Credo's Q1 2026 Earnings Call: Contradictions Emerge on Optical Dsp Market, Aecs, Customer Concentration, and Supply Chain Constraints

The above is the analysis of the conflicting points in this earnings call

Date of Call: None provided

Financials Results

• Revenue: $223.1M, up 31% sequentially and up 274% YOY
• Gross Margin: Non-GAAP gross margin 67.6%, up 20 bps sequentially; product GM 66.7%, up 12 bps sequentially and up ~517 bps YOY
• Operating Margin: 43.1%, compared to 36.8% in the prior quarter (+635 bps sequentially)

Guidance:

• Q2 revenue expected at $230M–$240M (~+5% q/q at midpoint)
• Q2 non-GAAP gross margin: 64%–66%
• Q2 non-GAAP operating expenses: $56M–$58M
• Q2 diluted share count: ~190M
• FY26: mid-single-digit sequential growth each quarter; ~120% YOY revenue growth
• FY26 non-GAAP OpEx to rise <50% YOY
• Non-GAAP net margin ~40% in coming quarters and for FY26
• Top 3 customers to grow significantly YOY; a 4th to exceed 10% of revenue for FY26
• Guidance based on current (fluid) tariff regime

Business Commentary:

• Revenue and Profitability Growth:
• Credo reported revenue of $223 million for Q1, marking a 31% sequential increase and 274% year-over-year growth.

• This growth was driven by strong demand for Credo's reliable high-speed connectivity solutions, supported by strategic partnerships with hyperscalers and increased investment in AI-driven infrastructure.

• ADC Product Line Growth:

• Credo's ADC product line continued its robust growth with three hyperscalers each contributing over 10% of revenue.

• The growth was fueled by a diverse customer base and significant year-over-year growth forecasts.

• Optical Business Momentum:

• The optical business sustained strong momentum, with the expectation to double optical revenue in fiscal 2026.

• This was attributed to delivering cutting-edge DSP solutions that meet the rigorous connectivity demands of next-generation applications.

• Pioneering AEC Market Expansion:

• Credo's innovation in active electrical cables (AECs) expanded with customer interest shifting from inter-rack to rack-to-rack solutions.
• The growth is driven by reliability, power efficiency, and system cost advantages over optical solutions, with all hyperscalers expected to adopt AECs eventually.

Sentiment Analysis:

• Management reported record revenue and profitability, with revenue up 31% q/q and 274% YOY, non-GAAP gross margin at 67.6%, and non-GAAP net income up 51% q/q. Q2 outlook calls for continued sequential growth and FY26 revenue up ~120% YOY with ~40% non-GAAP net margin. AEC adoption broadened, three customers >10% of revenue, and a fourth hyperscaler ramped ahead of plan.

Q&A:

• Question from Karl Ackerman (BNP Paribas): Are new hyperscaler ramps using 100G/lane AECs and do you have multiple programs per customer?
  Response: Recent ramps are at 100G/lane; engagements typically start with one platform and expand, and orders were pulled in with Credo delivering requested volumes.

• Question from Karl Ackerman (BNP Paribas): For scale-up interconnects (PCIe vs Ethernet/UAL/NVLink), where is AEC advantage and which market is larger?
  Response: Near term favors PCIe Gen6; Credo is protocol-agnostic as all use 200G SerDes, so it can serve any path; wants faster 200G/lane adoption to expand TAM.

• Question from Vivek Arya (Bank of America): Who were the 10% customers and how big were they; is AEC cannibalizing copper or optical?
  Response: Same three as last quarter at 35%, 33%, and 20% of revenue; AEC TAM is expanding with copper increasingly replacing optical in some rack-to-rack due to reliability and power.

• Question from Vivek Arya (Bank of America): What’s the next big growth driver and are you investing enough beyond AECs?
  Response: Building pillars in optical DSP/LRO, PCIe scale-up, and system-level solutions (incl. memory links); >50% of R&D spend is on optical; maintaining operating leverage.

• Question from Tom O’Malley (Barclays): Competitiveness at 1.6T optical and 3nm timing vs 800G?
  Response: Moved to 3nm for power; 200G/lane full DSP and LRO will enter production soon; expects 1.6T transceiver market to develop gradually.

• Question from Tom O’Malley (Barclays): Will PCIe retimer adoption broaden beyond the initial GPU vendor and one hyperscaler?
  Response: Yes; expects broad hyperscaler adoption near term; future protocol divergence is fine since Credo supports any 200G/lane protocol.

• Question from Joshua Buchholter (TD Cowen): How will LRO adoption ramp at 800G vs 1.6T?
  Response: Sees interest at 800G and stronger pull at 1.6T as power becomes critical; Credo’s LRO offers significant power savings with required performance.

• Question from Joshua Buchholter (TD Cowen): You noted non-linear customer ramps; how is visibility as the base broadens?
  Response: Individual customer ramps are non-linear around transitions, but diversification across customers/mediums/protocols is improving overall stability and visibility.

• Question from Tore Svanberg (Stifel): Use cases and maturity for row-scale (rack-to-rack) AEC networking?
  Response: AEC TAM is expanding from intra-rack to rack-to-rack, driven by reliability and power; still early, and copper is preferred wherever feasible.

• Question from Tore Svanberg (Stifel): Largest customer outlook given recent softness?
  Response: Purchases are inherently non-linear; the largest FY25 customer is expected to again be the largest in FY26 by a wide margin.

• Question from Christopher Rolland (Susquehanna): Do 1.6T optical/laser supply constraints affect optical DSP outlook or accelerate AEC?
  Response: 1.6T transition likely slower than forecasts but doesn’t change optical outlook; AEC adoption is independent—customers use AEC wherever possible.

• Question from Christopher Rolland (Susquehanna): Timing for a fifth hyperscaler and engagement with the major GPU vendor?
  Response: Fifth hyperscaler expected to show first traction by fiscal year-end; Credo is already connecting to NVIDIA GPUs and sees broader neo-cloud/sovereign opportunities.

• Question from Vijay Rakesh (Mizuho): What drove the faster-than-expected ramp at the fourth CSP; any ASP/margin change?
  Response: Accelerated qualification and deployment timing; no material change to ASPs or margins.

• Question from Vijay Rakesh (Mizuho): What’s behind the Q2 gross margin guide—tariffs or mix?
  Response: At current scale, quarterly GM variance is mainly mix; long-term GM model around 60% remains intact despite current levels above that.

• Question from Quinn Bolton (Needham & Company): Is the fourth hyperscaler program intra-rack or rack-to-rack, and will more programs follow?
  Response: Not a single-rack architecture; expects multiple follow-on programs with early co-architecture on next-gen racks and rows.

• Question from Quinn Bolton (Needham & Company): Do patent settlements (e.g., Amphenol, Volex) imply licensing revenue and more competition?
  Response: Credo expects its IP to be respected; settlements align with that. Market is large for multiple suppliers; Credo’s moat is full-stack ownership and first-to-qualify execution.

• Question from Suji Desilva (ROTH Capital): How broadly are intra-rack AECs adopted across your five customers?
  Response: Most start with intra-rack (≤3m); rack-to-rack usage is early but accelerating as customers prioritize reliability and power.

• Question from Suji Desilva (ROTH Capital): Why have disaggregated switch architectures been slower and what could catalyze them?
  Response: Front-end leaf/spine transitions lag; speeding to 100G/lane and beyond should catalyze; backend AI switch racks at 100G/lane are already ramping.