Schneider Electric gathered leading voices from the UK’s data centre and AI ecosystem in London this week for a summit on how to build AI-ready infrastructure at speed, cost, and carbon levels that the country can sustain. The Partnering for AI-Ready Data Centres event, held as a flurry of national announcements signalled a rapid expansion of AI capacity, focused on the practical steps to decouple AI growth from energy consumption and to deploy next-generation cooling and power systems at multi-megawatt scale.
The timing was deliberate. In the same week, the UK unveiled Stargate UK, OpenAI’s new supercomputer in Loughton, while NVIDIA confirmed what it described as its biggest GPU rollout in Europe. CoreWeave outlined a major UK expansion, and the UK-US Tech Prosperity Deal unlocked a reported £31 billion of investment including Microsoft’s £22 billion AI build-out, Google’s £5 billion UK data centre and a Northeast AI Growth Zone expected to create more than 5,000 jobs. The message from speakers was that infrastructure will decide how much of this momentum turns into long-term capability.
Liquid cooling moves from option to requirement
Discussions centred on three technical levers for AI estates. First, liquid cooling, including chip-level heat extraction, was presented as the path to run dense accelerator racks reliably and to enable operators to scale beyond air cooling limits. Second, modular data centre designs were highlighted to compress delivery times without locking operators into bespoke, slow-to-evolve architectures. Third, power systems and controls were discussed as part of a wider effort to reduce the energy intensity of AI workloads and to align new capacity with renewable sources.
Speakers from across the ecosystem set out the case for coordinated action. Schneider Electric was joined by senior figures from NVIDIA, Dell Technologies, Supermicro, Deep Green, JLL and Motivair, reflecting the mix of silicon providers, system builders, operators, real-estate advisers and thermal specialists now involved in AI projects. The practical emphasis was consistent throughout the agenda, from the mechanics of extracting heat at the chip to the realities of operating large data halls under tighter power constraints.
The event also underlined how industrial policy and technical deployment now intersect. Participants repeatedly returned to the need for sovereign capacity and resilient infrastructure if the UK is to support an expanding base of AI developers and adopters. That argument connects investment headlines to the data halls, switchgear and cooling loops where AI workloads run.
Collaboration becomes the competitive variable
The organisers framed the day as a collaboration forum rather than a product showcase, with government, industry and academia all referenced as necessary partners. The goal is to create an environment where operators can deploy liquid-cooled systems, expand modular capacity and access renewable power without piecemeal decision-making or duplicated effort.
Comments from contributors pointed to the scope of the opportunity and the scale of the task. Schneider Electric’s UK and Ireland leadership described the event as a platform for sharing strategies that keep infrastructure both sustainable and resilient as AI demand grows. NVIDIA’s UK and Ireland enterprise team cited a commitment intended to catalyse the UK start-up ecosystem, while Dell Technologies emphasised the role of generative AI in reshaping processes inside businesses. Supermicro focused on the need to future-proof data centres as critical infrastructure for AI-driven growth, and other contributors argued for investment in sovereign, sustainable capacity supported by a coordinated ecosystem.
The thread running through those perspectives was straightforward. AI’s near-term gains will depend on engineering choices now being made in plant rooms and at the rack, especially around liquid cooling. Chip-level heat extraction was presented as a prerequisite for the most demanding workloads, and to increase rack density without unacceptable thermal risk. Modular designs were treated as a hedge against uncertainty, allowing operators to add capacity in discrete blocks and to standardise deployment methods across sites.
None of the speakers pretended that these moves remove constraints entirely. Power availability, water stewardship and grid connections remain limiting factors, and the skills required to design, install and operate liquid-cooled environments are in high demand. However, the consensus in the room was that collaboration across suppliers, operators, investors and public bodies can accelerate delivery and reduce the risk inherent in one-off, bespoke builds.
The London event functioned as a signal that the UK’s AI ambitions will stand or fall on the quality of its infrastructure choices. If the sector can adopt chip-level liquid cooling where appropriate, compress build times with modular systems and align growth with renewable energy, then the country is better placed to convert headline investments into durable capacity. If it cannot, then announcements will outpace what the underlying power and cooling systems can support.
By convening technology providers, system builders and consultants, Schneider Electric sought to keep attention on those practical decisions. The company positioned itself as a convener of an ecosystem that includes advanced power systems, modular data centre solutions and liquid cooling innovations. The wider point lands beyond any single brand. AI will remain an abstraction until the facilities that power it are built to run dense, heat-intensive workloads safely and efficiently. The work now is to turn collaboration into deployment.
