Nvidia’s Networking Business: A Hidden AI Powerhouse
Nvidia CEO Jensen Huang demonstrated remarkable foresight when he initiated the development of AI-specific chips back in 2010, well before the current surge in artificial intelligence. A similarly strategic move in 2020 – a significant investment in data center networking through a key acquisition – has quietly become one of the company’s most profitable and rapidly expanding divisions. This growth is a key piece of the latest artificial intelligence news.
From Under the Radar to Revenue Leader
In just a few years, Nvidia’s networking business, designed to connect data centers, has evolved into the company’s second-largest revenue source, trailing only its compute division. Last quarter alone, it generated $11 billion in revenue, a staggering 267% year-over-year increase. For the full year, the division brought in over $31 billion, according to Nvidia’s recent earnings report. This impressive performance is fueled by the increasing demand for AI processing power.
The division encompasses technologies like NVLink, which facilitates communication between GPUs within a data center rack; Nvidia InfiniBand Switches, an in-network computing platform; Spectrum-X, an ethernet platform optimized for AI networking; and co-packaged optics switches. Collectively, these technologies provide everything needed to construct an “AI factory” – a data center specifically engineered for training AI models.
Outpacing the Competition
Kevin Cook, a senior equity strategist at Zacks Investment Research, highlights the exceptional nature of Nvidia’s networking business. “Nvidia’s networking business reported $11 billion for the quarter; that number is greater than Cisco’s networking business, almost as big as the full-year estimates,” Cook stated, adding that it achieves in one quarter what Cisco’s business accomplishes in an entire year. Despite this remarkable success, the segment receives less attention than Nvidia’s core chip business, which is considerably larger, or its original gaming business.
The Mellanox Acquisition: A Strategic Masterstroke
The foundation of Nvidia’s networking business lies in Mellanox, a networking company founded in Israel in 1999. Nvidia acquired Mellanox in 2020 for $7 billion. Kevin Deierling, a senior vice president of networking at Nvidia who joined the company through the acquisition, initially questioned Huang’s decision but now fully understands its strategic importance.
“Jensen said the first day when he acquired us, he said the data center is the new unit of computing. Networking is a lot more than just moving the smaller amounts of data between a compute node; it’s actually a foundation,” Deierling explained. Integrating a networking business with its GPU business allows Nvidia to offer a complete, optimized solution to its customers.
Cook adds, “When Jensen bought Mellanox in 2020, he saw that was the missing piece to make GPUs a complete package.” Deierling emphasizes that Nvidia’s approach of providing a full-stack solution, rather than individual components, and selling through partners, sets it apart from competitors.
Recent Innovations and the Future of AI Networking
Nvidia recently unveiled a series of updates to its networking system during Huang’s keynote at the Nvidia GTC technology conference on March 16. These included the launch of the Nvidia Rubin platform, featuring six new chips to power an “AI supercomputer,” as well as the Nvidia Inference Context Memory Storage platform and more efficient Nvidia Spectrum-X Ethernet Photonics switches.
“It’s no longer a peripheral to connect the printer, some other slow I/O device,” Deierling stated. “It’s fundamental to the computer. In the old days, we had what was called the back lining inside the computer. Today, the network is the back lining of the AI factory, and it’s super important.” This is a significant development in artificial intelligence news and the future of AI infrastructure.
Further Reading: Nvidia Official Website
Source: TechCrunch
