NVIDIA Investment in Marvell Deepens AI Tie
NVIDIA investment in Marvell ties the chipmaker into NVLink Fusion and AI-RAN, likely prompting investors to reprice Marvell's AI-infrastructure exposure.
KEY TAKEAWAYS
- NVIDIA invested $2.0 billion to fold Marvell into its AI factory and AI-RAN ecosystems.
- Marvell will supply custom XPUs and NVLink Fusion-compatible scale-up networking for rack-scale AI.
- Deal signals deeper commercial alignment reshaping Marvell's role in NVIDIA's rack-scale AI hardware.
HIGH POTENTIAL TRADES SENT DIRECTLY TO YOUR INBOX
Add your email to receive our free daily newsletter. No spam, unsubscribe anytime.
NVIDIA said in a press release on March 31, 2026, that its $2 billion investment in Marvell Technology, Inc. (MRVL) will integrate Marvell into NVIDIA’s AI factory and AI-RAN ecosystems through NVLink Fusion, accelerating joint AI and telecom infrastructure development.
Strategic Partnership and Integration
The partnership connects Marvell to NVIDIA’s AI factory and AI-RAN ecosystem via NVLink Fusion, a rack-scale platform for semi-custom AI infrastructure. The companies expect this collaboration to offer customers broader choices in AI infrastructure but noted that benefits depend on execution, market conditions, and potential regulatory or litigation risks.
NVIDIA founder and CEO Jensen Huang said, "The inference inflection has arrived," highlighting surging demand for AI inference and the race to build AI factories. This investment signals a deeper commercial alignment, folding Marvell into NVIDIA’s AI ecosystem and reshaping Marvell’s role in rack-scale AI hardware. By linking Marvell’s roadmap for custom accelerators and scale-up networking with NVIDIA’s architecture, the companies position Marvell to help assemble semi-custom AI systems.
Marvell will supply custom XPUs—accelerator processing units—and NVLink Fusion-compatible scale-up networking designed for rack-scale AI deployments. NVIDIA will provide its Vera CPUs, ConnectX network interface cards, BlueField data processing units, NVLink interconnects, and Spectrum-X switches to support the rack-scale compute architecture. The partners will also collaborate on silicon photonics and advanced optical interconnects to enhance high-speed AI networking.
The collaboration extends to NVIDIA Aerial AI-RAN, aiming to integrate AI infrastructure into telecom networks for 5G and 6G applications. This effort targets transforming carrier networks into AI-enabled platforms.
The companies cautioned that the partnership’s success depends on execution and regulatory approvals, referencing Marvell’s SEC filings for further details.
