kept “prudent” despite China uptick
• Mgmt stance: Cautious but optimistic — prudent guidance maintained despite observed strength

Q8 — Gary Wade Mobley

• Topic: Backlog accounting and H2 bookings expectations
• Key points:
  • Q2 ending backlog excludes China bookings that were reserved due to restrictions
  • Expects year-end backlog higher than start; book-to-bill >1 for the year
  • H2 bookings in Q3 and Q4 each exceed revenue in those quarters; renewal cycle strong
  • Record level of backlog expected at year-end
• Mgmt stance: Bullish — confident in record backlog and >1 book-to-bill

Q9 — Jay Vleeschhouwer

• Topic: Agentic AI platform requirements (JedAI)
• Key points:
  • Workflow automation needs a data structure/database to store multi-tool, multi-flow actions
  • JedAI (joint enterprise data & AI platform) is essential for capturing reasoning across tools
  • Flexible deployment: on-prem (sensitive data), cloud, or hybrid; “uniquely positions” Cadence
• Mgmt stance: Bullish — JedAI’s flexibility is critical to delivering full workflow solutions

Q10 — Joseph Michael Quatrochi

• Topic: China backlog impact and full-year guide increase drivers
• Key points:
  • China orders excluded from Q2 backlog due to restrictions; year-over-year backlog still expected to reach record level
  • Full-year revenue guide increased by $50M, EPS by ~$0.12 — driven by strength across all regions, not just China
  • H2 has high renewal activity naturally falling in Q3/Q4, contributing to >1 book-to-bill
• Mgmt stance: Bullish — broad-based global strength supports higher guide and record backlog

Q&A Batch (11-15 of 16)

Q11 — Yu Shi

• Topic: Recurring revenue percentage trajectory and hardware tariff exposure
• Key points:
  • Q2 recurring revenue was 78% (multi-year low), due to hardware strength and paused ratable revenue in China.
  • Full-year expectation is ~80% recurring, 20% upfront; long-term normalized level has shifted from 85-15 to 80-20.
  • Hardware manufactured in North America for domestic market, internationally for overseas; tariff exposure is "quite limited."
• Mgmt stance: Neutral — explains dip as temporary mix shift, no tariff disruption, but no explicit guidance change.

Q12 — Ruben Roy

• Topic: IP segment growth sustainability and long-term outlook
• Key points:
  • IP revenue growth: +30% last year, +40% last quarter, +25% this quarter.
  • Drivers: chiplet architectures, multiple advanced-node foundries (≥4), improved portfolio (e.g., HBM4 from Rambus).
  • Long-term: IP expected to grow faster than Cadence average, with slightly lower margins; "critical mass" achieved.
• Mgmt stance: Bullish — "much more optimistic" than 2–3 years ago; sees sustained growth from AI, new foundries, onshoring.

Q13 — William Clarke Jeffries

• Topic: Tax benefit from R&D expensing (OBBBA) and impact on investment appetite
• Key points:
  • Cash tax benefit of ~$140 million for remainder of fiscal 2025 (two quarters); annualized benefit roughly double.
  • Non-GAAP effective tax rate unchanged at 16.5%; benefit already incorporated into annual guide.
  • No change to R&D investment strategy or appetite.
• Mgmt stance: Neutral — benefit is a one-time cash impact; normalized over time, no behavioral change.

Q14 — Joshua Alexander Tilton

• Topic: Recurring revenue growth trajectory and durability of total growth
• Key points:
  • Current split is 80% recurring / 20% upfront; drift toward lower recurring driven by faster growth in IP, hardware, SD&A.
  • Core EDA software showing "good growth" across all geographies; 80/20 split expected to persist "for quite some time."
  • No guidance for next year yet; change reflects customer consumption patterns, not weakness.
• Mgmt stance: Neutral — pleased with portfolio balance; recurring growth muted by mix, but core EDA remains strong.

Q15 — Nay Soe Naing

• Topic: Agentic AI adoption barriers and customer willingness
• Key points:
  • EDA productivity has improved ~100x over 20 years (e.g., 500 people/5 years → 50 people/1 year).
  • Chip workload expected to grow 30–40x by 2030 (from 100–200B transistors to 1 trillion); headcount only 2–3x.
  • Customers (MAG 7, top 50 companies) are willing to adopt if AI delivers measurable ROI (e.g., 20% better PPA, 5–10x benefit).
• Mgmt stance: Bullish — workload growth forces automation; adoption depends on product efficacy, not human resistance; core EDA also benefits.

Q&A Batch (16-16 of 16)

Q16 — Sitikantha Panigrahi

• Topic: 传统半导体客户 vs 系统公司（超大规模）需求趋势
• Key points:
  • 系统公司（超大规模、物理AI如汽车）业务增长显著，表现超过以往。
  • 传统半导体方面，NVIDIA和Broadcom等大客户表现极好；更广泛的传统客户（如memory、mixed signal）出现复苏迹象，例如ADI在Q2有很好的扩张。
  • 复苏仍处于早期阶段，公司因客户和地理多元化，不依赖单一客户群复苏时间点，保持耐心。
• Mgmt stance: 中性偏乐观 — 看到传统半导体初步复苏信号，但尚不确定，公司因多元化布局而不过度依赖，采取等待态度。