AI-Driven Innovations Set to Transform the Energy Storage Sector by 2025

The Next Competitive Frontier in Energy Storage: AI-Driven Value Ocean

As of now, the energy storage industry is at a pivotal moment in its market transformation. The introduction of Document 136 marks the end of the era dominated by policies, transitioning the energy storage sector from policy-driven dynamics to a market-led approach. Document 394 sets a clear target for achieving comprehensive coverage of the electricity spot market by the end of 2025, signifying a rapid acceleration in China’s electricity market reform.

In this context, the focus of technological innovation in energy storage is shifting from a price-driven “volume competition” model to a more refined pursuit of value. The deep integration of artificial intelligence (AI) across various sectors of energy storage is fundamentally reshaping the industry’s value chain, positioning itself as the next level of competition within the industrial framework.

Emerging Energy Storage Companies Going Public: Betting on “AI + Energy Storage”

On April 28, a rising energy storage company from Jiangsu, Guoxia Technology, officially began its process for listing in Hong Kong, aiming to raise capital to accelerate its “AI + Energy Storage” strategy and establish a differentiated competitive edge. Founded in 2019, Guoxia originally focused on the European residential energy storage market, with over 70% of its revenue coming from Europe in 2022. As the domestic energy storage market rapidly expands, Guoxia is now simultaneously developing both domestic and international markets.

Thanks to the explosive growth in global energy storage demand over the past two years, Guoxia’s performance has surged. Their prospectus shows that revenue is projected to soar from 142 million yuan in 2022 to 1.026 billion yuan by 2024, achieving a compound annual growth rate (CAGR) of 168.9%. During the same period, gross profit is expected to rise from 35.6 million yuan to 155 million yuan, with a CAGR of 108.6%. Furthermore, Guoxia’s competitive position in the energy storage market continues to strengthen, ranking as the eighth largest supplier of multi-purpose energy storage systems globally in 2024 and the tenth largest for residential energy storage systems.

To support its growth in the energy storage sector, Guoxia is also accelerating its capacity expansion. However, similar to many manufacturing companies in their expansion phase, Guoxia faces the challenge of increasing revenue without corresponding profit growth. From 2022 to 2024, net profit is expected to rise only from 24.27 million yuan to 49.11 million yuan, while gross margin has sharply declined from 25.1% to 15.1%. To overcome this challenge, Guoxia is focusing on its “AI + Energy Storage” strategy.

As an industry pioneer, Guoxia has integrated seamless cloud solutions for energy storage system solutions and products, establishing a digital and AI energy research and development team in its first year. They are the first company to develop IoT platform solutions tailored for their operational industry and the first to create seamless energy storage industrial models based on AI technology. Currently, Guoxia has developed AI-optimized systems and tools, such as Safe ESS and Hanchu iESS, to enhance real-time energy optimization, predictive maintenance, and decision-making processes. After six years of focused development, Guoxia has become a leading provider of renewable energy solutions and products driven by platform technology and AI in China’s energy storage industry. In 2024, their intelligent energy storage system solutions will contribute 97.8% of their total revenue, representing the core driver of their growth and performance.

Consequently, Guoxia aims to enhance its competitiveness and differentiate its advantages by offering AI-optimized energy storage system solutions and products across all energy usage scenarios. The funds raised from its Hong Kong listing will primarily focus on enhancing R&D capabilities to strengthen its technological leadership in energy storage and building an overseas operational and service network to support its international growth strategy. This not only signifies a dual breakthrough for Chinese energy storage companies in technological innovation and capital operation but also heralds the arrival of a new era characterized by the deep integration of AI and energy storage.

New Trends in Energy Storage Technology by 2025: AI Empowering the Entire Industry Chain

In light of the accelerated global energy transition and the transformative impact of AI on industries, artificial intelligence is deeply integrating with the energy storage sector, fundamentally reshaping the industry’s value chain. If 2024 marks the initial stages of “AI + Energy Storage,” then 2025 is poised to be a turning point for their deep integration. Analyzing the direction of technological and product iterations within the industry this year, from battery technology advancements and precise forecasting to smart scheduling, AI’s integration with energy storage is significantly enhancing energy utilization efficiency while providing new solutions to the “impossible triangle” of a new energy system (safety, economy, and greenness).

Specifically, AI plays a transformative role in several key areas of the energy storage industry chain, including technology development, electricity trading, power plant operation and maintenance, and scenario expansion. It drives battery technology innovations, with leading energy storage companies leveraging intelligent and digital technologies to achieve qualitative leaps in key metrics such as battery energy density, cell and system capacity, charge-discharge efficiency, and cycle life. For instance, CATL has developed an intelligent design platform for battery materials, capable of completing material screening and closed-loop verification in just 90 days using AI algorithms. BYD is working to enhance design efficiency and identify opportunities for new materials and systems through AI technology in various aspects of battery design and manufacturing process control. LG Energy Solution is using AI to customize batteries for clients, while XINWANDA’s NoahX 2.0 system employs a digital twin platform to simulate cell aging trajectories in real-time, dynamically adjusting charge-discharge strategies to extend battery life by 20%.

Smart operation and maintenance are compressing the total lifecycle costs of storage systems. By applying deep learning algorithms, AI integrates data from multiple sources such as voltage, current, and temperature to accurately estimate the state of charge (SOC) and state of health (SOH) of batteries while accurately predicting their remaining lifespan using techniques like electrochemical impedance spectroscopy. Additionally, AI can intelligently adjust charge-discharge strategies based on grid load, renewable energy generation, and the status of energy storage systems to improve storage efficiency and reduce costs. For example, CATL’s “Tianheng Smart Storage” platform utilizes AI algorithms to advance fault warning times by seven days, enhancing overall efficiency by 3% and improving maintenance response times by 50%. Envision Energy has introduced the world’s first intelligent body storage system, EN 8 Pro, which integrates AI meteorological models and load forecasting algorithms to achieve over 90% accuracy in predicting electricity price peaks and troughs, leading to an annual revenue increase of 1.5 million yuan for a 100MWh power station. Kehua Data Energy has launched an integrated solution for photovoltaic energy storage and charging, optimizing energy scheduling with AI to reduce the data center’s Power Usage Effectiveness (PUE) to below 1.2.

Energy storage safety management is becoming proactive. The era of passive monitoring relying solely on voltage and temperature is over. The deep integration of AI with energy storage systems is reshaping the underlying logic of energy safety, primarily constructing a lifecycle safety protection system through intelligent prediction, dynamic adjustment, and proactive defense. For instance, Sungrow has developed an AI model for battery cells that achieves a thermal runaway warning accuracy of over 99% by integrating mechanistic models with deep learning algorithms, enhancing SOH estimation accuracy by 2-3% and increasing annual power generation by 7.3GWh for a 1GWh power station. Linyang Energy’s “Cloud-Edge Collaborative Multi-Modal Safety Chain Big Model” can provide fault warnings seven days in advance, reducing false positive rates by 40%.

Maximizing system benefits in electricity trading. The electricity market globally exhibits diverse regional demand and supply conditions, with varying storage needs across different areas. The deep integration of AI and energy storage is advancing energy management into a new intelligent phase, significantly improving the economic viability and reliability of energy storage systems through intelligent scheduling, safety protection, and lifecycle management. For instance, Sigrid New Energy has optimized energy storage system charge-discharge strategies with AI technology, resulting in a 42% reduction in users’ electricity purchase costs and a 100% increase in electricity sales prices, leading to a total cost reduction of 52%. Haier New Energy’s Star Engine 261 commercial energy storage system, built on “AI + Digital Twin + Big Data + Human-Machine Collaboration” technologies, achieves a high system efficiency of 90%, occupies only 1.19 square meters, offers 12 layers of safety guarantees, and provides intelligent maintenance across its entire lifecycle, making it the preferred choice for high-energy-consuming enterprises seeking cost reduction and efficiency enhancement.

Expanding the application boundaries of energy storage. AI enables energy storage systems to play crucial roles in more complex scenarios. In constructing new power systems, grid-connected energy storage technologies are increasingly gaining importance. AI can help address the inertia support issues associated with high proportions of renewable energy integration, while projects combining thermal power with energy storage for frequency modulation are on the rise, significantly shortening response times. On the user side, integrated photovoltaic storage projects are becoming more prevalent in industrial parks, facilitating cross-park electricity trading through Virtual Power Plant (VPP) aggregation. With the increasing penetration of residential energy storage, comprehensive energy solutions featuring “photovoltaics + energy storage + smart home” are gradually emerging. Additionally, data centers, which form the basis of AI operations, require highly reliable and stable power supply; thus, energy storage systems combined with intelligent scheduling can provide reliable backup power while optimizing their energy usage efficiency.

Driving Changes in Energy Storage Business Models

From a business model perspective, AI is introducing new profit growth points and operational models to the energy storage industry. In electricity market trading, leveraging AI’s precise predictions of market prices and supply-demand relationships allows energy storage systems to participate more efficiently in spot markets and ancillary service markets, thereby generating greater profits. For example, AI can analyze user behavior and market data to forecast energy storage demand across different regions and industries, assisting enterprises in adjusting production and market strategies to explore new market spaces. In the lifecycle management of energy storage systems, AI achieves comprehensive digital and intelligent management from planning and construction to operation and retirement, reducing operational costs and enhancing asset value.

Six Core Challenges Awaiting Resolution

As AI and energy storage systems move from conceptual collision to deep integration, this technological union, heralded as a “catalyst for the energy revolution,” faces multiple real-world tests. From conflicts between algorithmic logic and data silos to the lagging nature of the industrial ecosystem and standard systems, the process of empowering energy storage with AI is generating a series of challenges that require urgent solutions.

Firstly, AI algorithms demand high-quality and quantity data; however, many enterprises currently lack comprehensive lifecycle data accumulation, leading to issues such as scarce samples and non-standardized data, which affect the accuracy and reliability of AI models. Industrial collaboration also faces the “data silo” effect, with natural barriers to data sharing among battery manufacturers, system integrators, and operation service providers. Secondly, the construction and operational costs of energy storage systems remain high, particularly for large-scale deployments, where the procurement, installation, and maintenance costs of energy storage equipment can become bottlenecks for power storage projects. Thirdly, existing energy storage technologies may not fully meet the demands of high-intensity and high-frequency applications, necessitating further breakthroughs in the development and application of large-scale, long-duration energy storage and high-efficiency energy storage systems. Fourthly, there are security concerns regarding data. AI requires sensitive data such as cell-level voltage and internal resistance to enhance prediction accuracy; however, the leakage of such data could expose critical technologies like battery formulas and thermal management designs. Additionally, cross-border data flows face compliance barriers, with regulations like the EU’s AI Act categorizing energy storage AI systems as “high-risk AI” and requiring at least 70% of training data to come from the target market area; similarly, the U.S. CFIUS prohibits energy storage systems with Chinese AI components from connecting to the federal grid. These localized data sovereignty requirements compel multinational companies to establish localized AI training centers, increasing operational costs and risking decision inconsistencies due to delays in model parameter synchronization. Fifthly, there is a dilemma regarding the attribution of responsibility in algorithmic decision-making. When AI systems cause energy storage accidents due to prediction errors, legal accountability becomes ambiguous. A 2023 incident in Germany involving a storage power station led to a fire due to AI misjudgment, with courts struggling to determine whether the fault lay with the battery manufacturer’s hardware, software supplier’s algorithm, or operator’s parameter configuration. This “dispersed responsibility” phenomenon hinders insurance companies from developing targeted AI energy storage liability insurance products, raising industry risk costs. Lastly, the lagging and fragmented nature of standard systems has led to increased testing costs for enterprises due to the lack of comprehensive safety standards for AI in energy storage. Current global safety standards for AI in energy storage are still in a “patchwork” construction phase. For instance, IEC 63056 states that AI-involved energy storage systems must have interfaces for human intervention but does not clarify the thresholds for algorithmic decision explainability; similarly, China’s NB/T 42091 requires that AI diagnostic functions of Battery Management Systems (BMS) achieve an accuracy of ≥95% but does not define error tolerances under different working conditions.

In response to these challenges, the industry is exploring solutions. On a technical level, promoting “AI + Mechanistic Model” hybrid modeling, such as the battery aging prediction model developed by MIT that combines DFT calculations and neural networks, has reduced prediction errors to 0.8%. On an industrial level, establishing cross-enterprise AI energy storage safety alliances, such as the “Energy Storage AI Reliability Testing Platform” led by the China Electric Power Research Institute, which has covered 80% of mainstream manufacturers. On an institutional level, exploring “sandbox regulatory” models, with the EU planning to establish innovative pilot areas for energy storage AI by 2025 to allow for breakthrough technology testing within controlled limits.

Conclusion

As AI technology continues to evolve and innovate, its integration with the energy storage industry will deepen and broaden. In the future, competition within the energy storage sector will largely revolve around AI technology. Those who can seize the initiative in the AI-empowered energy storage arena and achieve breakthroughs in technology, application, and business will distinguish themselves in the fierce competition of the energy storage market, leading new trends in industry development. The convergence of AI and energy storage fundamentally represents a deep coupling of the digital and physical worlds. The challenges posed are not merely technical issues but also entail the restructuring of industrial ecosystems and institutional frameworks. Despite the immense potential, truly unlocking the ultimate value of “AI + Energy Storage” will require collaborative efforts across disciplines, policy support, and continuous technological advancements. In the short term, the focus can be on specific scenarios (such as battery life prediction), while long-term goals must involve constructing an “AI + Energy Storage” ecological system.