AI applied sciences which may remodel how individuals warmth and insulate their houses are among the 10 clear vitality improvements to have reached the ultimate of the £2 Manchester Prize.
The finalists embrace a non-intrusive system that makes use of AI to design bespoke panels that flip bricks into radiators that heat houses from the skin in, streamlining the method for older housing inventory transitioning to utilizing warmth pumps.
Different finalists embrace AI-enabled heat-mapping drones to assist councils and housing associations pinpoint social housing inventory in want of insulation upgrades extra effectively, and an AI know-how to determine supplies in end-of-life photo voltaic panels after which recycle them into new photovoltaic cells as an alternative of going to landfill.
The Manchester Prize, funded by the Division of Science, Innovation and Expertise and delivered by Problem Works (a part of the Nesta group), “is rewarding UK-led AI breakthroughs that assist the general public good, together with rising the financial system, bettering public companies and serving to to create a simply transition to Web Zero for everybody.”
Secretary of State for Science, Innovation and Expertise, Peter Kyle stated:
“AI is opening up transformative new methods to deal with local weather change and assist the UK’s ambition to turn into a clear vitality superpower.
“That features utilizing the know-how to maintain our houses heat, whereas additionally supporting tasks which is able to use AI to slash carbon emissions in our cement and metal industries – sectors which account for 16% of worldwide emissions.
“That is how we ship our Plan for Change – harnessing innovation to resolve main challenges, lower vitality payments, and enhance lives throughout Britain.”
Power Secretary Ed Miliband stated:
“Clear energy is the financial alternative of the twenty first century and these tasks will assist households and companies make the most of decrease payments, in a better and quicker means than ever earlier than.
“From specifically designed radiator partitions to a wise energy grid that flicks on and off as we’d like, AI has the potential to assist each house in Britain to really feel the advantages of hotter houses and homegrown clear vitality.”
Heating houses from the skin in
The ten finalists of the Manchester Prize embrace EnergyWall by Underheat, in partnership with College of Salford. EnergyWall upgrades a constructing’s partitions, gently warming or cooling houses from the skin, turning bricks into radiators that keep a cushty inside temperature all 12 months spherical.
Utilizing AI to analyse a constructing and off-site manufacturing, it designs and installs pipe methods into insulation panels for the partitions of a constructing, making retrofitting buildings with warmth pumps quicker, cheaper, and fewer disruptive. This strategy is right for social housing, serving to cut back carbon emissions, lower vitality payments, and deal with condensation that causes mould.
It’s a better, scalable solution to decarbonise heating and combat gasoline poverty throughout the UK.
Giving photo voltaic panels a second life
Inexperienced Loops by College of Wolverhampton, in partnership with ABCircular GmbH Berlin, tackles the problem of recycling end-of-life photovoltaic (PV) cells by creating high-efficiency photo voltaic panels from recycled supplies. Whereas there are tens of thousands and thousands photo voltaic panels within the UK, the specialist infrastructure to dispose and recycle of them is presently missing.
Inexperienced Loops makes use of machine studying to analyse the optical properties of supplies and constructions of photo voltaic cells. Utilizing extremely conductive artificially engineered MXene-based metamaterials, Inexperienced Loops optimises the design of photo voltaic cells to boost vitality efficiency whereas decreasing manufacturing prices.
With the rising e-waste drawback from previous photo voltaic panels, the know-how helps cut back waste, helps a round financial system, and makes photo voltaic vitality extra sustainable and accessible.
A “Google Maps” for house warmth loss
Fast Thermal Efficiency Evaluation algorithms (RaThPAs) by Kestrix makes use of AI and thermal drones to map warmth loss throughout total neighbourhoods, performing as quick, 3D vitality surveys from the sky.
This helps stakeholders like utilities, councils and housing suppliers plan vitality upgrades with fewer pricey, time-consuming web site visits. Like a “Google Maps of warmth loss,” the system exhibits the place buildings are leaking warmth and recommends fixes.
With a crew of consultants in laptop imaginative and prescient and physics, Kestrix goals to hurry up house retrofits, in flip reducing emissions, saving households cash, and making houses hotter and more healthy at scale.
Julia King, Baroness Brown of Cambridge, chair of the Manchester Prize judging panel stated:
“We’re at a important juncture within the journey to web zero, the following decade is make or break if the world is to maintain world temperatures rise to effectively under 2C by 2050. World emissions have to halve by 2030 in comparison with 1990 ranges if we’re to remain on monitor, whereas electrical energy manufacturing might want to double by 2050 to satisfy the calls for of an electrified financial system – clear vitality innovation is crucial. The speedy development of AI means we now have instruments like by no means earlier than to realize the purpose of decarbonising the financial system whereas supporting people, communities and companies to thrive.”
Different options within the operating embrace an AI utility to enhance the manufacturing of biogas the place accessible materials inputs can differ from daily, AI to dramatically lower the CO2 emissions from concrete manufacturing (chargeable for 8% of worldwide emissions) and from metal manufacturing (chargeable for an extra 8% of worldwide emissions), and wider emissions from trade.
Different finalists embrace AI applied sciences to assist the logistics trade lower its emissions, and AI to make sure the vitality grid stays balanced always as extra of our vitality comes from wind and photo voltaic.
The ten groups behind the superior AI options have every obtained £100,000 in seed funding, plus compute credit price £60,000, and extra non-financial assist to develop options able to successful the £1 million grand prize in spring 2026. The successful answer will display not solely technical innovation, but additionally an evidenced highway map to near-term (2030) adoption, scale and influence.
In March 2025, the Manchester Prize in its inaugural 12 months awarded £1m to Polaron – an AI know-how that dramatically accelerates the event of recent superior supplies from years to a matter of days. Rising the vitality capability of electrical automobile batteries by 10%.
A listing of the finalists is included under. Go to manchesterprize.org to search out out extra concerning the Manchester Prize and all 10 finalists.
Manchester Prize (12 months 2) finalists
Agent Web Zero by College of Sheffield and AMRC. Agent Web Zero is an progressive AI system that helps industrial firms turn into extra sustainable by analyzing their environmental influence in real-time. The system repeatedly screens vitality utilization and emissions by connecting to varied knowledge sources throughout operations. Utilizing superior AI strategies, Agent Web Zero identifies environmental hotspots and robotically suggests sensible enhancements. This offers companies clear, actionable insights to cut back their carbon footprint whereas sustaining productiveness and competitiveness, primarily offering a “sustainability assistant” that works 24/7 to assist firms obtain their net-zero targets.
BiofuelAi by College of Surrey. BiofuelAi brings cutting-edge AI and machine studying to the biofuel trade, optimising complicated, variable processes in actual time. Conventional biogas manufacturing usually depends on operator instinct attributable to unpredictable organic methods as a result of biofuels are constituted of a number of materials inputs. BiofuelAi solves this with superior predictive fashions that create a digital twin of every web site, enabling whole-system optimisation – from every day feedstock recipes to long-term acquisition methods. Developed by AI and sustainability consultants, the platform boosts effectivity, profitability, and environmental influence, providing a scalable answer for cleaner, data-driven vitality manufacturing worldwide.
Carbon Re by Carbon Re. Cement varieties the inspiration of our trendy world nevertheless it has a sustainability drawback – it’s chargeable for round 8% of worldwide CO₂ emissions. Carbon Re is tackling this problem by constructing AI course of management software program to chop emissions in cement manufacturing. Performing like self-driving for industrial crops, Carbon Re optimises industrial processes in real-time, serving to producers lower each prices and carbon whereas transitioning to low-carbon operations. A joint spin out of College School London and the College of Cambridge, Carbon Re was based to ship rapid local weather influence for heavy trade.
Cavolo by Kale AI. Cavolo makes use of superior AI to make metropolis deliveries extra environment friendly and eco-friendly. The system helps companies swap from conventional supply vans to Gentle Electrical Autos (LEVs), that are extra environment friendly in busy cities. By utilizing AI, Cavolo optimises supply routes in real-time, decreasing site visitors, vitality use, and emissions. The know-how helps make city logistics quicker and greener, permitting companies to ship items shortly whereas saving time and decreasing their environmental influence.
Deep.Optimiser-PhyX by Deep.Meta. Deep.Meta is tackling carbon emissions within the metal trade with an AI-powered Digital Twin – a wise digital duplicate of the manufacturing course of that mixes physics and machine studying to optimise furnace operations. By utilizing real-time sensor knowledge and materials science, Deep.Meta extra precisely predicts metal slab temperatures and improves scheduling, boosting vitality effectivity and considerably reducing emissions. Not like black-box AI, which might discourage adoption, Deep.Meta’s explainable, physics-based fashions supply clear reasoning, constructing belief with customers. Based by consultants in metallurgy and machine studying, Deep.Meta is already partnering with world steelmakers and goals to scale via broader trade collaboration.
DRIVE (Deep Re-enforcement studying for Clever Car and Power optimisation) by Versatile Energy Methods. Versatile Energy Methods (FPS) helps massive fleets like vans, vehicles, and buses swap to electrical by managing autos, chargers, and schedules with sensible software program. FPS makes use of superior AI known as Deep Reinforcement Studying to resolve complicated, fast-changing issues – like the place and when to cost – extra shortly and effectively. After coaching in a digital world, the AI could make sensible selections in actual time. First utilized in EV fleets, this know-how might additionally assist with greater vitality challenges sooner or later.
EnergyWall by Underheat, in partnership with College of Salford. EnergyWall upgrades a constructing’s partitions, gently warming or cooling houses from the skin, turning bricks into radiators that keep a cushty inside temperature all 12 months spherical. Utilizing AI to analyse a constructing and off-site manufacturing, it designs and installs pipe methods into insulation panels for the partitions of a constructing, making retrofitting buildings with warmth pumps quicker, cheaper, and fewer disruptive. This strategy is right for social housing, serving to cut back carbon emissions, lower vitality payments, and deal with condensation that causes mould. It’s a better, scalable solution to decarbonise heating and combat gasoline poverty throughout the UK.
Inexperienced Loops by College of Wolverhampton, in partnership with ABCircular GmbH Berlin. Inexperienced Loops tackles the problem of recycling end-of-life photovoltaic (PV) cells by creating high-efficiency photo voltaic panels from recycled supplies. It makes use of machine studying to analyse the optical properties of supplies and constructions of photo voltaic cells. Utilizing extremely conductive artificially engineered MXene-based metamaterials, Inexperienced Loops optimises the design of photo voltaic cells to boost vitality efficiency whereas decreasing manufacturing prices. With the rising e-waste drawback from previous photo voltaic panels, the know-how helps cut back waste, helps a round financial system, and makes photo voltaic vitality extra sustainable and accessible.
Grid Stability by College of Manchester. For electrical energy grids to perform, there have to be stability between the electrical energy going into the grid and the electrical energy leaving it. Grid Stability Monitor makes use of AI and machine studying to shortly analyse energy grid stability as extra low-carbon applied sciences like wind, photo voltaic, EVs and warmth pumps join. It replaces gradual, complicated simulations with speedy, AI-driven assessments, enabling real-time monitoring, quicker decision-making, and extra assured planning. This helps grid operators keep reliability whereas scaling up clear vitality options and reducing emissions.
Fast Thermal Efficiency Evaluation algorithms (RaThPAs) by Kestrix. Kestrix makes use of AI and thermal drones to map warmth loss throughout total neighbourhoods, performing as quick, 3D vitality surveys from the sky. This helps stakeholders like utilities, councils and housing suppliers plan vitality upgrades with fewer pricey, time-consuming web site visits. Like a “Google Maps of warmth loss,” the system exhibits the place buildings are leaking warmth and recommends fixes. With a crew of consultants in laptop imaginative and prescient and physics, Kestrix goals to hurry up house retrofits, in flip reducing emissions, saving households cash, and making houses hotter and more healthy at scale.
