Hedgehog Contributes Open Compute Project Reference Architectures for AI Training and Inference

New OCP Accepted™ designs accelerate deployment of open, scalable AI infrastructure

Hedgehog, the AI network company simplifying how AI infrastructure is built and operated, announced it has contributed its AI training fabric and AI inference fabric designs to the Open Compute Project (OCP) as reference architectures, and are immediately available through the OCP Marketplace. The OCP Accepted™ recognition applies to products that align/comply with approved OCP specifications

Announced in conjunction with the 2026 OCP EMEA Summit, taking place in Barcelona, Spain April 29 – 30, the contributions provide operators, system builders, and integrators with validated, production-ready blueprints for deploying open, Ethernet-based AI networks using disaggregated hardware and Hedgehog AI network software.

Our goal has always been to make AI networks easier to deploy and operate in the real world.

Open AI Fabrics Built for Real Deployments

Based on real-world production deployments supporting today’s most demanding workloads, The reference architectures emphasize interoperability across silicon vendors to prevent hardware lock-in. The designs address distinct environment requirements:

• AI Training Fabrics: Designed for large-scale GPU clusters, delivering predictable performance through congestion-aware routing, lossless Ethernet, and automated network lifecycle management.
• AI Inference Fabrics: Optimized for efficiency and low latency, featuring multi-tenant security, hybrid multi-cloud routing, simplified operations, and consistent performance at scale.

“Our goal has always been to make AI networks easier to deploy and operate in the real world,” said Marc Austin, CEO, Hedgehog. “By contributing these AI training and inference fabrics to OCP, we’re sharing already proven designs that help the community move faster while preserving choice across hardware and silicon.”

Also Read: AiThority Interview with Glenn Jocher, Founder & CEO, Ultralytics

Ecosystem Collaboration from Design to Production

The reference architectures reflect close collaboration across the OCP ecosystem, previewed initially during a joint panel at the OCP Global Summit 2025, where several companies shared lessons learned from deploying Ethernet-based AI fabrics at scale.

“Our long-term GPU customers need a network that keeps their clusters fed — no bottlenecks, no surprises,” said Jonmichael Hands, CEO, FarmGPU. “Hedgehog made our backend fabric setup straightforward and their support team has been rock solid. Getting these designs into OCP means more operators can run real AI infrastructure without reinventing the wheel.”

As an OCP hardware partner, Celestica worked closely with Hedgehog to validate the architectures across open systems.

“These reference architectures show how Celestica’s OCP-inspired switches and open source networking software can be combined to support modern AI workloads,” said Olivier Suinat,Chief Revenue Officer, Enterprise AI Platforms, Celestica. “Celestica is a Platinum OCP Member and proud to support contributions like these that give customers a clear, deployable path to building open and scalable AI infrastructure aligned with OCP standards.”

Advancing Open AI Infrastructure with OCP

The contribution aligns directly with OCP’s Networking Project to advance scalable, high-performance infrastructure.

“As AI workloads drive new demands on data center networks, the OCP Community is focused on enabling open, Ethernet-based solutions that can scale efficiently and operate reliably in production,” said James Kelly, VP Market Intelligence and innovation, Open Compute Project Foundation. “By publishing these validated AI fabric reference architectures through the OCP Contributions Database and associated products in the OCP Marketplace, we are giving operators and system builders direct access to designs that make it easier to adopt open, disaggregated AI networking with confidence and clarity.”