Equinix has expanded its collaboration with NVIDIA and established a new collaboration with Together AI to deliver Equinix Inference Exchange, a distributed AI inference program designed for global enterprises.
The initiative combines NVIDIA’s validated Enterprise Reference Architectures with Together AI’s inference platform, which supports more than 200 open-source models. Equinix will provide the underlying digital infrastructure through its global data centers, with Equinix Fabric connecting enterprises to clouds, networks and AI providers.
The development strengthens Equinix’s position in AI inference infrastructure as enterprises move AI workloads from experimentation into production. The platform is designed to bring inference closer to the data, users and applications it serves while addressing performance, cost and governance requirements across distributed environments.
Equinix Expands Distributed AI Infrastructure
Equinix Inference Exchange is designed to simplify the deployment of distributed AI workloads across different models, providers, clouds and geographic locations. The platform provides secure, low-latency connectivity to the infrastructure and ecosystems enterprises need as AI adoption expands.
Equinix operates more than 280 data centers across 77 metros, with 230 cloud on-ramps and more than 10,500 interconnected businesses. The company also said eight of the top 10 AI model providers and nine of the top 10 AI clouds are deployed with Equinix.
This scale provides the foundation for the company’s AI data centers and interconnection strategy, allowing enterprises to deploy inference closer to users and data. Equinix provides power, advanced cooling and operational support, while Equinix Fabric connects the infrastructure to clouds, networks and AI providers.
NVIDIA And Together AI Support Enterprise AI
NVIDIA anchors the platform with its Enterprise Reference Architectures and AI infrastructure, while Together AI provides its inference platform and access to more than 200 open-source models.
The architecture supports both multitenant deployments for shared efficiency and dedicated single-tenant environments where enterprises require dedicated capacity. Through Equinix Fabric, the solution will connect inference providers across major metropolitan areas and provide access to a wider ecosystem of clouds, networks and AI providers.
The combination gives the AI inference infrastructure offering a three-layer structure: Equinix provides the physical and connectivity foundation, NVIDIA supplies the accelerated computing architecture, and Together AI provides the inference platform.
This approach is also intended to give enterprises greater open-source AI model choice while reducing deployment complexity and supporting distributed AI environments.
AI Inference Infrastructure Targets Enterprise Use Cases
The AI inference infrastructure is designed around several enterprise use cases, including metro-edge inference, open-model migration and sovereign AI. Metro-edge inference allows workloads to run closer to users and data, supporting lower-latency AI applications.
For enterprises moving from proprietary models to open-source alternatives, Together AI’s platform provides a path to run those workloads through the same interconnected environment used to reach other providers. The platform can also support sovereign AI deployments where organizations need to control where data and inference are processed to satisfy residency and sovereignty requirements.
As distributed AI deployments become more complex, the AI inference infrastructure provided through Equinix Inference Exchange is intended to give enterprises greater flexibility over where inference runs and how those workloads connect with the wider digital environment. Equinix Inference Exchange is scheduled to become available in the first quarter of 2027.