Nvidia Earnings: Is Its Fastest-Growing Business Not Chips?

The Hidden Growth Engine Inside Data Centers

While Wall Street fixes its attention squarely on Nvidia’s graphics processing units, a faster-growing operational pillar quietly expands right beside them. According to Bloomberg and Yahoo Finance, compute still dwarfs networking inside the company’s Data Center business, but networking scales at a much sharper trajectory. Networking grew nearly 200% year over year during the last reported quarter, comfortably outpacing the 77% growth rate recorded by compute.

That momentum shows signs of normalization, though dominance remains. Analysts tracking the sector expect networking growth to register roughly 134% in the upcoming earnings report, according to Bloomberg and Yahoo Finance data. That projected expansion still easily beats compute, which analysts anticipate will grow by about 100% over the same timeframe.

Solving the Bottleneck in AI Infrastructure

This rapid shift addresses a fundamental engineering hurdle confronting the modern artificial intelligence industry. Major tech companies like Alphabet, Amazon, Microsoft, and Meta Platforms spend billions of dollars developing or commissioning custom processors capable of handling some workloads without an Nvidia GPU. Yet, those custom chips still need to communicate efficiently across massive server clusters.

When artificial intelligence data centers link together thousands of accelerators to form massive computing networks, transferring information between those distinct processors presents a significant challenge. Jessica Inskip, director of investor research at StockBrokers.com, highlighted this shift during an appearance on Yahoo Finance’s Opening Bid, describing Nvidia’s strategy as “investing in the bottlenecks that are the AI infrastructure.” Networking represents a primary bottleneck.

Selling the Plumbing and Rack-Scale Systems

Rather than merely shipping standalone chips, Nvidia increasingly supplies the plumbing that keeps huge AI systems running. The firm relies on its Spectrum-X line to link servers via high-speed Ethernet, whereas InfiniBand and NVLink facilitate rapid transmission of massive information packets among processors to prevent expensive hardware from going underutilized.

Nvidia has also broadened its technological reach through tools like NVLink Fusion, providing businesses the capability to integrate proprietary custom processors directly into Nvidia’s wider framework of high-speed links, networking infrastructure, electricity delivery, and thermal management. Instead of selling the chips one by one, this approach designs an entire server rack to work like one giant computer.

Market Resilience and Stock Performance

This diversification strategy alters how public markets price the company against traditional semiconductor peers. According to Yahoo Finance AlphaSpace coverage, Nvidia shares have traded sideways over roughly the past two months, while semiconductors fell into a bear market.

Ever since the semiconductor sector reached its peak near June 22, the value of Nvidia shares has remained virtually stagnant, whereas the iShares Semiconductor ETF (SOXX) has experienced a decline of approximately 20%. Even though networking operations alone do not fully explain this market strength, their swift expansion provides market participants with an additional rationale for perceiving Nvidia as something beyond a standard chip manufacturer.

What to Watch in Upcoming Earnings

When Nvidia reports earnings on Wednesday, Wall Street expects that figure to approach $17 billion. That figure keeps the business at roughly one-fifth of Data Center revenue.

Financial analysts and stakeholders will be paying close attention to Jensen Huang’s remarks concerning networking demand, proprietary silicon, and the extent to which Nvidia can monetize the broader AI ecosystem outside of just GPUs. Summarizing that overarching corporate ambition, Jessica Inskip noted on Yahoo Finance that Nvidia, she said, is trying to become a “one-stop shop for everything” in AI.

Frequently Asked Questions