Deep dive: AI for energy & emissions optimization — the fastest-moving subsegments to watch

The AI for energy and emissions optimization sector has grown from a niche technology category into a sprawling ecosystem encompassing dozens of distinct subsegments, each with its own competitive dynamics, capital requirements, and deployment trajectories. While the overall market reached $11.2 billion in 2025 according to BloombergNEF, growth has been far from uniform. Certain subsegments are accelerating at 40-60% compound annual rates while others have plateaued or entered consolidation. This deep dive identifies the fastest-moving subsegments, examines the structural drivers behind their momentum, and maps the implications for investors, operators, and policymakers in the UK and beyond.

Why This Analysis Matters Now

The UK's commitment to net zero by 2050, combined with the legally binding carbon budgets set by the Climate Change Committee, creates a regulatory environment that rewards early adoption of AI-driven energy optimization. The UK's Sixth Carbon Budget requires a 78% reduction in emissions by 2035 relative to 1990 levels, a target that cannot be achieved through hardware upgrades and fuel switching alone. AI optimization of existing infrastructure represents one of the few pathways to bridge the gap between current trajectory and legally mandated targets at acceptable cost.

Simultaneously, the global energy crisis triggered by geopolitical disruption has shifted corporate priorities from voluntary sustainability reporting toward operational cost control. UK industrial energy prices rose approximately 65% between 2021 and 2024, making energy efficiency investments with rapid payback periods increasingly attractive to CFOs who previously treated them as discretionary. This convergence of regulatory pressure and economic self-interest has created a receptive market environment that favours specific AI subsegments over others.

Subsegment 1: Grid-Edge Intelligence and Flexibility Markets

The fastest-moving subsegment in 2025-2026 is grid-edge intelligence, encompassing AI systems that optimise distributed energy resources (solar, batteries, EV chargers, and flexible loads) at the point of grid connection. UK-specific momentum is driven by National Grid ESO's transition to the National Energy System Operator (NESO) and the expansion of flexibility markets through the Demand Flexibility Service and the Balancing Mechanism.

The UK's flexibility market grew from approximately 3 GW of contracted capacity in 2023 to 7.4 GW by the end of 2025, with AI-optimised assets capturing a disproportionate share of value. Octopus Energy's Kraken platform manages over 6 million customer accounts globally and uses machine learning to orchestrate behind-the-meter assets (heat pumps, batteries, and EV chargers) in response to half-hourly pricing signals. In 2025, Kraken-managed assets earned an average of 23% higher returns from flexibility services compared to assets operating under simple time-of-use schedules.

Open Energy (formerly Electron) has deployed blockchain-integrated AI for local flexibility trading in multiple UK distribution network operator (DNO) regions, enabling assets as small as 5 kW to participate in automated flexibility auctions. Their platform processed over 2.1 million flexibility trades in 2025, up from 340,000 in 2024, representing a six-fold increase that reflects both platform maturation and market expansion.

Modo Energy's analytics platform provides independent intelligence on GB battery storage revenues and grid conditions, helping operators optimise bidding strategies across multiple revenue streams. Their data shows that AI-optimised battery storage assets achieved average revenues of £85,000-£110,000 per MW in 2025, compared to £55,000-£70,000 per MW for assets using rule-based strategies, a 50-60% premium directly attributable to algorithmic sophistication.

The structural driver behind this subsegment's acceleration is the UK's aggressive deployment of distributed energy resources. BEIS data indicates that domestic solar installations exceeded 1.3 GW of new capacity in 2025, while residential battery storage installations grew by 78% year-on-year. Each new distributed asset creates demand for AI orchestration, generating a self-reinforcing growth dynamic.

Subsegment 2: Industrial Process Emissions Monitoring and Reduction

The second fastest-moving subsegment combines continuous emissions monitoring systems (CEMS) with AI-driven process optimisation for heavy industry. This subsegment has accelerated sharply due to the UK Emissions Trading Scheme (UK ETS) price trajectory and the impending implementation of the Carbon Border Adjustment Mechanism (CBAM).

UK ETS allowance prices averaged £47 per tonne of CO2e in 2025, up from £35 in 2023. The UK government's consultation on aligning the UK ETS cap with net zero suggests prices could reach £100-£150 per tonne by 2030. At these levels, marginal emissions reductions become extremely valuable, creating a business case for AI systems that can extract 3-8% additional efficiency from processes already operating near design specifications.

Carbon Clean, a UK-headquartered company, has integrated AI-driven process optimisation with its modular carbon capture systems, reducing the energy penalty of carbon capture by approximately 15% compared to conventional controls. Their CycloneCC technology deployed at Heidelberg Materials' Padeswood cement plant in North Wales uses real-time AI adjustment of solvent flow rates and regeneration temperatures to minimise parasitic energy consumption while maintaining capture rates above 90%.

Peak AI, based in Manchester, provides decision intelligence for industrial manufacturers, with documented deployments at major UK chemical producers achieving 5-9% energy intensity reductions through AI-optimised process scheduling, raw material blending, and waste heat recovery coordination. Their platform processes over 3 billion data points daily across client operations.

Arborea, while primarily a biotech company, exemplifies the convergence of AI and emissions monitoring by using machine learning to optimise its BioSolar Leaf technology for CO2 absorption, deployed at Imperial College London and several UK commercial sites.

The momentum driver is regulatory clarity. The Environment Agency's updated permitting guidance now requires Best Available Techniques (BAT) assessments to consider AI-enabled monitoring and optimisation where technically feasible, effectively mandating consideration of AI solutions for new and substantially modified installations.

Subsegment 3: AI-Optimised Heat Networks and Building Decarbonisation

The UK's Heat Network Zoning Pilot, launched in 2025, and the forthcoming Heat Network Regulation under the Energy Act 2023 have created a policy framework that strongly incentivises AI deployment in district heating and building energy management. This subsegment is growing at approximately 35-45% annually in terms of deployment volume.

Utilita Energy and IRT Surveys have partnered to deploy AI-driven thermal imaging analysis across social housing portfolios, identifying retrofit priorities with 20-30% greater accuracy than traditional assessment methods. The AI system analyses drone-captured thermal imagery to classify heat loss severity, detect insulation gaps, and predict energy performance certificate (EPC) improvements from specific interventions.

Guru Systems (now part of SSEN) provides AI-powered heat network monitoring that detects faults, leaks, and suboptimal operating conditions in real time. Deployed across 47 heat networks serving approximately 40,000 UK homes, their platform has demonstrated 8-14% improvements in network efficiency through automated flow temperature optimisation and predictive fault detection.

Passiv UK uses machine learning for predictive building energy management, focusing on social housing and public sector buildings. Their deployment across 12 local authority portfolios has documented average heating energy reductions of 16%, equivalent to £180-£250 per dwelling annually, while maintaining or improving thermal comfort as measured by occupant surveys.

The structural accelerant is the UK government's target of connecting 600,000 homes to heat networks by 2028 and the legally mandated efficiency standards that accompany regulation. AI optimisation is not merely cost-effective for heat networks but increasingly mandatory for achieving the operating standards required by Ofgem's new regulatory framework.

Subsegment 4: Scope 3 Supply Chain Emissions Intelligence

While not traditionally classified under "energy optimisation," the convergence of supply chain emissions tracking with AI-driven optimisation has created a rapidly growing hybrid subsegment. UK companies subject to Streamlined Energy and Carbon Reporting (SECR) requirements and anticipating ISSB-aligned disclosure mandates are driving demand for AI systems that can both measure and actively reduce supply chain emissions.

Emitwise, founded in London, provides AI-powered Scope 3 emissions calculation that replaces spend-based estimates with activity-level data enriched through machine learning. Their platform uses natural language processing to extract emissions-relevant data from invoices, bills of lading, and supplier reports, achieving accuracy improvements of 60-80% compared to purely spend-based methods. As of 2025, Emitwise processes data for over 200 enterprise clients, including several FTSE 100 companies.

Normative, while Swedish-headquartered, has significant UK operations and provides science-based emissions accounting that uses AI to match procurement data against lifecycle emissions databases. Their engine was adopted by the UN-backed SME Climate Hub as the recommended calculation tool for small and medium enterprises.

Watershed, expanding aggressively into the UK market, combines granular emissions measurement with AI-driven scenario modelling that identifies the most cost-effective decarbonisation interventions across complex supply chains. Their platform enables procurement teams to simulate the emissions impact of supplier switching, logistics reconfiguration, and materials substitution.

This subsegment's growth is driven by regulatory convergence. The UK's proposed Sustainability Disclosure Standards, expected to mandate ISSB-aligned reporting for large companies from 2027, will require Scope 3 disclosures that only AI-enabled platforms can provide at reasonable cost for organisations with hundreds or thousands of suppliers.

Subsegment 5: EV Fleet Charging Optimisation

The UK's Zero Emission Vehicle (ZEV) mandate, requiring 80% of new car sales and 70% of new van sales to be zero emission by 2030, is driving explosive growth in AI-optimised fleet charging. The subsegment has grown approximately 55% year-on-year as commercial fleets transition from pilot programmes to full-scale electrification.

Driivz (acquired by Vontier) provides AI-powered smart charging management for fleet operators, optimising charging schedules against energy tariffs, grid constraints, and vehicle routing requirements. Their deployment with Royal Mail's electric delivery fleet demonstrated 28% reduction in charging costs through load shifting and tariff arbitrage, while ensuring all vehicles achieved sufficient charge for daily routes.

Evergreen Smart Power operates vehicle-to-grid (V2G) programmes in the UK, using AI to determine optimal charge and discharge cycles for fleet vehicles connected during idle periods. Their partnership with Octopus Energy and Nissan has demonstrated that AI-optimised V2G can generate £500-£800 per vehicle annually in grid services revenue while reducing fleet charging costs by 15-20%.

Connected Kerb, deploying on-street charging infrastructure across UK local authorities, uses AI algorithms to predict usage patterns and optimise power allocation across charger networks, reducing required grid connection capacity by 30-40% through intelligent load management.

The structural momentum comes from cost imperatives. UK fleet operators face an average electricity cost of 25-35p per kWh for workplace charging, and AI optimisation that reduces costs by even 5-8p per kWh translates to £800-£1,200 in annual savings per vehicle for typical duty cycles. Across fleets of hundreds or thousands of vehicles, these savings justify substantial technology investment.

Cross-Cutting Observations

Several patterns emerge across these five subsegments. First, UK regulatory frameworks are acting as primary accelerants. The combination of UK ETS pricing, heat network regulation, ZEV mandates, and disclosure requirements creates a uniquely dense incentive landscape that favours AI deployment.

Second, the fastest-moving subsegments share a common characteristic: they operate at the intersection of data abundance and decision frequency. Grid-edge flexibility requires decisions every 30 minutes. Industrial process optimisation involves continuous adjustment. Fleet charging must be optimised daily. These high-frequency decision environments are precisely where AI delivers the greatest marginal value over human operators or simple rule-based systems.

Third, value capture is shifting from technology providers to integrated platform operators. Companies that combine AI algorithms with market access (Octopus Energy), hardware integration (Guru Systems), or regulatory compliance (Emitwise) capture significantly more value than those offering standalone AI analytics.

What to Watch Next

Three emerging subsegments merit monitoring for potential acceleration in 2026-2027. Hydrogen production optimisation, where AI manages electrolyser operation in response to variable renewable generation, is approaching commercial viability with projects at Gigastack (ITM Power and Orsted) and BP's H2Teesside. Aviation decarbonisation AI, applying machine learning to flight path optimisation, fuel blending, and ground operations, is gaining traction following ICAO's adoption of a long-term aspirational goal. Agricultural emissions monitoring, using satellite imagery and AI to estimate and verify farm-level emissions for the new Environmental Land Management schemes, represents a significant but early-stage opportunity.

Sources

• BloombergNEF. (2025). AI in Energy Markets: Global Spending, Deployment, and Revenue Analysis 2025. London: Bloomberg LP.
• National Grid ESO. (2025). Future Energy Scenarios 2025: The Role of AI and Digitalisation. Warwick: National Grid ESO.
• Climate Change Committee. (2024). Progress in Reducing Emissions: 2024 Report to Parliament. London: CCC.
• Ofgem. (2025). Heat Network Market Framework: Regulatory Standards and AI-Enabled Compliance. London: Ofgem.
• UK Department for Energy Security and Net Zero. (2025). Smart Systems and Flexibility Plan: 2025 Progress Update. London: DESNZ.
• International Energy Agency. (2025). Digitalisation and Energy: AI Applications in Energy Optimisation. Paris: IEA Publications.
• Modo Energy. (2025). GB Battery Storage Revenue Benchmarks: Q4 2025 Report. London: Modo Energy.