Equinix and NVIDIA partnership pushes AI inference closer

The Equinix and NVIDIA partnership is expanding into a new phase focused on one of enterprise AI’s emerging challenges: where inference should run. Equinix has announced Equinix Inference Exchange, a distributed AI inference program that brings together NVIDIA’s Enterprise Reference Architectures, Together AI’s inference platform and Equinix’s global data centre and connectivity infrastructure.

The shift matters because enterprise AI is moving beyond experimentation. As workloads spread across models, clouds, providers and locations, the infrastructure supporting inference also becomes more distributed. Equinix’s approach is to bring inference closer to users, data and applications while giving enterprises more flexibility around connectivity and deployment.

Inference is becoming a location decision

The central idea behind Equinix Inference Exchange is that AI performance is no longer only about the model or computing hardware. The location of inference can influence performance, cost and governance, according to the announcement.

Managing those distributed environments, however, can also add operational complexity. The proposed solution aims to address this by connecting inference providers through Equinix Fabric, alongside access to clouds, networks and other AI providers.

Equinix says its infrastructure spans more than 280 data centres across 77 metros, with 230 cloud on-ramps and more than 10,500 interconnected businesses. The company also notes that eight of the top 10 AI model providers and nine of the top 10 AI clouds are deployed with Equinix.

NVIDIA and Together AI add two key layers

The NVIDIA enterprise AI solutions element provides the accelerated computing foundation through NVIDIA Enterprise Reference Architectures and AI infrastructure designed for AI factory throughput and token-cost efficiency.

Together, AI adds an open-model layer supporting more than 200 open-source models. Its platform can support both multitenant deployments and dedicated single-tenant environments, giving enterprises different deployment options based on workload needs.

Equinix provides the underlying infrastructure, including power, cooling and day-two operations, with Equinix Fabric connecting the environment to relevant clouds, networks and AI providers.

Read More: 

The Convergence of AI, Automation and Enterprise Execution

Adobe New CEO Anil Chakravarthy set to lead AI era

TCS subsidiary Hypervault makes a big AI data centre move in Hyderabad

HPE and Oracle Expand Alliance to Scale Global AI Data Center Infrastructure