Marvell Technology Posts Solid Q1, Guides for Continued Growth as Custom AI Efforts Expand

Marvell Technology (MRVL) reported fiscal first-quarter results that came in just ahead of expectations, supported by strong data center demand and accelerating momentum in its custom silicon programs. The Santa Clara-based chipmaker, whose AI infrastructure products power hyperscaler clients like Amazon, Microsoft, and Meta, posted a record quarter driven by 63% year-over-year revenue growth. Despite delivering results largely in line with consensus and guiding second-quarter revenue in line, shares slipped modestly in after-hours trading—reflecting investor demand for more pronounced upside amid intense focus on AI-related semiconductors.

For the April quarter, Marvell posted adjusted earnings per share of $0.62, beating the consensus by a penny. Revenue came in at $1.895 billion, just ahead of the $1.879 billion expected. Year-over-year, adjusted EPS rose 158%, while revenue surged 63%. Marvell’s data center segment was again the star, growing 76% annually to $1.4 billion, fueled by custom AI silicon shipments and strong electro-optic product demand. GAAP gross margin was 50.3%, while non-GAAP gross margin landed at 59.8%, comfortably within guidance.

Management guided fiscal Q2 revenue to $2.0 billion, at the midpoint of analyst estimates, with adjusted EPS expected between $0.62 and $0.72 (consensus: $0.66). Gross margins are forecast to remain steady between 59% and 60% on a non-GAAP basis, with strength expected to continue from the custom silicon pipeline. CEO Matt Murphy struck an upbeat tone, noting, “As the industry continues to move toward building custom AI infrastructure, Marvell is uniquely positioned at the center of this transformation.”

Several analysts emphasized Marvell's continued progress with hyperscaler clients, particularly Amazon Web Services (AWS), where the company is deeply involved in custom chip efforts for the Trainium line of AI accelerators. Both Needham and Wolfe Research cited management’s confirmation that Marvell has secured 3nm wafer and advanced packaging capacity for the next-gen Trainium chips, slated to ramp in 2026. Despite concerns about competition from Alchip and internal AWS design teams, Marvell reiterated that its custom silicon revenue with Amazon is expected to grow in fiscal 2026 and 2027.

Additionally, Marvell is expanding its engagement with Microsoft on the Maia AI accelerator family. The company confirmed it is working on the Maia 300 generation after securing wins with Maia 200, highlighting further diversification across major cloud providers. Meanwhile, the optical networking business is holding steady and maintaining leadership during the industry’s shift to 1.6T optical speeds.

Yet, despite this encouraging pipeline, multiple firms including Needham, Deutsche Bank, and Wolfe cut their price targets, citing valuation compression across the AI sector. Revised targets now range from $85 to $90, down from prior levels around $100–$115. While analysts remain bullish on Marvell’s multi-year trajectory, short-term investor sentiment is being constrained by muted upside and ongoing competitive noise, particularly around the AWS roadmap. Still, consensus remains confident that custom silicon demand—particularly for AI-specific workloads—will drive both revenue and margin expansion over the next several years.

Looking forward, Marvell’s June 17 virtual investor event—focused on the “Future of Custom Silicon”—is emerging as a key catalyst. The company plans to unveil more technical detail and roadmap commentary around its AI chips and long-term market share goals. With the AI infrastructure build-out still in its early innings, this event could help re-anchor investor conviction around the durability and differentiation of Marvell’s solutions.

Despite its strong earnings performance, Marvell stock remains down more than 40% year-to-date, caught in the broader AI-sector selloff and high investor expectations. The company entered 2025 trading at a forward P/E above 40x but now trades closer to 21x, below its three-year average. Analysts broadly agree that Marvell’s core narrative remains intact: the AI-driven shift toward custom chips across data center customers offers a compelling growth runway. But in the near term, the stock’s trajectory may depend less on quarterly beats and more on how well Marvell communicates the sustainability of its AI partnerships and product differentiation.

In summary, Marvell's Q1 report was solid—not shiny—but it showed clear progress on its strategic priorities. Strong results in the data center segment, along with steady guidance and confidence in AI custom silicon, provide a strong base. The June 17 investor event now looms large as the next litmus test for bulls and skeptics alike. Investors should keep an eye on how Marvell addresses competitive chatter, especially regarding AWS, and whether the company can reinforce its place at the center of the AI silicon revolution.