Marvell Technology, Inc. (NASDAQ: MRVL) Q2 2026 Earnings Call | 08/29/2025

Financial and Operational Highlights

• Data Center Performance: Data center revenue flat in Q3, with optics growing double digits. AI revenue (optics + custom silicon) on track to exceed 50% of total company revenue by fiscal year-end (January 2026), though exact timing (Q3 or Q4) not specified. Q4 FY25 mix: optics ~50%, custom silicon ~25%, other ~25%.

• Earnings Growth: Q2 EPS up 123% YoY; Q3 EPS guidance projects ~70% YoY growth, outpacing revenue due to operating leverage.

• Core Business Recovery: Enterprise networking and carrier segments expected to grow ~30% sequentially and >80% YoY in Q3, signaling a strong rebound.

• Capital Allocation: Closed automotive business divestiture to Infineon in August 2025 (ahead of year-end target) at an attractive valuation. Proceeds to support organic AI investments, potential tuck-in acquisitions, and increased share buybacks, driven by consistent free cash flow execution.

Strategic Initiatives

• AI and Custom Silicon: Secured 18+ design wins for custom silicon and XPU attach, with several in production since late 2024 and others ramping through 2026. Targeting 20% of a $94 billion TAM by 2028. Strong momentum across hyperscalers and emerging players, with a $75 billion pipeline expected to grow.

• Optics Business: Achieved significant growth over recent years, with Q3 guidance for double-digit growth. Strong execution and deep partnerships with module companies mitigate supply chain constraints (e.g., laser shortages). Marvell’s DSP-based pluggable modules dominate, with linear pluggable optics (LPO) as a niche but active use case.

• Scale-Up Switch Fabric: Investing in UA-Link and Ethernet-based products for XPU attach, leveraging low-latency switching IPs and SerDes. Product introductions expected within two years (by 2027), with near-term revenue from active electrical/optical cables (AECs/AOCs).

• R&D Focus: Over 80% of R&D now allocated to AI and data center (up from ~60% a few years ago), reflecting a strategic shift from consumer/enterprise to AI/data center leadership.

Market and Competitive Dynamics

• Supply Chain: Optics business insulated from ecosystem constraints due to robust planning and partnerships. No significant stops/starts in product ramps, despite industry noise around component shortages.

• 3nm XPU Programs: Lead customer’s 3nm XPU follow-on program remains on track for production in calendar 2026, with secured wafer and packaging capacity. Third 3nm XPU customer win also tracking for production in H2 2026. Marvell avoids commenting on competitive claims to minimize market noise, focusing on execution and market share goals.

• Market Opportunity: AI pipeline expansion driven by hyperscalers and emerging players. Strong demand for scale-up switch fabric and interconnect solutions, with Marvell well-positioned due to its IP portfolio and ecosystem engagement.

Q&A Highlights

• AI Revenue Mix: Optics and custom silicon growing as a share of revenue, but no quarterly mix updates provided beyond Q4 FY25 baseline (~50% optics, ~25% custom). AI expected to surpass 50% of revenue by fiscal year-end.

• Capital Allocation Strategy: Post-automotive divestiture, Marvell prioritizes organic AI investments and opportunistic buybacks. Tuck-in acquisitions considered to accelerate AI roadmap, but no plans for major M&A or further business sales (e.g., carrier, consumer) unless strategically compelling.

• Switch Fabric Timeline: UA-Link and Ethernet-based scale-up products slated for introduction by 2027, with AECs/AOCs driving near-term revenue. Combines Marvell’s switching and SerDes strengths.

• LPO Adoption: LPO modules are a small, niche segment with active wins, but DSP-based pluggables remain dominant for the foreseeable future due to scale and reliability.
  
  

Closing:

Marvell reported strong Q2 2026 performance, with significant EPS growth and a robust AI-driven outlook. The company is capitalizing on its leadership in optics and custom silicon, with 18+ design wins and a growing $75 billion AI pipeline. The core business rebound, disciplined capital allocation (including automotive divestiture proceeds), and focus on AI/data center (80%+ of R&D) position Marvell to capture a significant share of the $94 billion AI TAM by 2028.