EXECUTIVE SUMMARY:

• Schneider Electric has given a first look to its complete liquid cooling technology portfolio, including CDUs, HDUs (liquid-to-air heat dissipation units), and rear-door heat exchangers (RDHx)
• The move showcases how the company is moving forward after the acquisition of liquid cooling technology provider Motivair in February
• Vladimir Galabov, senior research director, enterprise infrastructure at Omdia, said the partnership was a ‘significant milestone for the industry’

Schneider Electric has unveiled its portfolio of end-to-end cooling solutions engineered for an AI-centric future – and in the process given a glimpse of how the company is moving forward after acquiring Motivair.

The acquisition of Motivair, which was finalised in February, was seen as a key move from Schneider Electric in the data centre market, with the company noting that liquid cooling, specifically applied to data centres and AI, was a ‘nascent market set for strong growth.’

The complete Motivair by Schneider Electric cooling solutions, now available globally, comprise data centre physical infrastructure, including CDUs (coolant distribution units), HDUs (liquid-to-air heat dissipation units), and rear-door heat exchangers (RDHx).

The CDU family scales from 105 kW to 2.5 mW and is certified for NVIDIA’s latest hardware while being ‘well-equipped for future increases in rack density’, the company notes, while the RDHx cools rack densities up to 75 kW. The AI thread is evident throughout, with the RDHx being claimed as ‘ideal’ for power-intensive GPUs, and the HDUs ‘perfect’ for AI accelerators.

“Together, Schneider Electric and Motivair are delivering something the market has never seen before – a complete end-to-end portfolio of liquid cooling solutions that can be deployed in new or retrofitted data centres anywhere in the world at scale,” said Rich Whitmore, Motivair by Schneider Electric CEO.

Andrew Bradner, senior vice president at Schneider Electric’s cooling business, further emphasised the rationale behind the acquisition, combining Schneider Electric’s global infrastructure, software and service expertise, with Motivair’s liquid cooling solutions. “We’ve created a capability that no-one else can match,” Bradner added.

An example of this portfolio in action was in September when Equinix and Dell Technologies, alongside Schneider Electric, announced the successful deployment of liquid cooling technology in Equinix’s Hong Kong data centre. The pilot utilised a 50kW in-rack direct-to-chip liquid cooling system, adapting to Equinix’s existing infrastructure.

Such collaborations are key. Alongside Dell, HPE and Supermicro were cited by Schneider Electric on the server side, Worldwide Technology on the system integrator side, and NVIDIA for chips to give customers a ‘truly integrated and proven path to AI performance’, as Whitmore put it.

“Successfully supporting AI in high performance computing workloads today isn’t just about new computing demands,” said Bradner. “It’s an infrastructure reckoning that demands transformation. AI demands compute that’s always-on, highly dense, and deeply interconnected as we drive towards power densities of one megawatt per rack.

“Deploying AI is complex, and no-one does it alone,” added Bradner. “That’s why partnership is at the heart of our solutions.”

Analysis: Why data centre operators must invest ‘urgently’ in next-gen AI infrastructure

Vladimir Galabov, senior research director, enterprise infrastructure at Omdia, stressed that data centre operators could not afford to wait to invest in infrastructure that will power AI workloads.

The numbers, Galabov noted, were compelling. Omdia forecasts that data centre capex is on track to surpass $1 trillion by 2030, as the firm said at its analyst summit in April. Alongside this, he noted the most cutting-edge AI models can require up to five times more compute with each passing year.

“At the rack level, we’re now delivering more power than data centres were ever designed to handle, so we must innovate to better remove heat and better supply the required electricity to the rack,” said Galabov. “That is why a new high voltage power delivery architecture, and extremely efficient cooling are crucial to sustaining this exponential growth in AI computing.”

Liquid cooling has emerged as the clear solution, with Galabov noting deployments have doubled two years in a row. This has been echoed by industry executives. Stuart Crump, global commercial director at LiquidStack, wrote for Data Centre Insight in September that while AI ‘may be driving the next industrial revolution… liquid cooling will ensure this revolution is sustainable.’

“Future progress in liquid cooling technology will require semiconductor computing, electrical and cooling equipment designers, manufacturers, integrators to work together,” noted Galabov.

The integration between Schneider Electric and Motivair was a ‘significant milestone for the industry’, Galabov added. “I think it’s fostering competition, accelerating innovation, and improving availability and affordability for liquid cooling infrastructure,” said Galabov, adding that the collaboration has potential to expand industry partnerships.

Galabov concluded that this space, for all its potential, had only just begun – and all boats needed to be lifted by the rising tide.

“First, we have only seen a glimpse of how essential and widespread AI will become,” said Galabov. “Second, data centre operators should invest urgently to prepare for next-generation AI infrastructure. And third, there is a very long tail [of] smaller operator and enterprise ecosystem partners… that want to deploy AI, but they need help. They need support from local partners, and they need support from managed service providers.”

You can take a look at the full portfolio by visiting the Schneider Electric website (PDF).

Read more: Schneider Electric announces new designs for power management, liquid cooling controls