The 2022 US CHIPS Act earmarked $52.8 billion, most for grants and subsidies to build giant semiconductor plants capable of producing AI chips. Europe’s 2023 Chips Act 1.0 mobilized about €11 billion in public funding, and the newly proposed Chips Act 2.0 shifts the focus toward building advanced foundries.
In contrast, the UK’s new AI Hardware Plan will unleash a mere pittance of public money — £1.1 billion. But the funds look well targeted, focusing on British strengths in chip design and research and on testing novel chips. The UK will support AI‑specific compute, AI‑relevant chips, and an AI talent pipeline, rather than a generic “chips for everything” agenda.
Design and research represent the UK’s comparative advantage. Arm-designed chips power almost all the world’s mobile phones. The plan will fund the design and testing of new chips to fit with the UK’s existing strengths: Arm‑style intellectual property, system‑on‑chip design, and cutting‑edge research in AI hardware and algorithms. In practical terms, every pound spent will go much further than vainly chasing a leading‑edge fab: it creates portable IP, exports, and high‑value jobs without locking the country into a single site or process node.
The contrast with previous UK chip strategies is striking. An earlier ten‑year semiconductor plan set out just £1 billion over a decade. But it recycled existing funds and lacked a convincing mechanism to translate research excellence into commercial success.
Unlike the European proposal to build a sovereign 1‑nanometer fab, the UK avoids promising to build a leading‑edge manufacturing champion, nor does it dangle tens of billions in subsidies to create a local copy of TSMC, Samsung, or Intel. The three companies dominate the global leading‑edge foundry market; launching a new entrant into that club from scratch would be an extraordinarily expensive and risky gamble, especially for a mid‑sized economy.
An interesting element of the British plan is a government commitment to buy chips and systems from “innovative start‑ups and British firms,” rather than limiting support to grants and R&D checks. By making “strategic purchases” of AI chips from UK‑based firms to keep them in the country, domestic companies receive guaranteed demand and a credible anchor customer.
That matters. Infant AI‑hardware companies often fail not for lack of technology, but for lack of customers willing to take the first risk. The UK government signaling that it will be a repeat buyer represents a powerful complement to venture capital and export markets.
The strategy also marks a cultural shift in how the UK thinks about ownership and control of strategic tech assets. When Japan’s SoftBank bought Arm in 2016, London let the purchase of a strategic tech champion breeze through, promoting “Britain is open for business.” If the new thinking were applied, it should have treated the sale as a national security concern.
Current Business Secretary Peter Kyle says he would have vetoed the sale of Arm had he been in government at the time, and has signaled a high “risk threshold” for allowing foreign takeovers of critical tech. The government now talks about increasing taxpayer investments in promising firms — an attempt to ensure that value created in the UK is not immediately sold offshore.
If the UK can anchor a cluster of AI hardware firms at home and retain more of their ownership, it becomes a reliable partner in a cutting-edge Western semiconductor supply chain — one that contributes valuable IP and products rather than simply buying chips designed and owned elsewhere. It would bolster a potential broad democratic alliance seeking to stay ahead of China.
Real risks remain. The UK’s £1.1 billion, while significant, remains modest compared to US and Chinese industrial programs; poorly targeted, it could vanish into a few failed bets. How procurement and “strategic purchases” are conducted will be crucial: if government buying is slow, bureaucratic, or steered towards politically connected firms rather than the best technology, it will do little to help genuine innovators.
Even if successful, the UK will continue to depend on foreign foundries — TSMC, Samsung, Intel — which means it must stay deeply engaged in alliances and export control regimes to ensure access to leading‑edge processes.
Yet, this strategy is the right model for a mid‑sized democracy, as opposed to the mega-fab dreams now circulating elsewhere in Europe. It puts real money behind AI‑specific hardware, uses the state as a customer to anchor new firms, focuses on design and innovation where the UK is strongest, and starts to treat ownership and control of critical tech as a strategic question rather than an afterthought.
It’s not a guarantee of success — but it improves the chances of building an AI hardware base that is innovative, resilient, and remains ahead of China’s innovation.
Bandwidth is CEPA’s online journal dedicated to advancing transatlantic cooperation on tech policy. All opinions expressed on Bandwidth are those of the author alone and may not represent those of the institutions they represent or the Center for European Policy Analysis. CEPA maintains a strict intellectual independence policy across all its projects and publications.
Tech 2030
A Roadmap for Europe-US Tech Cooperation
