Netris and a16z: $15 Million to Revolutionize AI Neoclouds

⚡

Key Takeaways

1Netris raised $15 million in a Series A round with Andreessen Horowitz to accelerate the launch of AI neoclouds.

2The startup offers network automation that reduces the time to deploy data centers for AI.

3Netris is already used in over 35 GPU clusters, backed by giants like Nvidia and Foxconn.

💡Why it matters — This funding round could transform the competitiveness of smaller operators against cloud giants by simplifying AI infrastructure.

The Rise of AI Data Centers and Their Challenges

With the explosion of artificial intelligence, many companies are venturing into the creation of specialized data centers. However, this complex task requires much more than acquiring hardware such as GPUs, network switches, and storage solutions. Indeed, configuring and optimizing these infrastructures to meet the specific needs of AI, particularly for inference and training, can take several months. This delay increases the cost of GPUs, which are often unused during this period.

Netris: An Innovative Solution for Neo-Clouds

Netris, a startup specializing in network automation, offers a solution to reduce these delays. Its software, designed to operate on network switches, allows neo-cloud operators to launch their services more quickly. By automating configuration and operations, Netris provides a network abstraction that facilitates the adaptation of hardware configurations according to needs. Additionally, it ensures isolation of servers and resources, which is essential for serving multiple clients simultaneously.

A Market Dominated by Giants

Historically, data center management was reserved for large operators like Equinix, NTT, Digital Realty, Oracle, Microsoft, AWS, and Google. These companies have addressed the challenges of network configuration and multi-tenancy through large engineering teams or by developing their own automation solutions. In contrast, small neo-cloud companies do not have such resources.

Technology Tailored to AI Needs

Alex Saroyan, CEO of Netris, explains that GPU cluster operators must constantly adjust their configurations. In traditional data centers, SDN (software-defined networking) was used, but this software technology is insufficient for AI. The high traffic requires hardware acceleration. Therefore, Netris offers a solution that combines the benefits of SDN with complete hardware acceleration, expertise developed over eight years.

Universal Compatibility

Netris's platform is designed to be vendor-independent, compatible with network equipment and data center standards, whether it involves Nvidia or AMD servers. This flexibility has convinced major players in the industry.

Strong Support and Rapid Expansion

Impressed by Netris's technology, Nvidia has recommended the startup to several of its clients. Today, Netris is deployed in over 35 GPU clusters worldwide, representing around one million GPUs. Companies like Lightning AI, Foxconn, Visionbay, Hewlett Packard Enterprise, TensorWave, and Telus are already using its services.

A Strategic Fundraising

To support its growth, Netris raised $15 million in a Series A funding round led by Andreessen Horowitz. This funding will allow the startup to hire more engineers and sales staff, expand support for more hardware vendors, and add new features to its algorithm.

An AI-Free Approach

Interestingly, Netris does not use artificial intelligence in its solutions. Saroyan emphasizes that the company relies on internally developed algorithms to manage automation and operations, preferring a deterministic and repeatable approach over AI solutions, which are often non-deterministic.

A Promising Future

With the entry of Guido Appenzeller, partner at a16z, onto the board of directors, Netris is preparing to strengthen its market position. The company aims to become a key player in the neo-cloud space, simplifying AI infrastructure for small and medium-sized enterprises.