Nvidia said Google and Meta, two of its largest customers and its sharpest rivals in custom AI silicon, are picking up its networking hardware.
The company made a set of announcements at the Open Compute Project Global Summit in San Jose on Monday.
Will Google and Meta use Nvidia’s networking technology?
Nvidia announced at the Open Compute Project Global Summit in San Jose today that Google and Meta will both adopt its networking hardware.
This announcement comes despite the fact that both firms have spent the past year trying to loosen Nvidia’s grip. Meta signed a multi-billion-dollar deal to rent Google’s Tensor Processing Units and has shipped its own MTIA chip designs. Google now sells TPUs to outside customers.
Meta is working with Nvidia on open AI networking built around its Spectrum-X Ethernet SuperNIC and Spectrum-6 switch chips, according to Nvidia’s blog post. Google will separately support Falcon Transport on Nvidia’s upcoming ConnectX-10 SuperNICs and BlueField data processing units.
Microsoft showed off its own Maia 300 inference chip at the summit, aiming to reduce its reliance on Nvidia and AMD. Still, Cryptopolitan reports that hyperscalers like Microsoft make up more than half of Nvidia’s quarterly revenue.
Nvidia began shipping Vera Rubin in August, after announcing it in March. Supermicro said on September 23 that it was shipping Vera Rubin racks, and CoreWeave said a week later that a customer was running production workloads on it.
Nvidia also shared its rack and compute-tray designs with the Open Compute Project, so other companies can build compatible hardware.
What else did Nvidia announce?
Nvidia released a specification called Open Data Center Specification v0.7, which is a shared blueprint for how AI factories should handle power, cooling, and layout. AMD, Google, and Meta joined Nvidia in publishing it.
The second published specification is the OCP Grid Resilience Specification, which Nvidia developed with Google, Meta, and Microsoft. It lays out how large AI sites should ride out power grid disturbances and get back to near-full load in under two seconds.
Nvidia also announced the Ultra-Dense Small Form Factor Pluggable (USFP) module with Cisco. It’s an optical design that runs at 400 gigabits per second per lane and reaches 12.8 terabits per second in total.
The companies say it offers four times the bandwidth density and about 1.3 times the power efficiency of 200-gigabit optics. TE Connectivity and Amphenol are working with them on the parts.
Nvidia’s revenue more than doubled in its fiscal second quarter (May to July 2026), rising 106% from a year earlier to $96.2 billion. Its data center business made up most of it at $89 billion. Meanwhile, Management predicts that the coming fiscal year will be supply-constrained.
The smartest crypto minds already read our newsletter. Want in? Join them.