UK + European semiconductor strategy
As governments treat semiconductors and AI as strategic infrastructure, most UK and European strategy work has concentrated on fabrication, packaging and skills. Design capability receives less attention. However, every chip starts with EDA software, processor IP and verification methods, and three companies supply most of the EDA tools used worldwide: Synopsys, Cadence and Siemens EDA. Two of them are American.
Sovereignty as capability
For the UK and Europe, semiconductor sovereignty is best measured in capability: the engineers, tools, methods and companies that can design and verify the chips a country depends on. Europe has real strengths to build on: strong universities, an active RISC-V community, and safety-critical industries, such as automotive, aerospace and industrial control, where verification, security and trust matter most.
What AI changes
As AI moves into design and verification flows, the models, the training data and the compute behind them become part of the same strategic question. Three issues stand out:
- Training data. AI tools for chip design need design data to learn from, and most of it is proprietary. Who owns it, and who can license it, will shape who can build these tools.
- Regulation. AI used in engineering will have to meet rules such as the EU AI Act, with measurable trust and independent verification.
- Open standards. RISC-V reduces dependence on proprietary processor architectures. It also moves the verification effort onto whoever builds each processor, which needs tools and skills that Europe has in only a few places.
How I contribute
I have spent my career building EDA and processor technology companies in the UK and the US, and now research AI+EDA at the University of Southampton. I am interested in contributing to this work with government, think tanks, universities and industry bodies, through workshops, roundtables, technical review of policy papers and independent analysis. More on the Strategy & Startups page.
Questions I am exploring
- What do semiconductor and AI sovereignty mean in practice for the UK and Europe, and which parts of chip design capability do they need to control themselves?
- What AI+EDA capability should the UK and Europe build for themselves, and where can their universities, RISC-V community and safety-critical industries lead?
- Who owns, and who can license, the design data that AI tools for chip design need to be trained on?