Everyone is focusing on Nvidia's chips, but Jensen Huang has already developed a "second pillar"

While the market's attention is focused on Nvidia's AI chips, Jensen Huang has quietly built a "second pillar" worth tens of billions of dollars. This data center networking business, established through a strategic acquisition in 2020, has now become one of Nvidia's most profitable and fastest-growing divisions.

In just a few years, the networking business aimed at connecting data centers has grown to become Nvidia's second largest revenue driver after its computing business. According to Nvidia's latest financial report, this division generated $11 billion in revenue last quarter, a staggering 267% year-on-year increase, and annual revenue surpassed $31 billion.

This astonishing growth rate and scale have directly reshaped the competitive landscape of the network equipment market. Kevin Cook, Senior Equity Strategist at Zacks Investment Research, pointed out that Nvidia's single-quarter networking revenue has already surpassed the annual networking revenue projections of long-standing network giant Cisco.

The Cornerstone of the “AI Factory”

The explosion of Nvidia's networking business directly benefits from the surging demand for AI processing. The division’s technology matrix includes NVLink, which connects GPU racks in data centers; the in-network computing platform InfiniBand Switches; the AI networking Ethernet platform Spectrum-X; and co-packaged optical switches, among others.

Together, these technologies provide all the necessary infrastructure to build “AI factories”—data centers dedicated to training AI models.

Kevin Deierling, Nvidia’s Senior Vice President of Networking, said: “People usually think of networking as ‘I have a printer and need to connect it.’ But from the very first day of the acquisition, Jensen Huang said that the data center is the new computing unit. Networking isn’t just about moving small amounts of data between compute nodes; it is, in fact, the foundation.”

The Vision Behind the $7 Billion Acquisition

The starting point for this massive business was Nvidia’s $7 billion acquisition of Israeli networking company Mellanox in 2020. Deierling joined Nvidia through this acquisition.

Initially, Deierling did not fully understand why Jensen Huang bought Mellanox at that time, but now he gets it. Possessing the networking business enabled Nvidia to bundle GPUs with the most compatible networking technologies for sale.

“When Jensen Huang acquired Mellanox in 2020, he saw this as the missing piece to make GPUs a complete solution,” said analyst Cook.

In addition to its technological leadership, the success of Nvidia’s networking business also owes to its unique business model. Deierling noted that Nvidia only sells these technologies as part of a full-stack solution, not as individual components, and brings them to market through partners.

“I can’t think of any other company that has the same full-stack capabilities we do,” said Deierling. “We’ve built the complete compute stack, the fully integrated stack, and then taken it to market with all of our partners.”

At the Nvidia GTC conference on March 16, Nvidia further consolidated this advantage by launching a series of networking system updates, including the Rubin platform (featuring six new chips), the Inference Context Memory Storage platform, and the more efficient Spectrum-X Ethernet Photonics switches.

“Networking is no longer an external device that connects printers or other slow I/O devices,” Deierling emphasized. “It’s the foundation of the computer. In the past, a computer had a backplane inside. Today, networking is the backplane of the AI factory—it's incredibly important.”