1. | EXECUTIVE SUMMARY |
1.1. | The scope of this report |
1.2. | Who should read this report? |
1.3. | Research methodology |
1.4. | Clarifying terms: machine learning vs artificial intelligence |
1.5. | Inefficiencies of overuse |
1.6. | Under- and over-fitting |
1.7. | Challenges facing the rechargeable battery industry |
1.8. | How AI can be applied throughout the battery lifecycle |
1.9. | AI disruptions to the battery supply chain |
1.10. | Use-case benchmarking |
1.11. | Use-case maturity comparison |
1.12. | AI in batteries for EVs |
1.13. | AI in batteries for BESS |
1.14. | Interest by region |
1.15. | Scope of forecasts |
1.16. | Methodologies |
1.17. | Diagnostics by capacity served |
1.18. | Diagnostics by market value |
1.19. | On-edge AI: diagnostics |
1.20. | On-edge AI: performance enhancement |
1.21. | Cell testing by market value |
1.22. | Second-life assessment by market value |
1.23. | AI will see significant usage throughout the battery industry |
2. | MACHINE LEARNING APPROACHES: AN OVERVIEW |
2.1. | An introduction to AI - shifting goalposts |
2.2. | Machine learning as a subset of artificial intelligence |
2.3. | Machine learning approaches |
2.4. | The importance of data - quality and dimensionality |
2.5. | Standardizing data structures |
2.6. | Supervised learning |
2.7. | Unsupervised learning |
2.8. | Problem classes in supervised and unsupervised learning |
2.9. | Reinforcement learning |
2.10. | Semi-supervised and active learning |
2.11. | The ɛ parameter: exploitation vs. exploration |
2.12. | Neural networks - an introduction |
2.13. | An artificial neuron in the training process |
2.14. | Types of neural network |
2.15. | Support vector machines |
2.16. | Decision tree methods |
2.17. | k-nearest neighbor (kNN) |
2.18. | k-means clustering |
2.19. | Principal component analysis |
3. | MATERIAL DISCOVERY |
3.1. | Overview |
3.1.1. | Material discovery in batteries - the attraction of AI |
3.1.2. | Traditional material discovery and DFT |
3.1.3. | An introduction to Materials Informatics |
3.1.4. | Property prediction and material grouping |
3.1.5. | Datasets and descriptors |
3.1.6. | The golden grail - inverting the process |
3.1.7. | Informed selection vs. novel material formulation |
3.1.8. | Virtual screening |
3.1.9. | De novo design |
3.1.10. | Integration of LLM interface |
3.1.11. | Electrodes |
3.1.12. | Electrolytes |
3.1.13. | Problem and algorithm classes |
3.2. | Players in materials informatics for batteries |
3.2.1. | BIG-MAP |
3.2.2. | Microsoft Quantum - Azure Open AI |
3.2.3. | Umicore |
3.2.4. | Wildcat Discovery Technologies |
3.2.5. | Schrödinger - an overview |
3.2.6. | Schrödinger technical details |
3.2.7. | Eonix Energy |
3.2.8. | Citrine Informatics |
3.2.9. | Morrow Batteries |
3.2.10. | Chemix |
3.2.11. | Aionics |
3.2.12. | SES AI |
3.2.13. | SES AI batteries |
3.3. | Business analysis for AI in battery material discovery |
3.3.1. | Business models/partnerships |
3.3.2. | Existing client-supplier relationships |
3.3.3. | Differentiation |
3.3.4. | Challenges |
3.3.5. | Materials informatics will see increasing use in the battery industry over the next decade |
4. | CELL TESTING AND MODELLING |
4.1. | Overview |
4.1.1. | Traditional cell testing - shortcomings and challenges |
4.1.2. | AI for high-throughput automated testing |
4.1.3. | Data forms for cell modelling |
4.1.4. | AI for design of experiment (DoE) and anomalous data identification |
4.1.5. | AI for lifetime modelling |
4.1.6. | AI for degradation modelling |
4.1.7. | AI for temperature and pressure simulation |
4.1.8. | Data driven cell architecture optimization |
4.1.9. | Algorithmic approaches for different testing modes |
4.2. | Players in AI for cell testing |
4.2.1. | Stanford, MIT and Toyota Research Institute |
4.2.2. | StoreDot - a data-first approach |
4.2.3. | StoreDot's batteries |
4.2.4. | Safion |
4.2.5. | TWAICE |
4.2.6. | Oorja Energy |
4.2.7. | Addionics |
4.2.8. | Monolith AI |
4.2.9. | Speedgoat |
4.2.10. | DNV Energy Systems via Veracity |
4.2.11. | NOVONIX and SandboxAQ |
4.2.12. | Cell testing players summary |
4.3. | Business analysis for AI in cell testing |
4.3.1. | Typical business models |
4.3.2. | Differentiation |
4.3.3. | Challenges |
4.3.4. | AI is well-placed to revolutionize the cell testing process for battery development, but it will take time |
5. | CELL ASSEMBLY AND MANUFACTURING |
5.1. | Overview |
5.1.1. | Overview of traditional manufacturing process |
5.1.2. | Data quality challenges |
5.1.3. | Data acquisition challenges in industrial settings |
5.1.4. | AI for defect detection and quality control |
5.1.5. | AI for manufacturing process efficiency |
5.1.6. | Algorithmic approaches in manufacturing and cell assembly |
5.1.7. | Digital twins |
5.1.8. | FAT/SAT |
5.2. | Smart battery manufacturing players |
5.2.1. | CATL - smart factories |
5.2.2. | CATL - manufacturing process optimization |
5.2.3. | Siemens Xcelerator |
5.2.4. | Samsung Robotic Laboratory: ASTRAL |
5.2.5. | Voltaiq |
5.2.6. | BMW Group and University of Zagreb |
5.2.7. | EthonAI |
5.2.8. | Elisa IndustrIQ |
5.2.9. | Smart battery manufacturing players summary |
5.3. | Business analysis for smart battery manufacturing |
5.3.1. | Types of smart battery manufacturing players |
5.3.2. | Challenges |
5.3.3. | Smart factories could become standard for larger players, but startups will struggle to adopt |
6. | BATTERY MANAGEMENT SYSTEM ANALYTICS |
6.1. | Overview |
6.1.1. | Battery management in mobility and ESS - the need for accurate diagnostics |
6.1.2. | Management of multi-cell battery packs - a basic example |
6.1.3. | The purpose of a BMS |
6.1.4. | The data pipeline - from BMS to AI |
6.1.5. | Data structures and forms for diagnostics |
6.1.6. | Fault detection and analysis |
6.1.7. | SoH and SoC determination for lifetime optimization |
6.1.8. | The genesis of 'prescriptive' AI |
6.1.9. | Algorithmic approaches to battery system management |
6.1.10. | The Battery Passport |
6.2. | Players in AI for battery diagnostics and management |
6.2.1. | ACCURE Battery Intelligence |
6.2.2. | TWAICE |
6.2.3. | BattGenie |
6.2.4. | volytica diagnostics |
6.2.5. | On-edge AI |
6.2.6. | Samsung: Battery AI in S25 |
6.2.7. | Eatron and Syntient |
6.2.8. | LG Energy Solution and Qualcomm |
6.2.9. | Tesla BMS: optimization over a journey |
6.2.10. | Cell diagnostics players summary |
6.3. | Business analysis for AI-assisted battery diagnostics and management |
6.3.1. | Business models |
6.3.2. | Differentiation |
6.3.3. | Challenges |
6.3.4. | Data-focused battery analytics will take off in Europe and see growth in the wider mobility industry |
7. | SECOND LIFE ASSESSMENT |
7.1. | Overview |
7.1.1. | Second-life batteries: an overview |
7.1.2. | Determining the second-life stream |
7.1.3. | Safety concerns and regulations |
7.1.4. | The battery passport |
7.1.5. | The use of AI |
7.1.6. | Algorithmic approaches and data inputs/outputs |
7.2. | Players in AI for second-life battery assessment |
7.2.1. | ReJoule |
7.2.2. | volytica diagnostics and Cling Systems |
7.2.3. | NOVUM |
7.2.4. | DellCon |
7.2.5. | Second-life assessment player summary |
7.3. | Business analysis for AI-assisted second-life assessment |
7.3.1. | Revenue streams - somewhat ambiguous |
7.3.2. | Types of players |
7.3.3. | Differentiation |
7.3.4. | Challenges |
7.3.5. | AI for second-life assessment in batteries will become the norm in Europe |
8. | FORECASTS |
8.1. | Diagnostics by capacity served |
8.2. | Diagnostics by market value |
8.3. | Cell testing by market value |
8.4. | Second-life assessment by market value |
9. | COMPANY PROFILES |
9.1. | ACCURE |
9.2. | Addionics |
9.3. | Aionics Inc. |
9.4. | BattGenie Inc. |
9.5. | Chemix |
9.6. | Eatron Technologies |
9.7. | Elisa IndustrIQ |
9.8. | Eonix Energy |
9.9. | EthonAI |
9.10. | Monolith AI |
9.11. | Oorja Energy |
9.12. | ReJoule |
9.13. | Safion GmbH |
9.14. | Schrödinger Update |
9.15. | SES AI |
9.16. | Silver Power Systems |
9.17. | StoreDot |
9.18. | TWAICE |
9.19. | Voltaiq |
9.20. | volytica diagnostics |
9.21. | Wildcat Discovery Technologies |
10. | APPENDIX A: DATA CENTRES DRIVING BATTERY DEMAND |
10.1. | A note on battery demand |