Netris Bags Fifteen Million From Andreessen Horowitz to Speed Up Artificial Intelligence Cloud Operations

The sudden explosion of artificial intelligence has convinced almost everyone to attempt launching a data center business. However, building out a functional cloud environment from scratch is incredibly difficult. Even if you manage to secure hard-to-find graphics processing units, network switches, and data storage cabinets, you still have to configure the entire system manually to make sure it runs smoothly. Setting up a data center to host heavy training and inference workloads typically takes several months of engineering work. The longer it takes you to launch your infrastructure in the market, the more money you lose while those expensive graphics cards sit idle in their racks.

A network automation startup named Netris wants to solve this exact headache for specialized cloud providers, often called neoclouds. The firm builds intelligent software that runs directly on network switches. It provides a central platform that connects hardware components to help data center operators launch their services much faster by automating configuration, setup, and daily management tasks. The system isolates the underlying hardware layer, allowing operators to change server configurations instantly and share resources across multiple separate corporate clients safely.

If this feels like a solution to a problem that larger companies already solved, you are partially correct. Historically, data centers belonged almost entirely to massive infrastructure giants like Equinix, NTT, Digital Realty, Oracle, Microsoft, AWS, or Google. Those multi-billion dollar corporations possess the capital to hire thousands of specialized engineers to build custom internal automation systems from scratch. Small, emerging cloud businesses simply do not have those kinds of resources at their disposal.

Netris Chief Operating Officer Alex Saroyan explained that running modern graphics chips requires daily configuration adjustments across thousands of switches. Traditional server facilities use software-defined networking, but that old method falls short because it handles data at software speeds rather than hardware speeds. Artificial intelligence workloads generate such a massive wave of data traffic that the network infrastructure requires direct hardware acceleration. Netris spent eight years building a system that delivers software-style control while maintaining raw hardware speeds.

The network map diagram shown in image_d9271f.jpg highlights how the company handles these intricate connections. The software operates independently of any single manufacturer, meaning it works seamlessly with standard networking gear and server frameworks from both Nvidia and AMD.

This hardware flexibility helped the startup win over prominent backing early on, including praise from Nvidia. Two years ago, the chip manufacturing giant witnessed a demonstration of the platform and immediately began recommending Netris to its own infrastructure clients. Today, Netris handles live operations across more than thirty-five distinct graphics processing clusters worldwide, managing close to a billion dollars worth of hardware assets. Their software powers systems for major firms like Lightning AI, Foxconn, Voltron Data, Hewlett Packard Enterprise, TensorWave, and Telus.

To build on this massive momentum, Netris just raised fifteen million dollars in a Series A funding round led by venture capital powerhouse Andreessen Horowitz. Interestingly, Netris does not actually use any artificial intelligence inside its own software code. Saroyan noted that the platform relies entirely on deterministic algorithms developed over years of testing. He emphasized that because network management requires absolute consistency across thousands of switch adjustments, relying on unpredictable AI models would be dangerous. Netris plans to use the fresh funds to expand its sales staff, hire more engineers, and build compatibility for an even wider range of hardware vendors.