en.Wedoany.com Reported - The UK government has announced a £1.1 billion AI hardware plan aimed at building a national computing strategy, including a £750 million investment in a supercomputer in Edinburgh and support for domestic chip startups.
In terms of funding and computing power, the UK government will allocate £750 million to build a new national AI supercomputer, a heterogeneous hybrid chip system, planned to be operational by 2030 at the University of Edinburgh in Scotland. This computer will operate as part of the UK's AI Research Resource, enabling researchers and businesses to share access to a national supercomputing network, including the Isambard-AI supercomputer in Bristol and the upgraded Zenith/DAWN supercomputer in Cambridge. According to the plan, £400 million is specifically earmarked for purchasing "next-generation" AI chips, including £150 million for inference chips and another £250 million for "more specialized chips." The government stated it will target procurement of new chips from innovative startups and UK companies. Additionally, £120 million will go to a new hardware innovation program, providing funding flexibility for startups. The tender process for the new supercomputer will soon begin, aiming to provide researchers, startups, and public services with the computing power needed to safely develop and run AI in the UK. At London Tech Week, US chipmaker AMD announced it is collaborating with Dell at the University of Cambridge to develop the Zenith supercomputer and the Sunrise fusion supercomputer system. Sunrise, in partnership with the UK Atomic Energy Authority (UKAEA), targets healthcare research, climate modeling, materials science, engineering simulation, fusion research, and scientific AI model development. AMD also launched a new project with Imperial College London to support scientific research reliant on large-scale computing resources.
In terms of capital and infrastructure, AMD has committed to investing approximately £2 billion in the UK over five years, supporting chip supply and new cloud capacity construction. Dutch new cloud data center provider Nebius has pledged £1.75 billion for capacity building. Nebius deployed its first GPU infrastructure in the UK last November and said it will add three new sites, adopting the latest NVIDIA full-stack AI factory technology, aiming to reach 65 megawatts by 2027. The company stated that UK businesses such as financial firm Revolut and biotech company Prima Mente are already running large-scale production workloads on its platform. The British Business Bank has also supported a new UK fund led by Silicon Valley investor Playground Global, investing £150 million in UK AI hardware companies—the bank's largest ever fund investment. Playground Global will open its first office outside the US in the UK. Technology Minister Liz Kendall said: "The UK is already a global leader in chip design, and I believe this is a race the UK can win." Additionally, the government has invested £20 million to expand a "Scaling Reasoning Lab" co-hosted by the Advanced Research and Invention Agency and CommonAI, along with £45 million for AI training, bringing the AI Skills Fund to £80 million, and plans to establish a new £12 million university doctoral training center.
In the network and model domain, UK startup Oriole Networks is collaborating with AMD at the reasoning lab, preparing to deploy a scaled external system that uses light instead of electrical signals to move data between data center chips. Its photonic network solution, PRISM, has been developed and tested in the lab as the "world's first fully photonic AI network," with broader industry rollout expected by 2027, potentially improving inference throughput and interactivity by an order of magnitude. The solution replaces electronic switches in the network core with optical circuit switching, reducing GPU power consumption by 81% and cutting GPU idle time from the typical 60% to less than 1%. Another startup, Cosine, has been selected to lead the UK's first fully sovereign frontier AI model, Lumen Sovereign. Supported by the UK government's Sovereign AI program, the project has convened organizations including BAE, Babcock, BT, Era4, HSBC, Leonardo, Lloyds, LSEG, NatWest, PwC, Telefónica, Thales, and the Alan Turing Institute to assist in its design. The open-weight Lumen Sovereign model will use Isambard-AI and is planned for release by the end of 2026, with approximately one-third of its total training compute time dedicated to optimizing complex software engineering problems and two-thirds focused on critical mission use cases. Industrial AI digital twin company PhysicsX announced an oversubscribed $300 million Series C funding round, valuing it at $2.4 billion. The round was led by Temasek, with participation from M&G Investments, Intrepid Growth Partners, Applied Materials, Atomico, General Catalyst, July Fund, NGP, and Radius, with NVIDIA and Siemens also reinvesting. Its AI models can predict physical behavior in seconds, used by companies in aerospace, defense, energy, semiconductors, automotive, and data centers. The company said revenue doubled year-over-year in the past year, booked revenue tripled, customer numbers doubled, and headcount reached 300.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
