UK government publishes AI Opportunities Action Plan

The Starmer government published the AI Opportunities Action Plan, a strategy document written by government AI adviser Matt Clifford at the request of technology secretary Peter Kyle, setting out 50 recommendations for expanding Britain’s AI infrastructure, adoption and domestic capability. The government’s accompanying response said it would accept or was already acting on the great majority of them.

The plan’s central commitment was to expand the UK’s AI Research Resource — the country’s public compute capacity — by “at least 20x” by 2030, alongside a ten-year plan for sovereign, domestic and internationally sourced compute. It proposed designating “AI Growth Zones” with streamlined planning approval for data centres and associated power infrastructure, naming the former Culham Science Centre nuclear-fusion site as an early candidate and prioritising post-industrial and coastal areas for the investment. On adoption, it set a “Scan > Pilot > Scale” framework for AI deployment across government, targeting a 20% time saving from AI assistants in public-sector work, and committed £21 million to an NHS AI diagnostics fund already under way. On talent, it proposed new fellowships and a scholarship programme aimed at training tens of thousands of additional AI specialists, and called for a new “UK Sovereign AI” unit to support frontier AI development partnerships domestically.

The plan positioned Britain’s approach as adoption- and infrastructure-led rather than regulation-led, in contrast to the EU’s AI Act, and followed the collapse of Britain’s own attempt at binding AI legislation earlier in the parliament. It set the framework the government worked from through the rest of 2025, including subsequent announcements on data centre sites and compute procurement.