
Nvidia's data center networking business generated a record $14.8 billion in revenue in its latest quarter, up 199% year over year, and has grown from $8.6 billion in fiscal 2024 to $31.4 billion in fiscal 2026.
As AI systems become larger and more communication-intensive, networking infrastructure is becoming as critical as raw computing power, allowing Nvidia to sell more technology into each data center regardless of whether customers use Nvidia's own AI chips.
If the company maintains its valuation multiple while earnings growth continues, Nvidia's market cap could reach $7.5 trillion.
What happened
Nvidia generated $14.8 billion in data center networking revenue in fiscal Q1 2027 (ended April 26), up 199% year over year, with networking revenue growing from $8.6 billion in fiscal 2024 to $13 billion in fiscal 2025 and $31.4 billion in fiscal 2026. The company captured 21.5% of data center Ethernet switching revenue in Q1 2026, making it the market leader ahead of Arista Networks at 20.7%.
Why it matters
As AI systems scale, high-performance networking has become critical infrastructure—without it, communication can bottleneck GPU utilization and slow AI workloads. Nvidia's networking technology (NVLink, Spectrum-X Ethernet, ConnectX adapters, BlueField data processing units) allows the company to benefit from AI infrastructure spending even when customers choose competing custom chips over Nvidia GPUs, and partnerships like the expanded one with Marvell Technology let Nvidia supply networking alongside third-party accelerators.
What to watch
If Nvidia maintains its current forward valuation (around 24.1× fiscal 2027 earnings) while earnings grow to approximately $12.90 per share in fiscal 2028 (from about $9 per share in fiscal 2027), its market capitalization could approach $7.5 trillion, compared to roughly $5.3 trillion today—though valuation compression could limit upside.
Nvidia's data center networking business has emerged as a hidden engine of the company's growth, generating $14.8 billion in revenue during its fiscal 2027 first quarter (ended April 26), a 199% year-over-year jump. This figure caps a remarkable three-year run: networking revenue climbed from $8.6 billion in fiscal 2024 to $13 billion in fiscal 2025, then jumped to $31.4 billion in fiscal 2026.
The opportunity exists because modern AI systems depend on more than raw computing power. Training and running sophisticated AI models requires massive data flows among thousands of accelerators and storage devices. If the network infrastructure cannot keep pace, communication becomes a bottleneck that reduces GPU utilization and slows down entire AI workloads. Nvidia's NVLink technology connects GPUs within powerful AI rack systems, while its Spectrum-X Ethernet solution connects servers and racks across larger data centers. As AI clusters scale further, Nvidia has more opportunities to sell this networking and interconnect technology alongside—or even independent of—its computing hardware.
Nvidia already holds the market leadership position. According to research firm IDC, the company captured 21.5% of data center Ethernet switching revenue in the first quarter of 2026, making it the market leader, though Arista Networks trails only marginally at 20.7%. This dominance reflects the strategic importance of networking as AI infrastructure matures.
The strategy extends beyond Nvidia's own chips. Custom AI chips from rivals pose a growing competitive threat to Nvidia's GPU business, but the company has positioned its NVLink Fusion rack-scale platform to work alongside custom processors. The expanded partnership with custom chip designer Marvell Technology exemplifies this approach: Marvell provides custom accelerators and networking hardware to connect those processors at high speed, while Nvidia supplies NVLink high-speed interconnect technology, Spectrum-X switches, ConnectX network adapters, and BlueField data processing units. This allows Nvidia to capture value from AI infrastructure spending even when it does not supply the primary accelerators.
The financial implications are substantial. Nvidia is currently trading at around 24.1 times Wall Street's fiscal 2027 earnings estimate of approximately $9 per share as of August 12. Analysts expect earnings to grow to roughly $12.90 per share in fiscal 2028. If Nvidia delivers on those expectations and maintains its forward valuation, the body notes, its market capitalization could approach $7.5 trillion, up from roughly $5.3 trillion today. However, the body cautions that valuation compression could limit upside even if earnings rise. Competitors remain significant—Arista's near-parity market share and customers' ability to choose alternative networking technologies present real constraints on Nvidia's pricing power and growth in this segment.
Nvidia's dominance in AI accelerators has long driven investor enthusiasm, but the company is increasingly diversifying its revenue streams through networking infrastructure. The body shows that as AI systems scale to thousands of accelerator chips, high-performance networking has shifted from a peripheral concern to a critical bottleneck—communication speed directly impacts GPU utilization and workload performance. This shift unlocks a new business model for Nvidia: rather than competing solely on GPU market share, it can sell networking and interconnect technology alongside (or even to customers who choose competing) custom accelerators. The fiscal 2026 data already shows this strategy working—$31.4 billion in networking revenue is substantial enough to merit serious attention from investors.
The partnership expansion with Marvell Technology is the clearest signal that Nvidia sees its future as a broader infrastructure company, not just a chip manufacturer. By supplying the connective tissue between custom accelerators and storage, Nvidia creates switching costs and lock-in effects that are harder for rivals to disrupt than GPU competition alone. However, the body notes meaningful headwinds: Arista Networks holds a near-equal market share (20.7% vs. Nvidia's 21.5%), and customers retain freedom to choose competing networking technologies. The valuation upside to $7.5 trillion assumes not only that earnings grow but that Nvidia's multiple does not compress—a constraint the body itself flags as a real risk.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
Google launched the Pixel 11, Pixel 11 Pro, Pixel 11 Pro XL, Pixel 11 Pro Fold, and Pixel Watch 5
Swedish AI coding startup Lovable raised $400 million in Series C funding, doubling its valuation from $6.6 bi…
Blacksmith Software raised $45 million in Series B funding led by Peak XV Partners, with Y Combinator and GV a…
Nvidia is assembling more than US$500 billion in third-party capital from major Wall Street firms to fund AI i…

Gallium arsenide (GaAs), a semiconductor material, is gaining adoption in short-reach optical links for AI dat…

Compal chairman Ray Chen said the company has moved from being a newcomer to AI servers two years ago into the…

The AI news that matters, in one minute each morning.
Sign up free