The UK Government is to spend £520m on the purchase or development of advanced semiconductors as part of a £1.1bn AI Hardware Plan, which will also fund the expansion of the Scaling Inference Lab and help train the next generation of engineers, programmers and chip developers.
Announced by the Technology Secretary Liz Kendall at London Tech Week, the plan will devote £750m to the construction of a national AI supercomputer, of which £400m will be directed towards providing said computer with “next-generation” semiconductors, including state-of-the-art inference chips. Another £120m will fund UK-based R&D into new chips, while the British Business Bank will invest up to £150m in UK-based providers of AI hardware.
The £50bn payback
At a time when the UK AI industry is attracting billions in private funding, the plan could help maintain or even accelerate the industry’s momentum, with some companies telling Raconteur that it will positively influence the direction of travel over the coming years. Yet with the plan not stipulating that chip purchases must be from UK-based companies, and with several prominent British semiconductor companies under foreign ownership, it will also boost non-domestic AI sectors
The British government estimates that, if the UK could capture only 5% of the global AI semiconductor market by 2030, it would create “tens of thousands” of well-paid jobs and provide $50bn in annual revenues. It’s within such a context that it has announced its hardware plan, with Liz Kendall framing the initiative as one that will support UK-based companies, enabling them to develop and sell chips for AI-related purposes.
With several British semiconductor companies under foreign ownership, it will also boost non-domestic AI sectors
“That is exactly what this plan does, backing the British firms developing the next generation of AI hardware, so we get more jobs, more growth, and more control over the technologies our future depends on,” the Technology Secretary said. “We are backing Britain because we believe in Britain.”
Such backing is likely to materialise in relation to the aforementioned £750m AI supercomputer, which the Department for Science, Innovation and Technology (DSIT) suggests will be one of “the most advanced in the world” when it comes online in 2030. It will operate using a heterogeneous mixed chip system, and will provide researchers, startups and also public services with the compute they’ll likely need to test future AI applications.
Compute and chips
Importantly, ministers say they “want” to see British-designed chips being used in this system, although DSIT offers no guarantees that the money it has earmarked for chip purchases will actually be spent on British semiconductors. Instead, its press release merely notes that £150m will be used to purchase inference chips, with British firms “well placed to compete” for any corresponding tenders. Meanwhile, another £250m will fund the purchase of “more specialised chips,” and here there’s no mention of UK-based companies having any advantage in this area.
Regardless of the possibility that much of the plan’s budget may be spent on non-UK companies, those British firms that are in line to receive some funding are generally positive about what the UK Government is doing to stimulate the nation’s AI and semiconductor industry. Speaking to Raconteur, Oriole’s VP of Business Development & Marketing, Joost Verbek, says that the plan is important because it takes a “full system approach” that connects the various layers of AI research, development, production and procurement.
“Government is becoming a customer as well as an investor”
“For companies like Oriole, the key shift is that government is becoming a customer as well as an investor, creating real demand for new technologies,” he says. “Alongside this, access to national compute infrastructure and dedicated scale-up capital makes it easier to build and deploy AI hardware in the UK, not just invent it.”
Based in London, Oriole describes itself as a photonic networking company, developing components that can transmit data between semiconductors using light rather than electrical signals. It’s receiving support from the Scaling Inference Lab, which will help British AI companies test their solutions and products at scale, and which is receiving £20m in extra funding to help it expand its operations.
Combined with the funding for the AI supercomputer and its semiconductors, and combined with an additional £45m in support for AI training (which is now receiving £80m in total), Verbek says that the plan provides a “strong signal” that the UK is serious about providing the infrastructure for a thriving AI sector. He adds, “It should make the UK more attractive, particularly for companies working on the next generation of AI systems.”
A longer-term shift from strategy to execution
The government’s announcement comes at what may be a pivotal moment for the global AI industry. The EU has recently announced its Technological Sovereignty Package, which includes measures to support and prioritise European AI and semiconductor companies. And across the pond, not only has the US already announced a private $500bn AI investment initiative, but the Trump administration is now considering taking federal stakes in labs such as OpenAI.
In such a context, the £1.1bn in funding (mostly for an AI supercomputer) may not seem all that momentous, yet Verbek affirms that the particular number is not of overriding importance at this stage, and that what matters most is the direction of travel. Similarly, the plan is good news because it shows that the government recognises that a healthy AI sector requires a solid infrastructure layer as its foundation, on which the model and app layers can then be built.
“It’s a strong foundation, but maintaining competitiveness will require sustained investment over time, particularly in large-scale deployment,” he concludes. “In our opinion, this is best seen as the start of a longer-term shift from strategy to execution, rather than a one-off intervention.”
Speaking to Raconteur, the DSIT also highlights the systemic structure of the plan, how it doesn’t simply focus on one stage of activity, but provides help across the entire lifecycle of new technologies.
A spokesperson for the department said, “Our AI Hardware Plan backs British chip firms at every stage, from invention through to procurement and scaling, targeted at sovereign capability in the technologies that matter most, rather than replicating every part of global supply chains.”
The department also affirms that the plan will operate in parallel with the £500m Sovereign AI Unit announced in April, meaning that the UK Government is now providing support to AI developers and infrastructure companies alike.
The UK Government is to spend £520m on the purchase or development of advanced semiconductors as part of a £1.1bn AI Hardware Plan, which will also fund the expansion of the Scaling Inference Lab and help train the next generation of engineers, programmers and chip developers.
Announced by the Technology Secretary Liz Kendall at London Tech Week, the plan will devote £750m to the construction of a national AI supercomputer, of which £400m will be directed towards providing said computer with “next-generation” semiconductors, including state-of-the-art inference chips. Another £120m will fund UK-based R&D into new chips, while the British Business Bank will invest up to £150m in UK-based providers of AI hardware.
