Netris Raises $15M Series A Led by a16z for GPU Cluster Networking

Standing up a GPU cluster is as much a networking problem as a hardware one, and for the neocloud operators building out AI infrastructure, the network engineering talent to do it cleanly is scarce. Netris, which makes hardware-accelerated network automation software that runs directly on network switches, closed a $15 million Series A led by Andreessen Horowitz, according to TechCrunch. The software automates setup, configuration, and operations for GPU clusters, and provides network abstraction and multi-tenancy capabilities so smaller teams can run production infrastructure without large internal engineering staffs.

The traction figures are notable for a company at this stage: Netris is reportedly active across more than 35 GPU clusters globally, covering approximately 1 million GPUs in total. Customers include Lightning AI, Foxconn, TensorWave, Telus, Visionbay, and Hewlett Packard Enterprise. Nvidia recommended Netris to its customers two years ago after a product demonstration -- a distribution channel that is harder to manufacture than a logo on a website.

CEO Alex Saroyan's case for the hardware approach is direct: "For AI, software is not okay, because the amount of traffic is so high, everything must be hard." The company has been developing its automation algorithms for eight years, well before the current wave of AI infrastructure demand. The new funding will go toward hiring engineers and sales staff, adding support for additional hardware vendors, and expanding algorithm functionality. Guido Appenzeller, the a16z partner on the deal, will join the board.

What the reporting does not give you is revenue or growth rate, so it is harder than usual to assess what this round buys in terms of runway. The honest caveat is also that Netris's fortunes are partly tied to the health of its neocloud customers, a segment that has seen real stress. The longer-range risk is that Nvidia or switch hardware vendors absorb comparable automation capabilities into their own platforms.

For AI infrastructure teams that lack deep networking expertise, the model Netris is selling is a plausible shortcut to faster deployments. If the neocloud segment keeps expanding, demand for exactly this kind of automation should grow alongside it.