CSA Catapult becomes Semiconductor Catapult

CSA Catapult becomes Semiconductor Catapult

Compound Semiconductor Applications Catapult will transition this summer into the Semiconductor Catapult, widening its remit as the UK Government seeks to strengthen AI hardware and semiconductor scale-up.

The change follows publication of the UK AI Hardware Plan, which sets out more than £1.1bn of targeted public and private investment across innovation, skills, procurement, and investment. The refocused Semiconductor Catapult will be built on the capability already established by CSA Catapult.

CSA Catapult was established in 2018 as a not-for-profit centre of excellence for compound semiconductor technology. Its work covers measurement, characterisation, integration, and validation across power electronics, advanced packaging, RF and microwave, and photonics applications.

The Semiconductor Catapult will broaden that work across the wider UK semiconductor sector. Its role will include support for companies moving from concept and validation into commercialisation, supply-chain access, and deployment in AI infrastructure.

Caroline O’Brien, CEO of CSA Catapult, said: “By refocusing, we will broaden our support for UK companies in design, development and commercialization across the wider UK semiconductor sector. As the Semiconductor Catapult, we will work to increase the adoption of UK-designed semiconductor technologies in AI infrastructure, including data centers across the UK.”

The AI Hardware Plan also includes a £750m heterogeneous AI supercomputer for the AI Research Resource, a £400m procurement opportunity for specialised chips within that programme, a £120m AI hardware innovation investment, and an £80m semiconductor and AI hardware skills package.

The transition moves the Catapult model closer to the deployment problem facing UK chip companies. The country has long-standing strengths in chip design, compound semiconductors, photonics, advanced materials, and university research. The harder stage is often the move from promising device or architecture into validated system, production route, and commercial adoption.

AI hardware makes that gap more demanding. Data-centre efficiency, inference acceleration, optical interconnects, high-density power conversion, memory security, packaging, and thermal management all need to be evaluated inside realistic systems. Isolated component performance is not enough when customers need proof that a technology can be integrated, supplied, cooled, powered, and supported.

The UK already has several semiconductor developments pointing toward that deployment challenge. Cambridge GaN Devices’ automotive ICeGaN sampling shows how device-level innovation has to be translated into qualified power-system designs, while rising semiconductor equipment billings linked to AI demand underline the scale of capital and manufacturing pressure behind the market.

The Semiconductor Catapult will also have to support a broader mix of technologies than the original compound semiconductor remit suggested. Power electronics, RF, photonics, advanced packaging, AI accelerators, sensors, and memory interfaces are increasingly interdependent. The strength of a chip technology often depends on the package, power stage, thermal path, test strategy, and software environment around it.

A more broadly focused Catapult gives the UK a clearer institutional route for companies trying to move from prototype to deployable hardware. Its value will come from connecting design validation with manufacturing partners, procurement routes, investors, and customers prepared to adopt UK-designed semiconductor technologies at scale.