Susan Bowen
CEO, Digital Catapult
AI-first startups could add £550 billion1 to the UK economy over the next decade, but only if the compute and connectivity gap is addressed.
These startups rely on high-performance compute and seamless connectivity to develop, train and deploy AI solutions. Without sufficient compute and connectivity, however, AI-first SMEs face barriers to scaling, slower innovation cycles and a weakened international standing.
Building the foundations to scale AI innovation
Digital Catapult is accelerating the practical application of AI across industries, helping AI-first startups turn innovation into commercial success. Accelerator programmes, for example, drive AI adoption across multiple sectors, including the creative industries where AI is driving the rise of ‘CreaTech’ and boosting the UK’s creative output. Closing the compute and connectivity gap will ensure that AI-first startups have the infrastructure they need to scale.
Closing the gap with targeted intervention
This infrastructure includes shared compute environments, open networks and simulation tools that bridge the gap between experimentation and real-world deployment. Digital Twin platforms allow startups to simulate and refine AI-driven systems in real-world environments. This May, the UK Digital Twin Centre opens in Belfast. It offers compute resources to bridge R&D and commercialisation, supporting the UK Government’s 31-point AI Action Plan to scale safe, effective AI innovation.
Closing the compute and connectivity
gap will ensure that AI-first startups have
the infrastructure they need to scale.
Digital Catapult also helps close the gap through cutting-edge facilities, innovation programmes and expert consultancy. Our testbeds provide SMEs with access to compute power and open networks, enabling them to develop and scale AI solutions. We also support AI-first SMEs through the Innovate UK BridgeAI programme. Over the past decade, we have facilitated £40 million of investment into CreaTech SMEs, turning bold ideas into commercial success.
Making AI innovation market-ready
Despite the availability of high-performance compute platforms like Isambard 3, CFMS, and the Co-Star Network, many AI-first startups still face barriers moving from R&D to commercialisation. High-cloud compute costs and limited digital support make scaling difficult, highlighting the need to close the compute and connectivity gap further, so startups can market new AI solutions. Targeted intervention will be necessary, pushing the boundaries of AI-driven innovation.
[1] Industry and Parliament Trust. 2025. Becoming an AI Superpower: Innovating for Economic Growth.
