The Netris Series A funding round of $15 million, led by Andreessen Horowitz, lands at a moment when everyone from telecoms giants to contract manufacturers is trying to stand up a GPU data centre before the AI window closes. Standing one up quickly, it turns out, is a different problem from standing one up at all.
Netris, founded in 2018 and headquartered in Santa Clara, California, sells software that runs directly on network switches and a platform that automates the setup, configuration, and ongoing operations of those switches. The idea is to strip out the months of manual network engineering that currently sit between a neocloud operator receiving its GPUs and the moment it can actually bill a customer.
CEO Alex Saroyan has a clear view of why the old tooling falls short. ‘As a GPU cluster operator, you need to make configuration changes to every link, every day. At traditional data centres, they were using something called SDN [software-defined networking] to do this, but SDN is falling short, because it’s a software technology,’ he told TechCrunch. ‘For AI, software is not okay, because the amount of traffic is so high, everything must be hardware accelerated. So you need something like SDN, but completely hardware accelerated. This is what we do, and this is what we’ve been doing for eight years.’
The company brands its approach NAAM, standing for Network Automation, Abstraction, and Multi-Tenancy, and positions it as the successor to SDN and intent-based networking, built specifically for the multi-layer network fabrics that GPU clusters require. According to FinSMEs, NAAM supports NVIDIA Spectrum-X Ethernet, NVIDIA Quantum InfiniBand, NVL72, NVIDIA BlueField DPUs, and virtual and edge networking, covering the full arc from Day Zero design and simulation through live operations.
What the Netris Series A Funding Actually Buys
The raise will go towards hiring engineers and sales staff, expanding support for more hardware vendors, and adding functionality to Netris’ core algorithm. Saroyan is emphatic that the algorithm involves no generative AI: ‘We started way before AI. We understood the challenge early on, and we started developing this algorithm early on. AI is not deterministic, right? Sometimes it likes to do things on its own. It’s good for creative work, but for changing many thousands of switch configurations, you don’t need to be creative. You need to be very persistent and repeatable.’
That philosophy has found a receptive audience. Netris is now live at more than 35 GPU clusters worldwide, covering roughly a million GPUs in total, operated by customers including Lightning AI, Foxconn, Visionbay, Hewlett Packard Enterprise, Tensorwave, and Telus. According to Andreessen Horowitz, some of those customers have built GPU data centres worth close to a billion dollars and have standardised on Netris software throughout, which a16z cited as a strong sign of confidence in the startup.
Two figures capture the commercial momentum, though they come from different sources and do not agree. The Business Wire press release puts ARR growth at 800% over the last 12 months; Netris’ own blog cites 622% year-over-year ARR growth in 2025. Both numbers are from the company; the Business Wire release, as the formal Series A announcement, is likely the more current figure. Either way, the direction is hard to argue with.
The NVIDIA relationship gives Netris a degree of third-party credibility that a press release alone cannot manufacture. Two years ago, Nvidia was apparently impressed enough by a demo to begin recommending Netris to customers. Then, in Q1 2025, the company became the first independent software vendor validated by NVIDIA for AI Network Automation, with multi-tenant NVIDIA Spectrum-X Ethernet live deployments enabled by NVIDIA Air AI factory simulation, according to the Netris leadership blog. Saroyan also confirmed the platform is vendor-agnostic and compatible with both NVIDIA and AMD server environments.
a16z Sees a Familiar Pattern
The a16z investment was authored by partners Guido Appenzeller, Raghu Raghuram, Jason Cui, and Martin Casado, with Appenzeller joining the Netris board. According to Forbes, the a16z team drew explicit parallels between Netris and the early software-defined networking wave that underpinned the first generation of cloud growth, and the investing team brings deep experience in networking and infrastructure automation.
The co-founders alongside Saroyan are Arsen Arakelyan and Tigran Martirosyan. The three built the company without a Series A for seven years, which probably explains why the algorithm prioritises being ‘persistent and repeatable’ over anything flashier.
The neocloud market is crowded and capital-intensive. Every month a GPU cluster sits unconfigured costs real money. If Netris can reliably compress that window from months to weeks, the Andreessen Horowitz bet looks straightforward. The proof will come in whether the 35-cluster footprint doubles before the next fundraise.
