Commercial returns from the industries that decide what Britain can build.

Demand is rising structurally – energy security, industrial heat, hydrogen, data centres and defence all pulling at once. The constraint is not appetite; it is the cost and friction of building here. First-of-a-kind risk, licensing and planning, grid queues and the political weather are the timeline, so we price them rather than assume them away. And we treat the delivered cost of industrial power as the number that decides whether much else on this list ever gets built.

• What we look for
• Revenue contracted or regulated far enough out to underwrite
• First-of-a-kind, consenting and grid risk priced explicitly
• A route to cheaper industrial power, not only lower carbon

Infrastructure rewards owners, not traders. The return comes from utilisation, from the difficulty of building a second one next door, and from value accruing quietly in land and permissions. The common mistake is to price it as an asset when what you are buying is a business.

• What we look for
• A position that cannot easily be built again nearby
• Utilisation upside under new ownership
• Value held in land and permissions, not only in the tenancy

Process knowledge is the most under-priced asset in British industry. It lives in a line, a team and a set of tolerances that took years to settle, and it does not travel well – which is exactly why we want it. The same logic reaches the harder end of the sector: qualification, packaging and the fabrication capability a country either holds or imports.

• What we look for
• Advantage in process, not in marketing
• Already qualified into customer supply chains
• Capability a country would rather not import

World-class engineering, and a thin pool of capital willing to fund it through certification. Uncrewed and high-altitude systems in particular sit in the gap between a research grant and a procurement cycle, which is a gap patient capital is built for. We back the whole stack, including the unglamorous ground segment where the recurring revenue actually lives.

• What we look for
• A funded route through certification
• Recurring revenue, not one-off programmes
• Civil and defence relevance in the same asset

AI is moving out of experiment and into deployment, and the limits of the dominant models are becoming visible exactly as the stakes rise: weak continual learning, little explicit reasoning, poor traceability, and awkward behaviour anywhere constrained, regulated or disconnected. We are sceptical of anything that loses its edge the week the model underneath it upgrades. The durable layer sits below and beside that one – power, land and compute; data that was hard to assemble and cannot simply be bought; and systems welded so far into a regulated workflow that replacing them is a project rather than a decision.

• What we look for
• An advantage that survives the next model release
• Auditable where a black box will not do
• Runs in constrained, regulated or disconnected settings

The valuable position in health is not the tool; it is the layer underneath. Data that is owned rather than licensed, relationships that work at national scale, and delivery infrastructure that keeps producing more of both. Hold that and the routes to revenue run in parallel rather than in sequence. The ambition is to move from owning the infrastructure to owning the therapeutic assets created from it, which is a different business from selling software to the people who do.

• What we look for
• Owned or exclusive data, not licensed access
• A repeatable path from data to validated assets
• Sensitive data stewarded, and governed, in the UK

Planning is the real variable, so we price it rather than assume it away. Structural demand, a genuine local relationship, and a team that has finished something comparable. Regeneration only counts if it still stands up commercially in year ten.

• What we look for
• Structural demand, not speculative
• A delivery team with a finished comparable
• Consenting risk priced, not ignored

The least glamorous line and the most useful. It funds our own delivery, bridges the gap between commitment and drawdown, and keeps us at the table when a project moves faster than an institution can. Selected venture positions also buy early sight of what will matter to the industrial platforms in three years’ time.

• What we look for
• Funds group delivery, not only its own returns
• Discipline in credit and structuring
• Early sight of technology the other verticals will need