HWNEWS
Liz Kendall's AI chip industrial strategy concedes the chip fabrication plants, the most capital intensive link, and bets on chip design IP, licensed circuit blocks that chip designers buy and sell, and research, where Arm, the UK chip designer
Liz Kendall stood at London Tech Week in June 2025 and pitched a £500 million (about $671 million) Sovereign AI fund aimed at capturing 5% of a global AI chip market the government projects at $1 trillion by the early 2030s, framed as a $50 billion revenue opportunity. The bet is the part that matters. The UK is conceding chip fabrication plants, the $20 billion-plus link in the chip stack, and trying to win on chip design, IP, and research, where Arm and UK universities give it an inherited base.
Kendall, the Technology Secretary, has argued the UK is better placed to compete on design and research than on building fabs. The rest of the AI Hardware Plan is policy scaffolding around that trade.
The UK's bet is the smallest by a wide margin, and the difference is not stylistic. The US has put $53 billion in CHIPS Act subsidies plus tax credits into onshore fab buildouts. The EU Chips Act785742), in force since 2023, targeted €43 billion (about $50 billion); capital actually drawn onto the continent has crossed €80 billion (roughly $93 billion), well past the original envelope. India's Semiconductor Mission 2.0 sits at Rs 1.2 lakh crore (about $12.5 billion), aimed at building a domestic fabrication base from a near-zero starting point. The UK has put £500 million (~$671 million) into a Sovereign AI fund, inside a £1.1 billion total package anchored on design, IP, and advanced packaging.
Read as a leaderboard, the UK loses. Read as a stack-segment bet, the comparison changes.
A modern leading-edge fab, the factory where chips are physically etched onto silicon wafers, now costs more than $20 billion and takes several years to bring online. For a mid-sized economy, the marginal return on a single such bet is negative once labor, tooling, and the talent pipeline are priced in. The US can absorb the cost in exchange for defense-grade supply security; the EU and India are using public capital to attract private fabs that would not otherwise land; the UK has chosen not to play that hand.
The UK's Council for Science and Technology, which advises the Prime Minister, has counseled the same niche/architectural approach Kendall is now executing: anchor on Arm, the UK-headquartered chip designer whose architecture underpins most mobile phones, and on the UK's university research strengths, rather than chase a full-stack rebuild the budget cannot fund. Industry coverage reads the same plan as a niche bet, a specialist's framing of a strategy the UK government presents as a market opportunity.
The fund flows into three buckets. Design and IP funding goes to fabless chip designers, the companies that design chips and license the IP but outsource manufacturing; the UK already hosts Arm and a wider set of fabless chip designers. Research funds university labs, doctoral training, and shared prototyping facilities, with the AI Hardware Plan committing to a national compute strategy and skills pipeline. Advanced packaging, the post-fabrication step where chips are assembled into modules, gets the rest. The UK has academic depth in packaging, and the capital intensity is a fraction of a fab's.
The right comparison is not the $20 billion fab. It is the design license, the prototype run, and the packaging line, all of which fit inside a sub-billion-dollar envelope.
Conceding the most capital-intensive link costs strategic leverage: defense supply, export controls, and the ability to set terms with hyperscalers. The Council's advice and Kendall's framing both acknowledge the trade in policy language. The harder question is whether the inherited base, Arm plus the university labs, can compound fast enough to make the bet pay before a US, EU, or Indian fab absorbs the design talent the UK is counting on.
A £15 billion-plus UK fab announcement in the next 24 months would settle whether the country can keep its hands off that link. The current bet, design and research over fabs, is a defensible specialization that is also partly a budget constraint dressed as one.