Redwood City, Calif. – September 07, 2026 -- Equinix, Inc. (Nasdaq: EQIX) will launch Equinix Inference Exchange in Q1 2027, a distributed AI inference program built with NVIDIA and Together AI that runs across the company's global data center footprint.
Equinix pairs NVIDIA's reference architectures with Together AI's model platform
The program combines NVIDIA's validated Enterprise Reference Architectures with Together AI's inference platform, which supports more than 200 open-source models. Delivery runs through Equinix's global data centers, with connectivity to clouds, networks and AI providers via Equinix Fabric.
The announcement came at Equinix Horizon, the company's first customer and partner event, alongside a companion product, Equinix Fabric One, designed to simplify cross-metro connectivity for distributed AI environments.
Equinix cites scale across 280 data centers and 77 metros to anchor the offering
Equinix operates more than 280 data centers across 77 metros, with 230 cloud on-ramps and over 10,500 businesses interconnected on its exchange. Eight of the top 10 AI model providers and nine of the top 10 AI clouds are already deployed on Equinix infrastructure.
Adaire Fox-Martin, CEO and President of Equinix, said the company's near three-decade history running enterprise interconnection positions it to combine NVIDIA's compute foundation with Together AI's open-model flexibility into architectures she described as "neutral by design, open by default."
NVIDIA and Together AI executives frame the deal around inference economics
Raj Mirpuri, NVIDIA's vice president of global AI clouds and infrastructure ecosystem, said the collaboration turns Equinix's interconnection platform into "a global fabric for AI inference" as accelerated compute becomes a strategic asset class. Together AI co-founder and CEO Vipul Ved Prakash said the partnership demonstrates that enterprises need not trade model performance for operational flexibility.
Three-layer architecture targets edge inference, open-model migration and data sovereignty
Equinix supplies the power, cooling and day-two operations layer connected through Equinix Fabric; NVIDIA anchors compute infrastructure aimed at maximizing AI factory throughput while minimizing token cost; Together AI runs the inference platform on top, offering both multitenant and dedicated single-tenant deployment options.
The solution targets three enterprise scenarios: metro edge inference for lower-latency deployments, migration from proprietary to open-source models without vendor lock-in, and sovereign AI deployments for regulated industries requiring data residency controls.
Futurum Group analyst says inference location has become a governance issue
Nick Patience, Vice President and Practice Lead for AI Platforms at The Futurum Group, said performance, cost and governance have become strategic considerations as AI workloads spread across more providers and environments, adding that solutions simplifying inference deployment while preserving flexibility will become increasingly important for enterprises.