Google Secures AI Compute Capacity at Nscale's West London Data Center
April 15, 2026
In a move highlighting the intense competition for advanced computing infrastructure in the AI era, Google has reportedly agreed to take up capacity at an Nscale data center facility in West London, United Kingdom. The deal underscores the strategic importance of securing high-performance GPU resources as major technology firms race to deploy and scale their artificial intelligence services.
According to a report from Bloomberg, citing a source familiar with the matter who requested anonymity, the agreement will grant Google access to Nscale's Grace Blackwell GPUs housed at the site. This development follows closely on the heels of OpenAI's decision to withdraw from a previous arrangement to lease space at several planned, but not yet built, Nscale data centers across the UK. The facility involved in the Google deal is understood to be a separate operation from those previously linked to OpenAI.
Nscale's current UK footprint includes a location in Slough, west of London, and another in Loughton, Essex, northeast of the capital. The Loughton site, acquired in early 2025, boasts a power capacity of 50 megawatts, scalable to 90MW, and is slated to become operational by the end of 2026. Microsoft has announced plans to build a supercomputer at that location, incorporating over 23,000 GPUs. Nscale has also committed £2.5 billion (approximately $3 billion) to construct multiple modular data centers in the UK, though specific locations for these remain undisclosed.
The agreement signifies a strategic expansion of Google's UK infrastructure, which already includes a data center opened in Waltham Cross, Hertfordshire, in September 2025, and reported development plans in Teesside and Essex. For Nscale, securing a tenant of Google's caliber validates its data center strategy and provides a significant anchor commitment for its UK operations. This deal reflects the broader industry trend where hyperscalers are aggressively locking in capacity with specialized operators to ensure a steady supply of the powerful computing resources essential for training and running next-generation AI models, a critical factor in maintaining competitive advantage.
Source: datacenterdynamics
