The archive · Developer & Business Tools · Product decision · 2018–2026
Netris' $15M a16z bet: hardware-accelerated networking gets AI neoclouds live fast
Netris raised $15M from a16z to automate switch configuration for AI neoclouds, cutting data-center go-live from months to days.
Netris
What the business is
Sells switch-resident software plus a management platform that automates configuration, operations and multi-tenancy for GPU data centers, live at 35+ clusters.
Starting capital：$15M Series A led by Andreessen Horowitz (June 2026)
How it started
Netris was started by Alex Saroyan and team roughly eight years before the AI boom, betting that data-center operators would need a better way to configure networks at scale. For years that market belonged to hyperscalers like Equinix, NTT, Oracle, AWS and Google, which solved network setup by hiring ranks of engineers or building their own automation; the AI boom then minted thousands of small neoclouds with GPUs but none of that expertise.
What happened
About two years before the round, Nvidia was impressed by a demo and recommended Netris to several customers. The platform went live at more than 35 GPU clusters worldwide (about a million GPUs), operated by Lightning AI, Foxconn, HPE, Visionbay, TensorWave and Telus. In June 2026 Netris raised a $15 million Series A from a16z, whose partner Guido Appenzeller joined the board; the funds are earmarked for engineers, sales staff, more hardware-vendor support and richer algorithm features.
How it ended up
Scaling: with a16z backing and a live footprint across 35+ clusters, Netris is expanding vendor support and hiring; no exit.
Background
Netris is a network automation startup betting that the AI data-center boom created a market of operators who cannot configure their own networks. Spinning up a GPU data center requires securing GPUs, switches and storage, then configuring everything for inference and training — a process that can take months, during which expensive GPUs sit idle. Netris sells software that runs on network switches plus a platform that automates setup, configuration and operations, compressing the go-live timeline for AI neoclouds.
The product bet is architectural: traditional software-defined networking (SDN) is a software technology, and CEO Alex Saroyan argues that at AI traffic volumes everything must be hardware accelerated. Netris offers SDN-like flexibility with deterministic, repeatable, hardware-accelerated configuration — and deliberately no AI, since Saroyan says changing thousands of switch configurations requires persistence, not creativity. The company has been building this since roughly eight years before the AI boom.
The bet found believers: about two years before its funding round, Nvidia was so impressed by a demo that it recommended Netris to customers. By June 2026 the platform was live at more than 35 GPU clusters worldwide — about a million GPUs — with customers including Lightning AI, Foxconn, HPE, Visionbay, TensorWave and Telus. That momentum produced a $15 million Series A from Andreessen Horowitz, with partner Guido Appenzeller joining the board.
The funding is earmarked for more engineers and sales staff, support for more hardware vendors, and additional functionality in the automation algorithm, as the company competes with hyperscalers' in-house systems for the neocloud build-out.
What has to be true
- The AI boom created thousands of small GPU-cloud operators that lack the engineering armies hyperscalers use to run networks, opening a tooling market.
- Configuration at AI scale must be hardware accelerated and deterministic, so a pre-AI algorithmic approach is defensible against newer AI-native competitors.
- Reference wins were concrete: Nvidia recommended the product, and 35+ clusters with ~1M GPUs were already live before the a16z round.
- The platform is vendor-agnostic across Nvidia and AMD servers, letting neoclouds adopt it without re-platforming.
What can be applied
An infrastructure bottleneck created by a boom is a startup wedge: the boom minted thousands of small operators who cannot replicate hyperscalers' in-house automation, so sell them the tooling.
Aftermath
As of June 25, 2026, Netris is scaling: a $15M Series A from a16z, 35+ live GPU clusters totaling about a million GPUs, and customers including Lightning AI, Foxconn, HPE and Telus. The company plans to use the capital to hire engineers and salespeople, add support for more hardware vendors and extend its automation algorithms, with a16z partner Guido Appenzeller on the board.
Sources
- Netris raises $15M Series A from a16z to help AI neoclouds go live faster
- Netris获a16z 1500万美元A轮融资，助力新云服务商快速上线
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card