NVIDIA Actually Invested in Its Competitor



NVIDIA invested $2 billion in $MRVL

In fact, what NVIDIA is interested in is not Marvell's chip business, but its core capabilities in AI infrastructure.

Marvell has experienced rapid growth in recent years, with data center operations becoming the company's biggest growth engine. Instead of simply selling chips, Marvell is better at helping cloud giants build complete AI infrastructure systems, including custom chips, high-speed networks, memory expansion, smart NICs, and optical interconnect technology.

Today, the biggest bottleneck in the AI industry is no longer computing power, but data transfer speed. GPUs are becoming more powerful, but if data transfer can't keep up, even the strongest computing power will be wasted. Therefore, the core of future AI competition will shift from "whose chips are stronger" to "who has faster data transfer."

This is also why Marvell is favored by the market. The company is focusing on silicon photonics, photonic interconnects, and co-packaged optics (CPO), driving AI infrastructure from traditional electrical signal transmission toward the era of optical communication. In the future, data transfer between chips will increasingly rely on light rather than copper wires.

The real purpose of NVIDIA's investment this time is to deeply integrate Marvell's custom chips and optical interconnect capabilities with its own NVLink ecosystem, jointly addressing bandwidth and energy efficiency issues in future AI system expansion.

More importantly, Jensen Huang sees a bigger change.

In the past, software companies made money from code, enjoying high profit margins. In the AI era, every content generation consumes massive computing power, network bandwidth, and electricity resources. Profits that once belonged to the software industry are now flowing into GPUs, data centers, network equipment, and infrastructure.

In other words, the biggest opportunities in the future may not lie in AI applications, but in the underlying infrastructure that supports AI operation.

Therefore, NVIDIA's investment in Marvell is not about cultivating a competitor, but betting on the next generation of global AI infrastructure.

Whoever masters chips, networks, optical communications, and data centers in the future will have the opportunity to become the biggest beneficiaries of the AI era.
View Original
post-image
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • 1
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pinned