Summary

Artificial intelligence is rapidly becoming a core technology across the battery value chain. It is changing how batteries are discovered, designed, tested, manufactured, managed, reused and recycled. The market for AI-driven battery technology spans AI software, AI-enabled services and AI-specific hardware, such as vision inspection systems and edge-AI battery management processors. It is set for strong growth over the next decade, as battery demand expands and AI becomes embedded in every stage of the lifecycle. The past twelve months have marked a turning point. AI has moved from pilot projects to company-wide transformation programmes at the world's leading cell makers, with targets for large productivity gains and AI governance overseen at board level. Recent developments include:

• Generative AI in R&D. Materials foundation models and generative design tools have entered commercial battery R&D. Battery developers now offer AI-driven materials discovery platforms that screen millions of candidate electrolytes and electrode materials.
• Sodium-ion. Sodium-ion batteries have entered mass production for passenger vehicles. Solid-state batteries are approaching commercial launch, and both chemistries rely heavily on AI-accelerated development.
• Regulation and trade. The EU battery passport, due from February 2027, is turning battery state-of-health data into a regulated asset. Export controls on advanced battery technology have made data, models and process know-how matters of national strategy.
• New demand. AI data centres, humanoid robots, drones and electric aircraft have emerged as new sources of battery demand. Each needs sophisticated battery intelligence.

Looking ahead, the fastest growth is expected in:

• Second-life assessment, as the first large wave of retired EV batteries arrives after 2030;
• Recycling, under circular-economy regulation;
• Digital twins, linking data across the battery lifecycle;
• Materials informatics, driven by next-generation chemistries.

Stationary storage and data centres are becoming the fastest-growing customer groups. On-edge AI will take a rapidly rising share of battery analytics as neural processors spread through battery management hardware. Adoption will follow three waves:

1. Analytics and BMS intelligence in the late 2020s.
2. AI-native manufacturing and accelerated materials discovery in the early 2030s.
3. Autonomous laboratories and fully digital-twin-based development by the mid-2030s.

Competitive advantage will go to companies that combine proprietary battery data, deep electrochemical expertise, validated physics-informed models and integration across the full battery lifecycle. Battery development is moving from an experiment-centred process to a data-centred one, and AI capability is becoming as strategically important as manufacturing scale.

The Global Market for Artificial Intelligence (AI) Battery Technology 2027–2037 is a comprehensive analysis of how AI is transforming every stage of the battery lifecycle, and of the commercial opportunities this creates. The report analyses nine AI use-case arenas:

• materials informatics;
• cell and pack design;
• cell testing and modelling;
• manufacturing and quality control;
• BMS and battery analytics;
• application operation;
• digital twins;
• second-life assessment;
• recycling.

Report contents include:

• Global market forecasts 2027–2037 by use case, AI technique, deployment, offering, chemistry, end use and region
• Bear, base and bull scenario analysis and value-pool distribution across the technology stack
• Analysis of machine learning, deep learning, physics-informed models, generative AI, foundation models and agentic AI for batteries
• Battery data infrastructure, public datasets, advanced sensing, BMS silicon and edge-AI processors
• AI-driven materials discovery, universal interatomic potentials and autonomous laboratories
• AI in cell and pack design, testing, lifetime prediction and gigafactory manufacturing
• Next-generation BMS, cloud analytics, on-edge AI and battery health certification
• AI for EV charging, V2G, BESS dispatch and data-centre battery management
• Battery digital twins across the lifecycle
• Second-life assessment, regrouping and AI in recycling
• AI opportunities by battery chemistry and end-use market
• Regional policy ecosystems, battery passports, AI regulation, safety standards and export controls
• AI adoption by cell makers, EV OEMs and platform companies
• Competitive landscape, partnerships, funding, M&A and patent analysis
• Profiles of 95 companies. Companies profiled include About:Energy, ACC (Automotive Cells Company), ACCURE Battery Intelligence, Addionics, Aegis Critical Energy Defence, Aionics, Altilium, Anabatic Semi, Analog Devices, Atinary Technologies, B2U Storage Solutions, BASF, BattGenie, Blue Solutions, BMW Group, Bosch, Breathe Battery Technologies, Brightfield AI/Voltus, Brill Power, CATL, Chemix, Circulor, Circunomics, Citrine Informatics, Cling Systems, Cognex, Connected Energy, CuspAI, Dassault Systèmes, DellCon, DNV (Veracity), DP Technology, Dragonfly Energy, Dukosi, Dunia Innovations, Eatron Technologies, EcoPro, Electra Vehicles, Elisa IndustrIQ, Elysia Battery Intelligence (Fortescue Zero), EnPower Greentech, enspired, Envision AESC, Eonix Energy, EthonAI, EVE Energy, Fluence Energy, Gaussion, GBatteries, Glimpse Engineering, Google DeepMind, Huawei Digital Power, Intellegens and more....

ページTOPに戻る

Table of Contents

1 EXECUTIVE SUMMARY 37
1.1 Scope of This Report 37
1.2 AI Growth Drivers in the Battery Industry 37
1.2.1 Net-zero targets and the electrification push 37
1.2.2 Energy density, cost and critical-material constraints 37
1.2.3 Regulatory pull: the EU Battery Passport 38
1.2.4 Regional demand patterns: Europe, North America and East Asia 38
1.3 Challenges Facing the Rechargeable Battery Industry 38
1.4 How AI Applies Across the Battery Lifecycle 39
1.5 Five Benefits of AI: Development Speed, Cost, Performance, Sustainability and Market Responsiveness 40
1.6 Paradigm Shifts: From Experiment-Centred to Data-Centred Development 41
1.7 AI Disruption of the Battery Supply Chain 42
1.8 Use-Case Benchmarking and Maturity 43
1.9 AI in Batteries for Electric Vehicles 46
1.10 AI in Batteries for Stationary Storage and Data Centres 46
1.11 Interest by Region 47
1.12 Market Findings 48
1.12.1 Battery diagnostics by capacity served and market value 49
1.12.2 On-edge AI: diagnostics and performance enhancement 49
1.12.3 Materials informatics, cell testing and second-life assessment 50
1.13 Three-Wave Adoption Framework 50
1.13.1 Wave 1: Analytics and BMS intelligence (2027–2030) 50
1.13.2 Wave 2: AI-native manufacturing and accelerated materials (2030–2034) 50
1.13.3 Wave 3: Autonomous discovery and fully digital-twin-based development (2034–2037) 50
1.14 The Investment Surge 51
1.15 The Past Twelve Months 52
1.15.1 Generative AI and foundation models enter battery R&D 52
1.15.2 Leading cell makers launch company-wide AI transformation 52
1.15.3 Battery passports approach enforcement amid calls for phasing 53
1.15.4 Export controls reshape access to advanced battery technology 53
1.15.5 Stationary storage and AI data centres become the growth engine 54
1.15.6 Robotics, drones and aviation emerge as new battery markets 54
1.15.7 EV demand slowdown sharpens focus on cost and productivity 54
1.16 Strategic Imperatives for Cell Makers, OEMs, Utilities and Software Vendors 55
1.17 Major Market Players 55

2 INTRODUCTION: AI AND THE BATTERY VALUE CHAIN 57
2.1 Defining AI-Driven Battery Technology 57
2.2 Machine Learning vs. Artificial Intelligence 58
2.3 The Battery Market Context 58
2.3.1 Global Li-ion demand by application 58
2.3.2 Li-ion vs. beyond-Li-ion demand 59
2.3.3 Chemistry mix: NMC, LFP/LMFP, sodium-ion, solid-state, silicon anode 60
2.3.4 Cost trajectory and pack-price outlook 62
2.4 Traditional Battery Development: Sequential, Experiment-Centred and Slow 63
2.5 What Has Changed: Why AI Is Deployable Now 64
2.5.1 Connected-fleet data volumes 64
2.5.2 Cloud and edge compute cost curves 64
2.5.3 Open battery datasets and benchmarks 64
2.5.4 Foundation models for chemistry and materials 64
2.6 The Economic Case for AI in Batteries 65
2.6.1 R&D cost and time savings 65
2.6.2 Manufacturing yield and defect reduction 65
2.6.3 Lifetime extension and warranty cost reduction 66
2.7 Market Architecture: Nine AI Use-Case Arenas 67

3 MACHINE LEARNING FUNDAMENTALS FOR BATTERY APPLICATIONS 70
3.1 AI as a Moving Target: Definitions and Scope 70
3.2 The Importance of Data: Quality, Dimensionality and Standardised Structures 70
3.3 Learning Paradigms 71
3.3.1 Supervised learning 71
3.3.2 Unsupervised learning 71
3.3.3 Problem classes: regression, classification, clustering and anomaly detection 72
3.3.4 Reinforcement learning and the exploration–exploitation trade-off 72
3.3.5 Semi-supervised and active learning 72
3.4 Classical Algorithms 73
3.4.1 Support vector and relevance vector machines 73
3.4.2 Decision trees, random forests and gradient boosting 73
3.4.3 k-nearest neighbour and k-means clustering 74
3.4.4 Principal component analysis 74
3.4.5 Gaussian process regression 74
3.5 Neural Networks and Deep Learning 74
3.5.1 The artificial neuron and the training process 74
3.5.2 Feedforward, recurrent (LSTM, GRU) and convolutional networks 75
3.5.3 Generative adversarial networks and transformers 75
3.5.4 Graph neural networks for materials 75
3.5.5 Universal machine-learned interatomic potentials 77
3.6 Natural Language Processing and Large Language Models 78
3.7 Physics-Informed and Hybrid Models 79
3.7.1 Electrochemical (P2D, SPM) and equivalent circuit models 79
3.7.2 Kalman and particle filtering 79
3.7.3 Physics-informed neural networks 80
3.8 Generative AI, Foundation Models and Agentic AI 81
3.9 Transfer, Federated and Domain-Adaptive Learning 82
3.10 Explainability, Uncertainty Quantification, Under-Fitting and Over-Fitting 82
3.10.1 Explainability 82
3.10.2 Uncertainty quantification 82
3.10.3 Fitting errors 82
3.11 The Inefficiency of Overuse: When Not to Apply Machine Learning 83
3.12 Edge vs. Cloud Deployment 83

4 BATTERY DATA INFRASTRUCTURE, SENSING AND COMPUTE 85
4.1 The Battery Data Problem 85
4.1.1 Data scarcity, cost of cycling data and proprietary silos 85
4.1.1.1 Laboratory data costs 85
4.1.1.2 Proprietary silos 85
4.1.2 Open datasets and benchmarks 85
4.2 The Data Pipeline: From BMS to AI 87
4.3 Advanced Sensing Hardware 88
4.3.1 Voltage, current and temperature sensing 88
4.3.2 On-board electrochemical impedance spectroscopy (EIS) 89
4.3.3 Embedded fibre-optic, pressure and gas sensors 89
4.3.4 Wireless BMS and chip-on-cell sensing 89
4.3.4.1 Wireless BMS 89
4.3.4.2 Chip-on-cell sensing 90
4.4 BMS Silicon and Edge-AI Processors 91
4.4.1 Battery-monitoring ICs and analogue front ends 91
4.4.1.1 Measurement accuracy 91
4.4.1.2 Synchronisation 91
4.4.1.3 Sampling rates and data resolution 91
4.4.1.4 Cell balancing 91
4.4.1.5 Communication and isolation 92
4.4.1.6 Functional safety and diagnostics 92
4.4.1.7 Integrated sensing beyond voltage and temperature 92
4.4.1.8 New application segments 92
4.4.1.9 Market structure 93
4.4.2 Microcontrollers and neural processors for on-edge AI 93
4.4.3 Battery-management ICs for AI-server backup units (BBUs) 93
4.5 Cloud Platforms and Connectivity 94
4.5.1 Telematics, OTA updates and vehicle data access 94
4.5.2 Battery-in-the-cloud architectures 94
4.6 Compute for Materials Discovery 95
4.6.1 HPC and GPU clusters for DFT and molecular dynamics 95
4.6.2 Machine-learned interatomic potentials 95
4.6.3 Quantum computing for battery chemistry 95
4.7 Data Standards and Interoperability 95
4.7.1 Battery passport data models 95
4.7.2 OPC UA and IT/OT convergence in gigafactories 96
4.8 Cybersecurity and Data Governance 96

5 AI-DRIVEN BATTERY MATERIALS DISCOVERY (MATERIALS INFORMATICS) 98
5.1 Why AI for Battery Materials 98
5.1.1 The attraction of AI: navigating a vast design space 98
5.1.2 Traditional material discovery and the limits of DFT 98
5.1.3 Materials informatics 99
5.1.4 The workflow of AI-based research and development 99
5.2 Property Prediction and Material Grouping 101
5.2.1 Datasets and descriptors 101
5.2.1.1 Unsupervised grouping 102
5.2.2 DFT plus AI interpolation: lithium diffusion, voltage window and stability 102
5.3 Inverting the Process: Inverse Design 102
5.3.1 Informed selection vs. novel material formulation 102
5.3.2 Virtual screening 103
5.3.3 De novo and generative design 104
5.3.3.1 Large-scale generative results 104
5.3.4 Large language model interfaces and literature mining 104
5.4 Reducing Material Screening Time 105
5.4.1 The case for AI in material screening 105
5.4.2 Case studies of AI-based screening-time reduction 105
5.5 Cathode Materials 106
5.5.1 High-nickel and cobalt-free layered oxides 106
5.5.2 LFP and LMFP optimisation 106
5.5.3 Disordered rock-salt and lithium-rich cathodes 106
5.6 Anode Materials 107
5.6.1 Graphite and silicon-carbon composites 107
5.6.2 Lithium-metal anodes and dendrite suppression 107
5.6.3 Hard carbon for sodium-ion 107
5.7 Liquid Electrolytes 108
5.7.1 Formulation and additive design 108
5.7.2 Solvent stability prediction 108
5.8 Solid Electrolytes 109
5.8.1 Sulfide, oxide, polymer and halide systems 109
5.8.2 Case study: Microsoft and PNNL AI-screened solid electrolytes 110
5.9 AI-Based Optimisation of Cell Material Combinations 111
5.10 Autonomous (Self-Driving) Laboratories 111
5.10.1 Robotic synthesis and characterisation 111
5.10.2 Closed-loop active learning 112
5.10.3 Agentic AI lab assistants 113
5.10.4 European initiatives: BIG-MAP and Battery 2030+ 114
5.10.5 Commercial self-driving laboratories 115
5.10.6 Academic milestones and mobile robots 115
5.10.7 Big-technology autonomous laboratories 115
5.11 Critical-Mineral Substitution and Supply-Risk Mitigation 116
5.12 Players in Materials Informatics for Batteries 116
5.13 Business Analysis 118
5.13.1 Business models and partnerships 118
5.13.2 Existing client–supplier relationships 118
5.13.3 Differentiation 118
5.13.4 Challenges: data quality, interpretability, compute and qualification 119
5.13.4.1 Data quality and availability 119
5.13.4.2 Gap between computation and experiment 119
5.13.4.3 Interpretability and trust 119
5.13.4.4 Computational resources 120
5.13.4.5 Model generalisation 120
5.13.4.6 Qualification and time to market 120
5.13.4.7 Intellectual property and data ownership 120
5.13.4.8 Regulatory and safety requirements 121
5.13.4.9 Skills 121
5.13.5 Pricing of materials informatics platforms 121
5.13.6 Risks for SaaS business models 122
5.13.7 Barriers to profitability 122
5.13.8 Big-technology platforms and competition for dedicated players 123
5.13.9 In-house development and consolidation 123
5.13.10 Outlook 123
5.14 Market Forecast: Materials Informatics for Batteries 2027–2037 123

6 AI APPLICATIONS BY BATTERY CHEMISTRY 125
6.1 Advanced Li-Ion: High-Nickel, LFP/LMFP and the Path to 350 Wh/kg 125
6.2 Silicon and Silicon-Carbon Anodes 127
6.3 Lithium-Metal and Anode-Less Cells 127
6.4 Solid-State and Semi-Solid-State Batteries 128
6.4.1 Commercial status 128
6.4.2 Demand outlook 128
6.5 Lithium-Sulfur Batteries 130
6.6 Sodium-Ion Batteries 130
6.6.1 Commercial status 130
6.6.2 Demand outlook 131
6.7 LTO and Niobate Batteries 132
6.8 Zinc-Based and Aluminium-Ion Batteries 132
6.9 Redox Flow Batteries 132
6.10 Structural, Flexible, Printed and Other Emerging Formats 133
6.11 PFAS-Free Binders and Additives: Regulation-Driven Reformulation 133
6.12 AI Opportunity by Chemistry 134

7 AI IN CELL AND PACK DESIGN 136
7.1 From Trial-and-Error to Simulation-Led Design 136
7.2 Multi-Variable Cell Design Optimisation 136
7.2.1 Electrode thickness, particle size distribution, binder content and electrolyte ratio 136
7.2.2 Genetic algorithms combined with reinforcement learning 137
7.2.3 Coupling P2D physics models with AI to test design scenarios virtually 137
7.3 Electrode and Microstructure Design 138
7.3.1 Image-based microstructure reconstruction 138
7.3.2 Porosity, tortuosity and loading optimisation 138
7.3.3 Thick-format electrodes 138
7.4 Cell Formats and Architectures 139
7.4.1 Larger cell formats and tabless designs (Tesla 4680) 139
7.4.2 Bipolar and dual-electrolyte architectures 139
7.4.2.1 How bipolar cells work 139
7.4.2.2 Benefits 139
7.4.2.3 Challenges 140
7.4.2.4 Commercial status 140
7.4.2.5 How AI helps with bipolar designs 140
7.4.2.6 Dual-electrolyte architectures 140
7.4.2.7 Variants 141
7.4.2.8 Dual-electrolyte challenges 141
7.4.2.9 How AI helps with dual-electrolyte designs 141
7.4.3 Commercial examples 141
7.4.3.1 Large-format tabless cylindrical cells 141
7.4.3.2 Long prismatic "blade" cells 142
7.4.3.3 High-integration cell-to-pack systems 142
7.4.3.4 Multi-layer electrodes 142
7.4.3.5 Three-dimensional current collectors 142
7.4.3.6 Thick electrodes and simplified cell designs 142
7.4.3.7 Dry-electrode processing 142
7.4.3.8 Bipolar designs 142
7.4.3.9 The common thread 143
7.5 Cell Performance and Energy Density 143
7.6 AI-Based Battery Pack Design Structure Optimisation 144
7.6.1 Why pack structure optimisation matters 144
7.6.2 Key considerations in optimal pack design 145
7.6.3 AI-based optimisation workflow 145
7.7 Research on AI-based pack design optimisation 146
7.8 Cell-to-Pack, Cell-to-Chassis and Hybrid Packs 148
7.8.1 Cell-to-pack and cell-to-chassis 148
7.8.2 Hybrid battery packs 148
7.9 Generative Design, Lightweighting and Thermal System Design 149
7.10 Design for Recycling and Sustainability 149
7.11 Battery Design and Simulation Software Landscape 150
7.12 Market Forecast: AI Cell and Pack Design 2027–2037 151

8 AI IN CELL TESTING, MODELLING AND LIFETIME PREDICTION 152
8.1 Traditional Cell Testing: Shortcomings and Challenges 152
8.2 AI for High-Throughput Automated Testing 153
8.3 Data Forms for Cell Modelling 153
8.4 AI for Design of Experiments and Anomalous Data Identification 154
8.5 Lifetime Modelling and Early-Life Prediction 154
8.6 Degradation Modelling 155
8.6.1 SEI growth, lithium plating and active-material loss 155
8.6.2 Incremental capacity, differential voltage and EIS health indicators 156
8.7 Remaining Useful Life (RUL) Prediction Methods 157
8.7.1 Model-based approaches 157
8.7.2 Stochastic process methods: Wiener, Gamma and Markov 157
8.7.3 Classical machine learning approaches 158
8.7.4 Deep learning approaches 158
8.7.5 Hybrid fusion approaches 158
8.8 Temperature and Pressure Simulation 160
8.9 Data-Driven Cell Architecture Optimisation 161
8.10 AI Classification of Cell Cross-Section Defect Types 161
8.11 Algorithmic Approaches by Testing Mode 162
8.12 Test Laboratory Automation and Certification Implications 163
8.13 Players in AI for Cell Testing 164
8.14 Business Analysis 165
8.14.1 Typical business models 165
8.14.2 Differentiation 165
8.14.3 Challenges 165
8.14.3.1 Trust and acceptance 165
8.14.3.2 Transferability across chemistries and protocols 166
8.14.3.3 Capturing rare and late-emerging failure modes 166
8.14.3.4 Data access and confidentiality 166
8.14.3.5 Data quality and consistency 166
8.14.3.6 Integration with laboratory infrastructure 166
8.14.3.7 Regulatory and certification barriers 167
8.14.3.8 Demonstrating commercial value 167
8.14.4 Outlook 167
8.14.4.1 Near term (2027–2030) 167
8.14.4.2 Medium term (2030–2034) 167
8.14.4.3 Longer term (2034–2037) 167
8.14.4.4 Structural change 167
8.15 Market Forecast: AI Cell Testing 2027–2037 168

9 AI IN CELL ASSEMBLY AND MANUFACTURING 169
9.1 Overview of the Traditional Manufacturing Process 169
9.2 The Gigafactory Ramp-Up Problem 170
9.2.1 Scrap rates during ramp-up 170
9.2.2 The "valley of death" between start of production and target yield 170
9.3 Data Quality and Data Acquisition Challenges in Industrial Settings 171
9.4 AI Use Cases by Process Step 172
9.5 Vision AI for Defect Detection and Quality Control 173
9.5.1 Coating non-uniformity, electrode surface defects and stacking misalignment 173
9.5.2 X-ray and CT inspection 173
9.5.3 Ultrasound and acoustic inspection 174
9.6 Streaming Process Analytics and Preventive Control 175
9.6.1 Electrolyte injection volume, calendering pressure and drying conditions 175
9.6.2 Predicting defects before they occur 175
9.7 Electrode Coating and Drying Simulation 176
9.8 Formation and Ageing: AI-Shortened Protocols 176
9.9 Algorithmic Approaches in Manufacturing and Cell Assembly 177
9.10 Factory Digital Twins, Virtual Commissioning and FAT/SAT 177
9.11 Predictive Maintenance, MES and IT/OT Integration 178
9.11.1 Predictive maintenance 178
9.11.2 Manufacturing execution systems 178
9.12 Robotics and Physical AI in Cell and Pack Assembly 178
9.13 Smart Battery Manufacturing Players 179
9.14 Business Analysis 180
9.14.1 Types of smart manufacturing players 180
9.14.1.1 Vertically integrated cell makers 180
9.14.1.2 Industrial software and automation providers 180
9.14.1.3 Equipment makers 181
9.14.1.4 Inspection-technology providers 181
9.14.1.5 Specialist AI analytics companies 181
9.14.1.6 Battery data platforms 181
9.14.1.7 System integrators and engineering firms 181
9.14.1.8 Research institutions 181
9.14.2 Challenges 182
9.14.2.1 Upfront investment and uncertain payback 182
9.14.2.2 Integration with heterogeneous equipment 182
9.14.2.3 Data quality and genealogy 182
9.14.2.4 Model maintenance and drift 182
9.14.2.5 Skills shortages 183
9.14.2.6 Organisational and cultural barriers 183
9.14.2.7 Cybersecurity 183
9.14.2.8 Data ownership and vendor lock-in 183
9.14.2.9 Scaling beyond pilots 183
9.14.3 Outlook 183
9.14.3.1 Large cell makers 183
9.14.3.2 New entrants and start-ups 184
9.14.3.3 How solutions will adapt 184
9.14.3.4 Policy support 184
9.14.3.5 Next-generation chemistries 184
9.14.3.6 Overall outlook 185
9.15 Market Forecast: AI in Battery Manufacturing 2027–2037 185

10 NEXT-GENERATION BMS AND BATTERY ANALYTICS 187
10.1 Battery Management Systems: Purpose and Multi-Cell Pack Management 187
10.2 Limitations of Existing BMS and the Case for Next-Generation BMS 187
10.3 Next-Generation BMS Integrated with Big Data and AI 188
10.4 Data Pre-Processing for AI Model Development 190
10.4.1 Pre-processing procedures 190
10.4.2 Why health indicators must reflect operating environment and data-collection conditions 190
10.4.3 Methods for extracting health indicators 190
10.4.4 Limitations and worked examples 191
10.5 Data Structures and Forms for Diagnostics 192
10.6 State Estimation: SoC, SoH, SoP, SoE and Internal Temperature 193
10.7 Fault Detection and Anomaly Diagnosis 193
10.7.1 Selecting deep learning models for battery time-series prediction 193
10.7.2 Selecting deep learning models for anomaly detection and fault diagnosis 194
10.7.3 Case study: data patterning and anomaly diagnosis for safe operation 194
10.8 Prescriptive AI: From Prediction to Recommended Action 195
10.9 AI-Optimised Charging: Adaptive and Degradation-Aware Protocols 196
10.10 Thermal Runaway Early Detection and Advanced Pack Sensors 196
10.10.1 Advanced sensor technologies 197
10.10.2 Sensor market outlook 197
10.10.3 Integration and the path to predictive maintenance 198
10.11 Wireless BMS and Remote Monitoring 198
10.12 On-Edge AI 199
10.12.1 Case study: real-time lifetime prediction on embedded Linux 199
10.12.2 Samsung Galaxy S25 battery AI 199
10.12.3 Eatron and Syntiant 199
10.12.4 LG Energy Solution and Qualcomm 199
10.12.5 Tesla BMS: optimisation over a journey 199
10.13 Cloud Battery Analytics: Fleet, Warranty, Residual Value and Insurance 200
10.14 The Battery Passport 200
10.15 Players in AI for Battery Diagnostics and Management 201
10.16 Business Analysis 202
10.16.1 Business models 202
10.16.1.1 Software as a service (SaaS) 202
10.16.1.2 Embedded software licences 202
10.16.1.3 Certification and reporting services 203
10.16.1.4 Outcome-based models 203
10.16.1.5 Bundled hardware and software 204
10.16.1.6 Hybrid and platform models 204
10.16.2 Differentiation 205
10.16.2.1 Model accuracy and validation against real field outcomes 205
10.16.2.2 Scale of data 205
10.16.2.3 Physics-informed approaches 205
10.16.2.4 Integration 206
10.16.3 Challenges 207
10.16.3.1 Data access 207
10.16.3.2 Edge constraints 207
10.16.3.3 Certification 207
10.16.3.4 Price pressure 208
10.16.3.5 Proving value 208
10.16.4 Outlook 209
10.16.4.1 Expansion across mobility 209
10.16.4.2 Convergence of analytics and BMS 209
10.16.4.3 Expansion across the lifecycle 210
10.16.4.4 Global diffusion 210
10.16.4.5 Market evolution 210
10.17 Market Forecast: Battery Diagnostics and BMS Analytics 2027–2037 210

11 AI FOR BATTERY OPERATION IN EVs, ENERGY STORAGE AND DATA CENTRES 213
11.1 AI for EV Operation Within the Power Grid 213
11.1.1 Why grid-aware EV operation needs AI 213
11.1.2 Case studies 213
11.1.2.1 OEM joint platforms 214
11.1.2.2 Energy-retailer tariffs with automated charging 214
11.1.2.3 Utility managed-charging programmes 214
11.1.2.4 Vehicle-to-grid services 214
11.1.2.5 Large-scale pilots in China 214
11.2 AI-Based Charging for Optimal EV Battery Operation 216
11.2.1 Why AI-based charging is needed 216
11.2.2 Case studies 216
11.2.2.1 Journey-level charging optimisation 216
11.2.2.2 Ultra-high-power charging 216
11.2.2.3 Battery swapping at scale 217
11.2.2.4 Battery-as-a-service in emerging markets 217
11.2.2.5 Adaptive charging software 217
11.2.2.6 Depot charging for commercial fleets 217
11.2.2.7 Research on AI-optimised fast charging 218
11.3 Smart Charging and Vehicle-to-Grid Platforms 219
11.3.1 Smart charging and V1G 219
11.3.2 Bidirectional charging 219
11.3.2.1 The role of AI 219
11.3.3 Degradation cost 219
11.3.4 Scaling 219
11.4 AI-Based Environmental Control for Energy Storage Systems 220
11.4.1 Why environmental control must reflect battery ageing characteristics 220
11.4.2 Limitations of existing control strategies 220
11.4.3 EIS image-based classification of operating environments 220
11.4.4 Designing AI-based control strategies for optimal ESS operation 221
11.5 BESS Dispatch, Energy Trading and Augmentation Planning 222
11.5.1 The economics of storage 222
11.5.2 AI dispatch and trading platforms 222
11.5.3 Augmentation planning 222
11.5.4 Long-duration storage 222
11.6 Big Data Management in Cloud Servers 223
11.6.1 Why cloud data management matters as data volumes grow 223
11.6.2 Operational pattern images for large-scale EV data 223
11.6.3 Data compression for storage efficiency 224
11.6.4 Optimising lifetime prediction models in the cloud 224
11.7 Battery Storage for Data Centres, Commercial and Industrial Sites 225
11.7.1 AI data centres as a driver of battery demand 225
11.7.1.1 Demand outlook 225
11.7.2 Battery backup units for AI servers 226
11.8 Battery Swapping, Battery-as-a-Service and New Business Models 227

12 BATTERY DIGITAL TWINS 228
12.1 Digital Twin Concepts and Technologies 228
12.1.1 Concept and expected benefits 228
12.1.2 Components of a digital twin 228
12.1.3 Implementation and use 229
12.1.4 Key enabling technologies 230
12.1.5 Optimising the digital twin 230
12.2 Digital Twins in Battery Development 230
12.2.1 Applications in batteries 230
12.2.2 Hierarchical structure of the battery digital twin 231
12.2.3 Vision for the battery digital twin 231
12.2.4 Battery modelling with digital twins 232
12.2.5 Challenges in applying digital twins to batteries 232
12.2.5.1 Limited observability 232
12.2.5.2 Parameterisation 232
12.2.5.3 Model fidelity versus computational cost 233
12.2.5.4 Data continuity across the lifecycle 233
12.2.5.5 Validation 233
12.2.5.6 Model drift and maintenance 233
12.2.5.7 Integration and interoperability 233
12.2.5.8 Uncertainty quantification 234
12.2.5.9 Cost and business case 234
12.2.5.10 Data ownership and privacy 234
12.2.5.11 Security 234
12.2.5.12 Skills 234
12.3 Digital-Twin Functions 234
12.3.1 SoX estimation and cell balancing 234
12.3.2 Advanced fault diagnosis and RUL estimation 235
12.4 Extending Across the Lifecycle: Manufacturing, Thermal Management, Battery Passport and V2G 236
12.5 Battery Digital Twin Platforms 236
12.5.1 Building a battery digital twin platform 236
12.5.2 Comparison of integrated platforms 236
12.6 Digital Twins and Cloud BMS 237
12.6.1 Trends in digital-twin-based state estimation 237
12.6.2 The virtual battery model 238
12.6.3 Digital twin BMS case studies 238
12.7 Market Forecast: Battery Digital Twins 2027–2037 239

13 SECOND-LIFE ASSESSMENT AND BATTERY REUSE 241
13.1 Second-Life Batteries Overview 241
13.1.1 Determining the second-life stream 241
13.1.2 End-of-life volumes 241
13.2 Safety Concerns, Regulation and the Battery Passport 242
13.2.1 Safety risks 242
13.2.2 Regulation and standards 243
13.2.3 The battery passport 243
13.3 The Role of AI 243
13.4 AI-Based Rapid Diagnostics to Cut Reuse Costs 244
13.4.1 Why rapid diagnostics are needed 244
13.4.2 Key considerations when receiving used batteries 244
13.4.3 Rapid SOH diagnosis using one-hot encoding 245
13.4.4 Improving rapid SOH diagnosis with AdaBoost 245
13.4.5 Research on AI-based rapid diagnostic technologies 245
13.5 AI-Based Regrouping for Battery Reuse 246
13.5.1 Limitations of existing reuse processes 246
13.5.2 Removing defective cells using RLS-deviation fault diagnosis 246
13.5.3 RUL-based regrouping algorithms 246
13.6 Second-Life Applications and Economics 247
13.6.1 Economics 249
13.6.2 AI's contribution 249
13.7 Players in AI for Second-Life Assessment 249
13.8 Business Analysis 250
13.8.1 Revenue streams 250
13.8.2 Types of players 251
13.8.2.1 Diagnostics specialists 251
13.8.2.2 Marketplace operators 251
13.8.2.3 Second-life system integrators 251
13.8.2.4 OEMs and cell makers 251
13.8.2.5 Recyclers 251
13.8.2.6 Energy companies and utilities 252
13.8.2.7 Insurers, financiers and certifiers 252
13.8.3 Differentiation 252
13.8.3.1 Speed and cost of assessment 252
13.8.3.2 Accuracy and certified reliability 252
13.8.3.3 Breadth of coverage 252
13.8.3.4 Access to first-life data 252
13.8.3.5 Integration across the value chain 252
13.8.3.6 System design and operating capability 253
13.8.3.7 Warranty strength 253
13.8.3.8 Regulatory compliance and documentation 253
13.8.4 Challenges 253
13.8.4.1 Access to battery history 253
13.8.4.2 Heterogeneity of batteries 253
13.8.4.3 Economics under falling new-battery prices 253
13.8.4.4 Competition with recycling 254
13.8.4.5 Safety and liability 254
13.8.4.6 Regulation and certification 254
13.8.4.7 Logistics and scale 254
13.8.4.8 Warranty and performance risk 254
13.8.4.9 Access to finance 254
13.8.5 Outlook 254
13.8.5.1 Volume growth after 2030 255
13.8.5.2 Convergence of first-life and second-life analytics 255
13.8.5.3 Standardisation and certification 255
13.8.5.4 Evolving business models 255
13.9 Market Forecast: AI Second-Life Assessment 2027–2037 256

14 AI IN BATTERY RECYCLING 257
14.1 Limitations of Existing Recycling Processes and the Case for AI 257
14.2 Recycling Technologies 258
14.2.1 Hydrometallurgy 258
14.2.2 Pyrometallurgy 258
14.2.3 Direct recycling 258
14.2.4 Other methods 258
14.3 AI Across the Recycling Process 259
14.3.1 Sorting and chemistry identification 259
14.3.2 Robotic and vision-guided disassembly 260
14.3.3 Black-mass characterisation 260
14.3.4 Hydrometallurgical and direct-recycling process control 260
14.4 Recycling Specific Components and Beyond-Li-Ion Chemistries 262
14.4.1 Component recovery 262
14.4.2 Beyond-lithium-ion chemistries 262
14.5 Traceability, Battery Passports and EU Targets 262
14.6 Recycling Players Using AI 263
14.6.1 Chinese players 264
14.6.2 Technology providers 264
14.7 Market Forecast: AI in Battery Recycling 2027–2037 264

15 END-USE MARKETS 266
15.1 Electric Vehicles 266
15.1.1 Sales in 2025 266
15.1.2 Early 2026 and outlook 266
15.1.3 Battery demand 266
15.1.4 Passenger vehicles 267
15.1.4.1 Development 267
15.1.4.2 Manufacturing 267
15.1.4.3 Operation 267
15.1.4.4 Residual value and warranty 268
15.1.4.5 Adoption 268
15.1.5 Commercial vehicles and heavy trucks 268
15.1.5.1 Market growth 268
15.1.5.2 Distinctive battery requirements 268
15.1.5.3 Infrastructure 269
15.1.6 Two- and three-wheelers 269
15.1.6.1 Swapping and BaaS 269
15.1.6.2 AI requirements 269
15.1.6.3 Chemistry and cost 270
15.1.7 Regional dynamics 270
15.1.8 Used-EV market and battery health certification 270
15.1.8.1 The confidence gap 270
15.1.8.2 Variation within models 270
15.1.8.3 Certification in practice 270
15.1.8.4 Toward mandatory certificates 271
15.1.8.5 Assessment methods 271
15.2 Stationary Energy Storage 272
15.2.1 Growth in 2025 272
15.2.2 Outlook 272
15.2.3 Drivers 272
15.2.4 Utility-scale BESS 273
15.2.4.1 Why AI matters 273
15.2.4.2 Obstacles 274
15.2.5 Commercial, industrial and residential storage 274
15.2.5.1 New business models 274
15.2.6 Data-centre backup and on-site storage 274
15.2.7 Long-duration storage 275
15.2.8 Regional dynamics 275
15.2.8.1 China 275
15.2.8.2 Other regions 275
15.3 Consumer Electronics and Wearables 276
15.4 Off-Highway Machines: Construction, Agriculture and Mining 277
15.5 Robotics, Drones and Physical AI Platforms 277
15.5.1 Humanoid robots 277
15.5.1.1 Energy demand 277
15.5.1.2 Current battery packs 278
15.5.1.3 The weight constraint 278
15.5.2 Autonomous mobile robots and warehouse automation 279
15.5.3 Drones 279
15.5.4 AI applications 279
15.5.5 Outlook 280
15.6 Aviation and eVTOL 281
15.6.1 eVTOL aircraft 281
15.6.1.1 Commercial progress 281
15.6.1.2 Battery requirements 281
15.6.1.3 AI applications in eVTOL 282
15.6.1.4 Validation and explainability 282
15.6.2 Hybrid-electric aircraft 282
15.6.3 Battery technology roadmap for aviation 283
15.6.4 Outlook 283
15.7 Marine and Rail 284
15.7.1 Marine 284
15.7.2 Rail 284
15.8 Defence 284
15.9 Medical Devices 285
15.10 Comparative Application Analysis 285

16 GLOBAL MARKET FORECASTS 2027–2037 287
16.1 Scope of Forecasts and Methodology 287
16.2 Total Market 287
16.3 By Use Case 288
16.3.1 Leading segments 288
16.3.2 Fastest growth 289
16.4 Battery Diagnostics by Capacity Served 289
16.5 By AI Technique 290
16.6 By Deployment Mode: Cloud, On-Edge and On-Premises 291
16.7 By Offering: Software, Services and AI-Enabled Hardware 292
16.8 By Battery Chemistry 292
16.8.1 LFP and LMFP 292
16.8.2 High-nickel NMC/NCA 292
16.8.3 Next-generation chemistries 293
16.9 By Region 293
16.10 Scenario Analysis: Bear, Base and Bull Cases 294
16.11 Value Pool Distribution Across the Stack 296

17 REGIONAL MARKETS AND BATTERY POLICY ECOSYSTEMS 298
17.1 Regional Demand Patterns for Battery AI 298
17.1.1 Europe: analytics and second-life assessment led by the Battery Passport 298
17.1.2 North America: materials informatics and AI-assisted cell testing 298
17.1.3 East Asia: manufacturing and development applications 298
17.2 China 299
17.2.1 Institutional mapping and 15th Five-Year Plan priorities 299
17.2.2 Solid-state and sodium-ion innovation policy 300
17.2.3 Export controls on battery technology 300
17.2.4 AI adoption by Chinese cell makers 301
17.2.4.1 Scale of the ecosystem 301
17.2.4.2 Demand drivers 301
17.2.4.3 AI use 301
17.2.5 Materials Genome Engineering 301
17.3 South Korea 302
17.3.1 K-Battery strategy and next-generation R&D funding 302
17.3.1.1 Institutions 302
17.3.1.2 Strategy over time 302
17.3.1.3 Demand measures 302
17.3.2 Battery hubs: Pohang and Cheongju–Ochang 302
17.3.2.1 Current conditions 303
17.3.3 The "full-stack" AI, energy and battery strategy 303
17.3.4 AI-based development strategies of Korean cell makers 303
17.3.5 AI-driven science programmes 304
17.4 Japan 304
17.4.1 Battery Industry Strategy and Green Transformation targets 304
17.4.1.1 Progress and setbacks 305
17.4.1.2 Industry consolidation 305
17.4.2 NEDO programmes and the Advanced Battery Collaboration 305
17.4.3 NIMS and AIST data-driven materials platforms 305
17.5 European Union 305
17.5.1 Batteries Regulation, BATT4EU and Battery 2030+ 305
17.5.1.1 Research programmes 306
17.5.2 National strategies: Germany, France, Hungary, Poland 306
17.5.2.1 Europe's position at end-2025 306
17.5.3 Industrial Accelerator Act and Battery Booster Facility 307
17.5.3.1 Scale-up policy 307
17.5.3.2 Implications for AI 307
17.6 United States 307
17.6.1 DOE, ARPA-E and national laboratory programmes 307
17.6.2 Policy shift toward stationary storage and defence 308
17.6.2.1 Capacity build-out 308
17.6.2.2 Policy shift 308
17.6.2.3 Implications for AI 308
17.6.3 The Materials Genome Initiative 308
17.7 United Kingdom 308
17.7.1 AI strengths 308
17.7.2 Market drivers 309
17.8 Canada and India 309
17.9 Rest of World: Morocco, Australia and Southeast Asia 309
17.10 Global cell manufacturing capacity 310

18 REGULATION, STANDARDS AND TECHNOLOGY SOVEREIGNTY 312
18.1 Battery Safety Standards and AI Validation 312
18.2 Battery Passport and Data-Sharing Rules 313
18.2.1 Scope of the passport 313
18.2.2 What it means for AI 314
18.2.3 Wider data-access rules 314
18.3 AI Regulation: EU AI Act, US and China Frameworks 314
18.3.1 The EU AI Act 314
18.3.2 Implications for battery AI 315
18.3.3 United States 315
18.3.4 China 315
18.4 Functional Safety of AI-Based BMS (ISO 26262, ISO 21448, ISO/PAS 8800) 315
18.5 Standards Development: IEC, SAE, IEEE and National Programmes 316
18.6 PFAS Restrictions and Materials Regulation 317
18.7 Export Controls and Technology Sovereignty 317

19 AI ADOPTION BY BATTERY MAKERS, EV OEMs AND PLATFORM COMPANIES 320
19.1 Battery Cell and Materials Companies 320
19.2 EV OEMs 323
19.4 Platform Companies and Research Institutions 325

20 COMPETITIVE LANDSCAPE, INVESTMENT AND OUTLOOK 329
20.1 Competitive Dynamics by Use Case 329
20.2 Types of Players: Start-ups, Software Incumbents, Cell Makers and OEMs 330
20.3 Business Models and Revenue Streams 331
20.4 Strategic Partnerships and Joint Ventures 332
20.5 Investment and Funding 333
20.5.1 Strategic investors 333
20.6 M&A Landscape 334
20.7 Patent Landscape 335
20.7.1 Growth 335
20.7.2 By jurisdiction 335
20.8 Barriers to Adoption 336
20.8.1 Technical: data scarcity, generalisation and interpretability 336
20.8.2 Economic: ROI, payback periods and fragmented buyers 336
20.8.3 Organisational: skills shortages in data science and electrochemistry 337
20.9 Emerging Frontiers 2030–2037 338
20.9.1 Smart cells with embedded sensing and self-healing functions 338
20.9.2 Quantum–AI hybrid materials discovery 338
20.9.3 Agentic AI across the battery lifecycle 338

21 COMPANY PROFILES 339 (95 company profiles)

22 APPENDIX 434
22.1 Research Methodology 434
22.2 Forecast Assumptions 437
22.3 Glossary of Terms and Abbreviations 439

23 REFERENCES 441