- Lead. London Tech Week 2026 concluded this week with more than £6 billion in new AI investment commitments, led by AMD’s £2bn five-year partnership, Nebius’s £1.7bn UK compute build-out, and a £1.1bn government AI Hardware Plan unveiled by Prime Minister Keir Starmer.
- Fact. British AI startups raised £8.2 billion in venture capital in the first half of 2026, with London ranking third globally for VC behind the Bay Area and New York—a position the UK government is explicitly trying to lock in through a sovereign compute strategy targeting 5% of the global chip market.
- Stake. A £750m national supercomputer due by 2030 sits at the centre of the plan. Starmer framed the ambition in terms of commercial retention as much as industrial policy: “start here, scale here, and stay here.”
More than 30,000 attendees from 130 countries gathered in London this week for London Tech Week, where artificial intelligence featured in roughly half of more than 600 sessions. The investment announcements that emerged came with unusual specificity. AMD, the US chipmaker, pledged up to £2bn over five years for UK high-performance computing partnerships with Cambridge and Imperial universities, a new Sovereign AI Innovation Lab developed jointly with Dell, and stakes in British startups. Nebius, the cloud provider, committed approximately £1.7bn to build UK AI infrastructure through three separate Nvidia-hardware deployments reaching 65 megawatts by 2027, alongside an expansion of its London R&D hub.
The government’s £1.1bn compute plan
The centrepiece of the UK government’s response was the AI Hardware Plan announced by Starmer. Its components, according to The Next Web’s summit roundup: £750m for a national supercomputer due in 2030; £400m to acquire specialist AI chips; £150m for next-generation inference chips to be purchased from British firms as early as this summer; £120m for chip design; and £45m for engineering training. A separate £200m government adoption package targets public-sector AI deployment. London Mayor Sadiq Khan added a distinct £12m commitment to help the city’s small businesses conduct AI readiness checks and access mentoring programmes.
The cumulative ambition is a 5% share of the global chip market—an aggressive target for a country that currently holds a fraction of that figure and relies heavily on US and East Asian supply chains for advanced semiconductors. Whether the investment timeline is sufficient to close that gap before the next technology cycle is a point of genuine debate among industry observers, particularly given that the £750m supercomputer is not due until 2030.
Why the UK is pushing sovereign compute now
The push arrives as US export controls on advanced chips have tightened, and the question of who controls AI infrastructure has moved from a technical concern to a geopolitical one. British AI startups have raised £8.2bn in VC in the first half of 2026 alone, and the country’s tech sector is valued at £1.2 trillion. That momentum creates political pressure on government to ensure companies benefiting from British research and talent do not simply relocate once they achieve scale—the retention problem Starmer’s “start here, scale here, and stay here” framing explicitly addresses.
The regulatory dimension runs in parallel. The UK’s Competition and Markets Authority has an active probe into how dominant cloud and AI players bundle their services—covered in UK watchdog launches Strategic Market Status probe into Microsoft’s Office and Copilot bundle—reflecting the same underlying concern about infrastructure control. The London Tech Week pledges represent the investment face of that same question: whether the UK can build enough sovereign capacity to have a credible alternative to depending entirely on a small number of US hyperscalers for its AI future.