UK Space Agency roadmap says Britain should lead on trusted AI for space, not computing scale

The UK Space Agency on October 27, 2026, published a 10-year Artificial Intelligence Roadmap for Space, setting out which AI technologies the British space sector should develop through 2036 and how. Its core argument is that the UK is unlikely to win on scale: it “may not out-scale global AI or constellation leaders, but it can out-integrate and out-govern them.” The agency published a companion Quantum Technology Roadmap for Space on the same day.

The agency commissioned the roadmap, and Space Professionals Partnership Ltd developed it for the agency and the Department for Science, Innovation and Technology (DSIT), after consulting more than 100 stakeholders across industry, academia and government. It is a set of recommendations and puts no figure on the investment it calls for.

Eight technologies, eight recommendations

The roadmap narrows a longer list of 12 AI functions to eight priorities: computer vision and image analytics; AI for simulation and testing, including digital twins; analytics, forecasting and discovery; machine learning, including learning from very few examples; planning, optimization and scheduling; agentic or autonomous AI for behavior and control; human-AI interaction; and edge computing. It maps these onto seven investment themes and turns the result into eight recommendations:

• A high-assurance national capability for space surveillance and satellite tracking, built on AI.
• AI to make the UK's own positioning, navigation and timing more resilient.
• AI across satellite communications, for security, control and quality.
• A national AI-for-Earth-observation ecosystem to support commercial and government decisions.
• Faster progress on autonomy and hybrid systems that combine physics models with AI.
• AI in launch economics and spaceport operations.
• Trusted national infrastructure for space data and AI.
• A national AI hub to coordinate the sector's AI work.

Assurance instead of scale

The document says the UK lags in investment in hyperscale computing infrastructure and lacks control over critical space assets in navigation, satellite communications and access to space. Instead, it argues that the country's advantage lies in “trusted, certifiable, resource-efficient AI” built into safety-critical systems. Its first immediate priority is for the UK to invest in verifying, validating and certifying AI for autonomous space systems, drawing on the AI Security Institute, the Alan Turing Institute and standards work linked to BSI. Other priorities include writing AI requirements into the agency's own funding and procurement, and low-power AI hardware, from neuromorphic computing to radiation-tolerant designs.

The roadmap is blunt about how the agency funds innovation today. That funding, it says, “operates mainly at project-level not programme-level, and lacks strategic prioritisation.” It says a national program that takes technologies from technology readiness level 3-4 toward market deployment could provide sustained investment. Otherwise, it warns, the UK risks remaining a strong contributor to others' programs rather than a leader.

Where the UK starts

The roadmap cites European missions as evidence that AI already works in orbit, while noting that these efforts remain small and experimental: ESA's PhiSat-1 CubeSat carried an AI chip that discarded cloudy images on board, and ESA's OPS-SAT served as a testbed for machine-learning experiments.

On the AI side, the UK has a large base to draw on. DSIT's AI sector study for 2024, published in September 2025, counted 5,862 AI companies, with £23.9 billion in AI revenue, £11.8 billion in gross value added and 86,139 AI-related jobs. The roadmap notes that most of these companies are not working on space, and that investment goes mainly to life sciences, transport, financial services and health.

To judge progress by 2036, it suggests tracking measures such as the number of AI-enabled missions deployed, reductions in manual intervention on satellites, the number of UK AI-space companies that grow beyond Series B funding, and how widely UK-developed AI certification frameworks are adopted abroad. The publication does not say how the agency will act on the recommendations or what funding, if any, will follow.