Key Highlights
- Nvidia announced a $2 billion equity investment in Marvell Technology alongside a strategic collaboration
- The partnership brings Marvell’s custom XPU chip designs into Nvidia’s expanding AI infrastructure ecosystem
- Shares of Marvell surged up to 11% during premarket hours; Nvidia gained 1.6%
- The collaboration leverages Nvidia’s NVLink Fusion technology to bridge both companies’ hardware offerings
- Marvell specializes in custom AI accelerators known as XPUs, serving major cloud providers like Amazon
In a major strategic move announced Tuesday morning, Nvidia revealed it has acquired a $2 billion equity position in Marvell Technology. The transaction includes a comprehensive partnership that integrates Marvell into Nvidia’s rapidly expanding AI ecosystem.
This collaboration enables enterprises to deploy hardware from both organizations to create what the companies describe as “semi-custom AI infrastructure.” Essentially, customers gain flexibility to combine Marvell’s specialized chip architectures with Nvidia’s processing units and networking solutions.
Nvidia’s CEO Jensen Huang characterized the move as a strategic expansion. “Together with Marvell, we are enabling customers to leverage NVIDIA’s AI infrastructure ecosystem and scale to build specialized AI compute,” Huang noted in an official statement.
Marvell Technology, Inc., MRVL
Marvell has built its reputation developing custom AI accelerators—referred to as XPUs—primarily for hyperscale cloud operators such as Amazon. These processors have traditionally served as alternatives to Nvidia’s GPU offerings, making this partnership something of a strategic shift.
Instead of positioning Marvell as competition, Nvidia has chosen to embrace it. The reasoning is straightforward: since enterprises are deploying XPUs regardless, Nvidia prefers having its networking and processing technology integrated into those configurations.
NVLink Fusion Powers the Partnership
The technical foundation enabling this collaboration is Nvidia’s NVLink Fusion infrastructure, which debuted last year. This platform enables third-party silicon—including Marvell’s XPUs—to interface seamlessly with Nvidia’s processing cores and networking equipment.
This capability unlocks fresh opportunities for Nvidia beyond its traditional GPU-focused offerings. For Marvell, the arrangement means its specialized chip architectures can now be marketed as components of a comprehensive, Nvidia-compatible AI solution.
Marvell stock skyrocketed as high as 11% in premarket activity following the disclosure. Shares were recently trading up approximately 8.6% at $95.28. Nvidia climbed 1.6%.
The $2 billion transaction represents a pure equity stake—Nvidia becomes a shareholder, and Marvell enters its ecosystem. This is not an acquisition.
Marvell had experienced challenging trading conditions recently. Shares had declined roughly 7.45% in sessions prior to Tuesday’s announcement. The Nvidia partnership provided substantial upward momentum.
Amazon Relationship Adds Strategic Layer
Marvell’s client roster includes Amazon as one of its most significant custom chip partners. Amazon has been developing proprietary processors—the Trainium and Inferentia product families—with engineering support from Marvell.
This existing relationship now operates within a Nvidia-compatible framework, potentially enhancing the appeal of Marvell’s solutions to additional cloud infrastructure providers seeking adaptable AI hardware configurations.
Nvidia shares advanced 1.6% in premarket sessions. While the $2 billion investment represents a relatively modest outlay for a corporation of Nvidia’s scale, the strategic implications—particularly expanding NVLink Fusion adoption—matter considerably more than the transaction value itself.
Marvell shares had been trading significantly below their 52-week peak levels prior to Tuesday’s announcement.


