CRDO earnings call summary

Prepared Highlights

• Q1 FY2026 收入: 2.231 亿美元，环比增长 31%，同比增长 274%，创历史新高 (CFO)。
• 产品收入: 2.171 亿美元，环比增长 31%，同比增长 279%；AEC 产品线再次实现健康的两位数环比增长，创下新纪录 (CFO)。
• 非 GAAP 毛利率: 67.6%，高于指引上限，环比增长 20 个基点；产品非 GAAP 毛利率为 66.7%，主要得益于规模扩大 (CFO)。
• 非 GAAP 营业利润: 9620 万美元，营业利润率 43.1%，环比增长 635 个基点，得益于收入增长带来的经营杠杆 (CFO)。
• 非 GAAP 净利润: 9830 万美元，创历史新高，环比增长 51%；非 GAAP 净利率为 44.1% (CFO)。
• 现金流: 经营活动现金流 5420 万美元；资本支出 280 万美元，主要用于生产设备；自由现金流 5130 万美元 (CFO)。
• 现金与库存: 期末现金及等价物 4.796 亿美元，环比增加 4830 万美元；期末库存 1.167 亿美元，环比增加 2660 万美元 (CFO)。
• 客户集中度: 三家超大规模客户各自贡献超过 10% 的收入；第四家超大规模客户在 Q1 首次贡献了实质性收入 (CEO)。
• AEC 市场进展: AEC 采用从机架内扩展到机架间解决方案；一家客户因 GPU 密度提升四倍，可在长达 7 米的扩展网络上全部采用 AEC (CEO)。
• 光学业务: 光学 DSP 势头强劲，有望在 FY2026 再次实现收入翻倍；提供支持高达 1.6Tbps 的全 DSP 和线性接收光学 (LRO) 解决方案 (CEO)。
• Retimer 业务: 以太网 Retimer 业绩增强；PCIe Retimer 系列获得显著客户关注，预计 2025 年获得设计订单，2026 年产生量产收入 (CEO)。

Official guidance

• Q2 FY2026 收入指引: 2.30 亿至 2.40 亿美元，按中点计算环比增长 5% (CFO)。
• Q2 FY2026 非 GAAP 毛利率指引: 64% 至 66% (CFO)。
• Q2 FY2026 非 GAAP 运营支出指引: 5600 万至 5800 万美元 (CFO)。
• Q2 FY2026 稀释加权平均股数指引: 约 1.90 亿股 (CFO)。
• FY2026 全年展望: 预计收入同比增长约 120%，季度收入环比中个位数增长；非 GAAP 运营支出同比增长低于 50%；非 GAAP 净利率预计约为 40% (CFO)。

Mgmt quotes

• “在第一季度，我们实现了 2.23 亿美元的收入，环比增长 31%，同比增长 274%。我们的非 GAAP 毛利率为 67.6%，并实现了近 1 亿美元的非 GAAP 净利润。” (CEO)
• “AEC 提供了无与伦比的可靠性、信号完整性、能效和系统成本组合……与光学解决方案相比，它们的可靠性高出 1000 倍，功耗仅为其一半。” (CEO)
• “我们的非 GAAP 净利润为 9830 万美元，创历史新高，环比增长 51%……非 GAAP 净利率为 44.1%，这体现了我们业务中显著的经营杠杆。” (CFO)
• “我们预计第四家客户将在本财年超过 10% 的收入门槛，从而进一步加强收入多元化。” (CFO)
• “Credo 处于行业独特地位，是全球少数几家能够在我们目前正在认证的先进速度下提供尖端 SerDes 技术的公司之一。” (CEO)

Prepared Metrics

Q&A Batch (1-5 of 5)

Q1 — Karl Ackerman

• Topic: AEC adoption at new hyperscalers, 100G per lane, and program breadth
• Key points:
  • Recent ramps confirmed at 100 gig per lane; multiple programs in flight per customer.
  • Pull-in of orders relative to previous schedule; volumes delivered as requested.
  • Near-term scale-up opportunity is PCIe protocol (Gen5 to Gen6); AECs suitable for inter-rack and rack-to-rack.
  • Long-term (2–5 years) scale-up TAM significant; company is agnostic to 200G per lane protocol (Ethernet, UAL, NVLink, PCIe) because all use same IEEE 200G SerDes.
• Mgmt stance: Bullish — encouraged by pull-in and expanding customer engagements; sees scale-up as major TAM driver.

Q2 — Vivek Arya

• Topic: AEC market size, customer concentration, and next growth pillars
• Key points:
  • Q1 had three 10% customers (same as Q4): largest 35%, second 33%, third 20% of revenue.
  • Expect continued diversification; fourth hyperscaler expected to become a 10% customer for full FY2026.
  • Largest FY2025 customer is the largest driver of FY2026 growth.
  • AEC TAM expanding: moving from inter-rack to rack-to-rack; copper replacing optical in some connections due to reliability (1,000x fewer link flaps).
  • Next big opportunities: additional protocols, scale-up network, optical (system-level), GPU-memory connections, and inference.
• Mgmt stance: Bullish — early innings for AEC; multiple growth pillars ahead; investing R&D (more than half of OpEx) to support future waves.

Q3 — Tom O'Malley

• Topic: Optical DSP competitiveness at 1.6T (3nm platform) and PCIe retimer proliferation
• Key points:
  • 200G per lane optical DSP solution: will offer full DSP and LRO (linear receive optical); moved directly to 3nm for power efficiency (initial 1.6T transceivers >20W).
  • 1.6T market will take time to develop; product to enter production in upcoming months.
  • PCIe retimer opportunity broader than two current customers (large GPU maker + one hyperscaler); most hyperscalers expected to pursue PCIe in near term.
  • Same 200G per lane SerDes serves all protocols.
• Mgmt stance: Bullish — confident in 3nm product; PCIe retimer market significant and expanding.

Q4 — Joshua Buchalter

• Topic: LRO adoption curve (800G vs 1.6T) and customer ramp linearity
• Key points:
  • 800G LRO seeing interest; 1.6T LRO expected to become more popular due to power efficiency needs.
  • Customer ramps are not linear: surge for new generation deployment, then balance, then upward trend until next transition.
  • Diversification across customers, mediums (copper/optical), and protocols is on track.
• Mgmt stance: Neutral — acknowledges non-linear customer patterns; sees LRO success at both 800G and 1.6T.

Q5 — Tore Svanberg

• Topic: AEC use cases (row-scale, inter-rack, rack-to-rack) and FY2026 customer concentration
• Key points:
  • Inter-rack market still early; four hyperscalers adopting; expects all hyperscalers to adopt AECs eventually.
  • Rack-to-rack is expanding TAM (scale-out network near-term; scale-up network long-term, potentially 10x connections).
  • Largest FY2025 customer expected to be largest in FY2026 by far; its purchasing pattern is not consistently up-right (pauses/ramps).
  • Drivers for AEC: reliability, power efficiency, system cost vs. optical.
• Mgmt stance: Bullish — early stages for row-scale; copper preferred where possible; largest customer remains strong growth driver.

Q&A Batch (6-10 of 10)

Q6 — Christopher Rolland

• Topic: 1.6T optical supply constraints, AEC acceleration, and fifth hyperscaler timeline
• Key points:
  • Optical business seeing great results with 400G/50G per lane ramp, recent success with 800G; expects long growth with 100G per lane.
  • 1.6T transceiver adoption will take more time than industry analysts predicted; no slowing of expected growth from that delay.
  • AEC TAM stands alone; customers will use AEC if they can, supply constraints in optical won’t affect AEC adoption.
  • Fifth hyperscaler: through qualification, expects first traction toward end of this fiscal year; all hyperscalers have potential to be 10% customer.
  • Connecting to many GPUs including the “big guy” (NVIDIA); sees opportunities beyond hyperscalers in Neo Cloud and sovereign markets.
• Mgmt stance: Bullish on optical and AEC growth; neutral on 1.6T timing (delay doesn’t impact outlook); confident in hyperscaler pipeline.

Q7 — Vijay Rakesh

• Topic: Fourth CSP ramp acceleration and gross margin guidance
• Key points:
  • Accelerated ramp with fourth customer driven by qualification coming together quickly; no big change in ASPs or margins from that program.
  • Q4 to Q1 gross margin saw ~20 bps sequential increase from scale; any fluctuations due to product mix, not linear.
  • Long-term gross margin model unchanged; currently at or above model.
• Mgmt stance: Neutral on margin (short-term mix-driven, long-term model intact); positive on fourth customer ramp.

Q8 — Quinn Bolton

• Topic: (No content provided in input)
• Key points: N/A
• Mgmt stance: N/A

Q9 — Quinn Bolton

• Topic: Multiple programs at new customer and patent lawsuit settlements
• Key points:
  • Once first program is successful, it leads to next-generation planning; CRDO gets involved early in architecting next-gen racks and row designs.
  • Announced patent lawsuit settlements with Amphenol and Volex; no specifics on royalties, but expects IP to be respected.
  • Market for AECs is large enough for multiple winners; CRDO’s competitive edge is delivering next-gen solutions first, achieving qualification first, and helping customers ramp.
  • Owns all aspects of product; feels good about expanding moat.
• Mgmt stance: Bullish on customer relationship expansion and IP position; neutral on competition (market large enough for multiple winners).

Q10 — Suji Desilva

• Topic: Intra-rack vs. rack-to-rack AEC adoption and disaggregated architectures
• Key points:
  • First step is typically intra-rack (≤3 meters); traction seen over last 6–9 months as customers realize cluster reliability and power benefits.
  • Early stages of market expanding into rack-to-rack solutions; expects acceleration in AEC usage.
  • Disaggregated architectures: first uptake is in 100G per lane back-end networks for AI switch racks; front-end leaf/spine moving more slowly to next-gen speeds.
  • Speed will be a catalyst for broader adoption.
• Mgmt stance: Bullish on AEC adoption acceleration; neutral on disaggregated architecture timing (slower than expected, but speed catalyst ahead).