NVIDIA’s networking revenue is no longer a footnote to its GPU business.
Multiple financial sources report the division generated approximately $11 billion in its most recently reported quarter. Phemex reports 267% year-over-year growth for the period, a figure that also appears in ainvest.com’s analysis. Neither source specifies the fiscal quarter by name. Until NVIDIA’s investor relations materials confirm the precise period, these figures should be treated as reported rather than audited. The T1 verification source – NVIDIA’s earnings release, wasn’t surfaced in this package.
The annualized implication is notable. At $11 billion per quarter, NVIDIA’s networking division is running at roughly $31 to $44 billion annually, depending on quarterly variation. At that scale, observers have noted the division rivals Cisco’s total networking revenue, a comparison that underscores how thoroughly AI infrastructure demand has transformed NVIDIA’s business mix.
This is what AI infrastructure investment looks like in revenue terms. Every data center rack built for AI training or inference requires networking fabric to connect the compute. NVIDIA’s InfiniBand and Ethernet networking products are the connective tissue of the AI build-out. The 267% growth figure, if it holds against NVIDIA’s official reporting, is a direct measure of how fast that build-out is accelerating.
For investors tracking the AI infrastructure thesis, NVIDIA’s networking revenue is a cleaner signal than GPU revenue alone: it measures the density of the infrastructure being built, not just the chips being purchased.