Latos unveils swift AI-ready edge data centres in the UK

Latos, a data centre operator in the UK, has launched a new capability to design, build, and operate small-scale volumetric data centres at the network edge, which can be constructed in as little as six months.

The company's new facilities are designed to support real-time AI applications, including autonomous vehicles and banking, as they enable AI and other high-performance systems to operate faster and more securely.

The edge-based data centres, scalable from 24 racks, aim to facilitate what consultancy firm Accenture describes as the "binary big bang", the transformation of businesses and public services through AI. These facilities will augment Latos' previously announced 50,400 sq metre hyperscale data centre in Cardiff.

Mike Carlin, CEO of Latos Data Centres, said, "While we applaud the government for its ambition over the UK's AI infrastructure, a handful of large-scale data centres in far-flung locations won't be enough. We'll need more compute facilities closer to end users – and that means the network edge."

Latos asserts that the new volumetric data centres, which will be constructed to Tier III standards and offer energy delivery of up to 100kW per rack, are suitable for intensive AI training tasks. For example, training a generative model such as ChatGPT generally requires around 80kW per rack.

These data centres will be constructed from modular components and will integrate the latest techniques in net-zero building, with features such as high-performance power and cooling systems. Latos plans to handle the design, construction, and operation of these data centres for clients across the UK and will also make the designs available for licensing to partners worldwide.

"We've designed Latos volumetric data centres specifically to help organisations capitalise on AI. They are state of the art, quick to build, and cost-effective to run," added Mike Carlin.

The introduction of Latos' edge data centre capability comes as the UK government outlines its intention to amplify the nation's computing capacity by 20-fold by 2030, as detailed in the AI Opportunities Action Plan. This involves creating AI Growth Zones (AIGZs) to promote private investment in data centres and other AI-related infrastructure.

Furthermore, the Bank of England indicates that over 70% of businesses are currently investing in or planning to invest in AI technology within the next three years, underscoring the increasing demand for data centre facilities that can support AI workloads.

Latos has set a target of opening 40 data centres throughout the UK by 2030, marking a significant expansion in its capabilities and service offerings. The company's move into the construction phase of a hyperscale data centre in Cardiff aligns with its broader strategy of building scalable and efficient computing facilities at the network edge.