Marvell Technology (MRVL) Stock News

Marvell Technology (NASDAQ: MRVL) has unveiled its custom Ultra Accelerator Link (UALink) scale-up solution, expanding its AI compute platform capabilities. The solution includes comprehensive IPs featuring 224G SerDes, UALink Physical Layer IP, Configurable Controller IP, Scalable Switch Core, and advanced packaging options. This open standards-based technology enables efficient scaling of AI infrastructure, supporting hundreds to thousands of AI accelerators in scale-up deployments. The solution addresses hyperscalers' challenges in scaling AI infrastructure while maintaining high performance and low latency. As a member of the UALink Consortium, Marvell is contributing to developing open specifications for accelerator connectivity, with support from industry leaders like AMD for advancing scale-up networks.

Marvell Technology (NASDAQ: MRVL) reported strong Q1 FY2026 results with record revenue of $1.895 billion, representing a 63% year-over-year increase. The company posted GAAP net income of $177.9 million ($0.20 per diluted share) and non-GAAP net income of $540.0 million ($0.62 per diluted share). Cash flow from operations was $332.9 million. The growth was primarily driven by strong AI demand in the data center market, benefiting from custom silicon programs and electro-optics products. For Q2 FY2026, Marvell forecasts revenue of $2.000 billion (±5%) with expected non-GAAP diluted EPS of $0.67 (±$0.05). The company plans to host a Custom AI Investor Event on June 17, 2025, to showcase technological advances and long-term market share goals in custom AI infrastructure.

Marvell Technology (NASDAQ: MRVL) has unveiled an innovative multi-die packaging platform for custom AI accelerators, marking a significant advancement in AI infrastructure semiconductor solutions. The platform enables 2.8x larger multi-chip accelerator designs compared to conventional single-die implementations. Key features include a novel RDL interposer that integrates 1390 mm² of silicon and four HBM3/3E memory stacks. The solution offers improved die-to-die interconnect efficiency, reduced power consumption, enhanced chiplet yields, and lower product costs. Already production-qualified with a major hyperscaler, the platform supports current HBM3/HBM3E memory and is being qualified for future HBM4 designs. Notably, Marvell is collaborating with all four top hyperscalers to develop custom XPUs, CPUs, and other devices for cloud and AI infrastructure optimization.

Marvell Technology (NASDAQ: MRVL) has unveiled an innovative multi-die packaging platform for custom AI accelerators. The solution enables 2.8x larger multi-chip accelerator designs compared to conventional single-die implementations, featuring a novel RDL interposer that integrates 1390 mm² of silicon and four HBM3/3E memory stacks. The platform has been qualified with a major hyperscaler and is entering production.

Key advantages include more efficient die-to-die interconnect, lower power consumption, increased chiplet yields, and reduced total cost of ownership. The modular RDL interposer design offers an alternative to traditional silicon interposers, providing supply chain flexibility. Marvell is currently collaborating with three of the top four hyperscalers to develop custom XPUs and CPUs for clouds and AI clusters.

Marvell Technology (NASDAQ: MRVL) has announced a strategic collaboration with NVIDIA to integrate NVIDIA's NVLink Fusion technology into Marvell's custom cloud platform silicon solutions. NVLink Fusion offers up to 1.8TB/s of bidirectional bandwidth and enables seamless integration of custom XPU silicon with NVIDIA's rack-scale systems. The partnership leverages Marvell's comprehensive semiconductor expertise, including SerDes, die-to-die interconnects, silicon photonics, and advanced packaging. This collaboration aims to provide hyperscalers with accelerated deployment of custom scale-up solutions for AI model training and inference, while allowing them to leverage existing NVLink architecture investments. The integration will enable cloud providers to scale out AI factories to millions of custom XPUs efficiently.

Marvell Technology (NASDAQ: MRVL) has announced a webinar focused on custom silicon technology for AI infrastructure, scheduled for June 17, 2025, from 10:00 am to 12:00 pm PT. The company has postponed its previously scheduled Investor Day from June 10, 2025, to 2026 due to macroeconomic uncertainty.

The company has also reaffirmed and narrowed its fiscal Q1 2026 revenue guidance to approximately $1.875 billion (±2%), compared to the previous range of ±5%. The webinar will coincide with Marvell's annual internal senior technical leadership conference and will feature presentations on AI technology advances, followed by a Q&A session.

Marvell Technology (NASDAQ: MRVL) announced significant changes to its Board of Directors. Michael Strachan, Lead Independent Director and Audit Committee Chair, and Robert Switz, Executive Compensation Committee Chair, will not seek reelection at the Annual Meeting on June 13, 2025. Both directors have served since 2016 and played key roles in Marvell's transformation. Brad Buss, a director since 2018, will become the new Lead Independent Director effective June 13, 2025.

The company has nominated eight current directors for reelection, including Sara Andrews, Tudor Brown, Brad Buss, Daniel Durn, Rebecca House, Marachel Knight, Matt Murphy, and Rick Wallace.

Marvell Technology (NASDAQ: MRVL) has announced successful interoperability testing of its Structera Compute Express Link (CXL) portfolio with both AMD EPYC CPUs and 5th Gen Intel Xeon platforms. The testing demonstrated seamless integration across various configurations and workloads.

The Structera product family, built on CXL 2.0, includes two key products:

• Structera A: Near-memory accelerators with Arm Neoverse V2 cores and DDR5 support
• Structera X: Memory expansion controllers supporting DDR4 and DDR5 DIMMs

This interoperability enables cloud data center providers to increase memory capacity and bandwidth, add computing cores, and optimize resource utilization. The technology supports memory-intensive applications like AI, machine learning, and high-performance computing while promoting sustainable data center architectures.

Infineon Technologies AG (OTCQX: IFNNY) has announced the acquisition of Marvell Technology's Automotive Ethernet business for US$2.5 billion in an all-cash transaction. The deal strengthens Infineon's position as the global leader in automotive semiconductors and enhances its capabilities for software-defined vehicles.

The acquired business is projected to generate revenue of US$225-250 million in 2025 with a gross margin of approximately 60%. Marvell's Brightlane™ Automotive Ethernet portfolio includes PHY transceivers, switches, and bridges supporting network data rates up to 10 Gbps. The business serves over 50 automotive manufacturers and has a design-win pipeline of around US$4 billion until 2030.

The transaction will be financed through existing liquidity and additional debt, with closing expected within calendar year 2025, subject to regulatory approvals.