Equinix launched its Distributed AI Hub platform, which is designed to simplify and secure complex, distributed AI ecosystems for enterprises.

The Hub aims to provide a single, unified framework for enterprises to connect to and use various AI infrastructure providers through private, low-latency connectivity via Equinix's 280 global data centers.

The hub, which is set to be showcased at Nvidia’s upcoming GTC event, will provide vendor-neutral access to model developers, GPU clouds, data platforms, network, and security services.

“We are providing enterprises the freedom to build and scale AI wherever their data, partners, and teams already live, while running inference close to the data and users that depend on it, without the operational drag that comes from stitching together complex, distributed systems,” Equinix Chief Business Officer Jon Lin explained.

AI infrastructure is undergoing somewhat of an evolution, with the shift from training to inference meaning computational power needs to be closer to where it’s actually needed.

Equinix argues its new distributed platform will give enterprises a decentralized means to run AI workloads without the need to rebuild their architecture each time.

And in a jab at hyperscale rivals, the colocation giant said its hub is “open and vendor-neutral by design,” contending AI marketplaces from the big boys tend to favor their own services.

First in line to integrate with the platform was Palo Alto Networks, which has lent its Prisma AIRS real-time AI security offering to ensure enterprises gain visibility and control over applications running via the Hub.

“With our Distributed AI Hub, we're giving customers a simpler, smarter, and far more connected way to run and scale their AI today,” Lin added.