Equinix opens Sydney centre for liquid-cooled AI tests

Equinix has opened a network of Enhanced Performance Innovation Centres, including a site in Sydney, to let companies test liquid-cooled AI infrastructure before wider deployment.

The programme is now live in Sydney, Frankfurt, London, Osaka and Paris, targeting enterprises moving AI projects from experimentation into production. Customers can use the sites, located inside operating data centres, to trial GPU and CPU systems, run training and inference workloads, and examine operational issues before committing to larger installations.

Rising demand for AI computing has increased power use and rack density in data centres, pushing more operators towards liquid cooling. That shift has also created practical challenges for companies that need to integrate cooling systems with power, networking and day-to-day facility operations.

The new centres are intended to address those issues by providing production-ready environments where customers can assess equipment under live conditions. Enterprises work with Equinix staff and hardware partners to test how different parts of a liquid-cooling setup function together.

Live testing

Each EPIC location operates within an Equinix data centre rather than as a standalone demonstration lab. That allows customers to evaluate systems in the same type of setting where they are expected to run, including the operational constraints that come with a live facility.

The centres also give enterprise teams direct exposure to liquid-cooling technology before they make investment decisions. Equinix is targeting organisations that may have limited in-house experience deploying and operating high-density AI infrastructure at scale.

Brian Stein, Senior Vice President, Infrastructure Products and Services at Equinix, outlined the company's view of the market. "Liquid cooling is an essential component of modern enterprise infrastructure and a hard requirement for customer's deploying AI," said Brian Stein, Senior Vice President, Infrastructure Products and Services at Equinix.

"EPIC provides businesses with a place to test, validate and build with confidence before they commit to large-scale deployments. We're putting our global infrastructure footprint and unmatched ecosystem to work to accelerate the path from concept to production for our customers," Stein said.

Broader build-out

The launch extends Equinix's existing approach of creating dedicated spaces for customers and partners to test infrastructure before wider use. In 2022, the company introduced a Co-Innovation Facility for partner collaboration on new technologies before those systems entered Equinix data centres.

It later added Solution Validation Centres in 2024 to help customers validate deployments and review how Equinix supports distributed environments. EPIC adds liquid cooling and dense AI workloads to that model.

The Sydney site puts Australia among the first wave of locations for the programme, alongside several European centres and Osaka in Japan. More locations are expected to open in 2027, indicating sustained customer demand for environments focused on pre-deployment testing of AI systems.

The initiative reflects a wider issue facing corporate IT teams and data centre operators as AI adoption spreads beyond pilot projects. Systems built around advanced GPUs can place far greater thermal and power demands on facilities than conventional enterprise workloads, making infrastructure design a more immediate concern for organisations looking to scale AI use.

For some companies, that creates a skills gap as much as a technology challenge. Liquid cooling may already be established in parts of the data centre market, but many enterprises have not operated it directly across distributed environments and need to understand maintenance, consistency and integration before introducing it more broadly.

One customer cited by Equinix described that process in practical terms. "Liquid cooling was new territory for us, so we built our confidence step by step," said Matthieu Blumberg, SVP of Engineering Infrastructure, Security, Internal IT & CISO at Criteo.

"Our AI runs across a distributed footprint, and the infrastructure must perform the same way wherever it sits. We saw how liquid cooling performed in practice, learned what it took to run it well, and then committed to scale. Equinix was with us at every stage, with the expertise and the live environment to prove it out first. That's what turned liquid cooling from an open question into a core part of our distributed AI infrastructure," Blumberg said.