AI Vision for Energy

AI Vision for Energy: Startup Coalition’s response to the DESZN call for evidence on their Vision for the use of Ai in the energy sector.

Britain is building a decentralised energy system faster than it is learning to run one. Heat pumps, EVs, home batteries and plug-in solar are arriving in homes at scale. Wind farms are paid to switch off while demand sits idle behind the same constraint. The connections queue holds back gigawatts of generation, storage and demand. AI can coordinate all of this, but only if the firms building it can get to market.

The Government’s Vision for an AI-enabled clean energy system rests on three principles: lower bills, efficiency and decarbonisation, and security. We support all three, and add a fourth: competition and innovation. AI reaches consumers fastest when many firms compete to deliver it, not when a few incumbents control the data and the routes to market. Every intervention should be tested against one question: does it widen or narrow the field of firms able to innovate in the energy system?

The UK starts from a strong position. Its competitive retail market has already produced energy software platforms exported worldwide. It has an independent system operator, advanced flexibility markets and world-class AI research. Our advantage when applying AI to the energy sector today is not hardware, but the software, data and market-design layer that sits on top of a decentralised grid. That advantage holds only if the UK is the easiest place to deploy these products. If startups prove their products here but scale elsewhere, Britain loses the industrial prize.

Today it is not the easiest place to do this though. Startups in our network report the same barriers again and again:

• Data is locked away. The dataset that matters most, half-hourly consumption joined to the devices behind the meter, does not exist. Investment in UK Open Energy startups has never exceeded £16m a year, against a peak of £836m for Open Banking and Open Finance.
• The regulator has not prioritised innovation. The FCA’s sandbox received 142 applications in 2025/26 alone. Ofgem’s sandbox page records no new sandboxes in four years.
• Pilots do not become products. Networks are funded to trial, not to adopt, and are rarely rewarded when AI or flexibility avoids reinforcement.
• New rules risk penalising smarter optimisation. The Smart Secure Electricity Systems regime, as drafted, could capture AI-driven products that simple incumbent tariffs escape.
• Markets do not yet pay for flexibility by place and time. Constraint payments have already passed £1bn this year.

This report captures our response to the UK Government’s November 2026 call for evidence on its Vision for an AI-enabled clean energy system. In it, we set out 20 recommendations across six areas: 

1. Open data
2. Regulatory culture
3. Pilot-to-production
4. Market design and flexibility
5. Household control 
6. Agents, assurance, and risk

Taken and acted on together, we contend that this blueprint would make the UK the place where energy AI is tested, proven and scaled first, and where the benefits show up as pounds off bills and gigawatts connected.