Rebecca PY MALAYSIA, 4 SEPTEMBER 2026 – Equinix, Inc. has announced a major expansion of its long-standing collaboration with NVIDIA to introduce Equinix Inference Exchange, a distributed AI inference program designed to accelerate how global enterprises move AI from experimentation into production.
Unveiled at the company’s inaugural customer and partner event, Equinix Horizon, alongside the new Equinix Fabric One interconnect platform, the initiative combines NVIDIA’s validated Enterprise Reference Architectures with Together AI’s open-model inference platform, which supports over 200 open-source models. Delivered across Equinix’s global footprint of more than 280 data centers in 77 metros, the solution provides low-latency, high-performance connectivity via Equinix Fabric to major clouds, networks, and AI providers.
As AI adoption shifts from initial testing to full-scale enterprise production, determining where inference runs has become a critical strategic decision impacting operational performance, cost, and governance. Equinix Inference Exchange addresses these challenges by offering a flexible, neutral-by-design foundation that brings intelligence closer to enterprise data, applications, and end users. The integrated architecture combines Equinix’s physical infrastructure, power, and cooling capabilities with NVIDIA’s accelerated compute foundation and Together AI’s hybrid platform, supporting both multitenant efficiency and dedicated single-tenant environments.
The solution targets key modern enterprise requirements, including metro edge inference for ultra-low latency, seamless migration from closed to open-source models to avoid vendor lock-in, and sovereign AI deployments tailored for strict data residency and regulatory compliance. Scheduled to become available in Q1 2027, the platform reinforces Equinix’s central role in global AI infrastructure, where eight of the top ten AI model providers and nine of the top ten AI clouds are currently deployed.