Netris raises $15M Series A from a16z to help AI neoclouds go live faster
Summary
Network automation startup Netris has secured \$15 million in Series A funding from Andreessen Horowitz to accelerate its solution for "neoclouds" – new data centers specializing in AI inference and training. The company addresses the significant challenge of rapidly configuring and operating GPU clusters, a process that can take months for smaller operators lacking the engineering resources of hyperscalers. Netris provides software that runs on network switches and a platform for automated setup, configuration, and multi-tenancy, reducing go-live times. Its vendor-agnostic platform supports both Nvidia and AMD servers and is currently deployed across more than 35 GPU clusters globally, totaling about a million GPUs, for clients like Lightning AI and Foxconn. Notably, Netris achieves this automation using proprietary algorithms developed over eight years, explicitly avoiding non-deterministic AI for critical network configuration tasks.
Key takeaway
For AI Architects or MLOps Engineers building or expanding GPU clusters, Netris' approach offers a compelling solution to accelerate deployment. You should evaluate hardware-accelerated network automation platforms to significantly reduce months-long setup time and operational overhead. This allows you to bring compute resources online faster, minimizing costly GPU idle time. It also enables robust multi-tenancy for diverse customer needs, without relying on non-deterministic AI for core network functions.
Key insights
Hardware-accelerated network automation is critical for rapidly deploying and scaling AI-focused data centers.
Principles
- Network configuration for AI demands hardware acceleration.
- Deterministic algorithms are preferred for critical infrastructure.
- Multi-tenancy requires hardware-layer isolation.
Method
Netris automates network setup, configuration, and operations via software on switches and a connected platform, providing network abstraction.
In practice
- Automate GPU cluster network setup to reduce idle time.
- Implement hardware-accelerated SDN for AI traffic.
- Use vendor-agnostic network solutions for flexibility.
Topics
- Network Automation
- GPU Clusters
- AI Infrastructure
- Neoclouds
- Hardware Acceleration
- Andreessen Horowitz
Best for: Investor, CTO, VP of Engineering/Data, AI Architect, MLOps Engineer, Director of AI/ML
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by AI News & Artificial Intelligence | TechCrunch.