Nvidia’s networking business is now a multibillion-dollar engine

Nvidia’s networking business has grown to become the company’s second-largest revenue driver, positioned behind its compute segment. The division’s financial performance, detailed in Nvidia’s most recent earnings, highlights this growth:

• Last quarter, the networking business reported $11 billion in revenue.
• This quarterly figure represents a 267% year-over-year increase.
• For the full year, the division reported $31 billion in revenue.

The focus on networking mirrors Nvidia’s earlier long-term bets. In 2010, the company began building AI-specific chips, a decade before the current market demand. Similarly, the foundation of its networking business segment dates back to 2020, when Nvidia acquired Mellanox for $7 billion. Mellanox, a networking company founded in Israel in 1999, provided the technology that now underpins Nvidia’s networking portfolio. This portfolio includes NVLink, which is GPU communication technology; InfiniBand, a networking platform; and Spectrum-X, an Ethernet platform. Together, these technologies allow Nvidia to package its graphics processing units (GPUs) with the networking infrastructure they work best with, enabling customers to purchase the technology as a full-stack, integrated solution. Kevin Cook, a senior equity strategist at Zacks Investment Research, noted that when Nvidia acquired Mellanox in 2020, the technology represented the missing piece to make its GPUs a complete package.

According to Cook, Nvidia’s networking business is one of the most impressive new segments from the company. Cook noted that the division does in one quarter what competitor Cisco’s networking business does in a year.

Nvidia recently updated its networking systems during CEO Jensen Huang’s keynote address on March 16 at the company’s annual Nvidia GTC technology conference. Kevin Deierling, the senior vice president of networking at Nvidia, explained that networking has evolved beyond legacy perceptions of simply connecting basic devices like printers. Instead, it serves as the foundation for the “AI factory”—defined as a data center designed for training AI models. Deierling noted, “It’s no longer a peripheral to connect the printer, some other slow I/O device. It’s fundamental to the computer. In the old days, we had what was called the backplane inside the computer. Today, the network is the back lining of the AI factory, and it’s super important.”

Why it matters

Nvidia’s networking division has grown into its second-largest revenue driver, providing the essential infrastructure for AI factories and allowing the company to sell its GPUs as part of a full-stack solution.