When Consumption Meets AI | The Impact of Artificial Intelligence on the Energy Storage Industry: Symbiosis or Conflict?
First Financial – March 9, 2025 – Author: Lu Ruyi, Editor: Le Yan
Artificial Intelligence (AI) is transforming energy management from large power plants to smart devices, driving improvements in intelligence and efficiency. However, challenges remain, such as the stark contrast between the rapid increase in electricity demand driven by AI and the stability of power supply, as well as the high energy consumption of AI technologies versus the goals of green transformation.
With the rise of platforms like Deepseek and Manus, AI is making significant inroads into the new energy sector, heralding an imminent revolution in energy storage. Yet, the high energy consumption and safety concerns associated with AI warrant careful consideration.
The Surge of AI in Energy Storage
AI is propelling changes in the energy storage industry, particularly by enhancing safety and efficiency in storage power plants and expanding market demand for energy storage. According to reports from First Financial, domestic renewable energy companies are gradually integrating AI technologies into their photovoltaic and energy storage operations, shifting energy management towards greater intelligence and efficiency.
“Large photovoltaic power plants face constantly changing generation data and market fluctuations. Without intelligent management systems, operators might struggle to make optimal energy dispatch decisions in a timely manner, which can negatively impact revenue maximization,” said a representative from LONGi Green Energy.
Cui Jian, President of Xiamen Kehua Data Technology Co., Ltd., expressed to First Financial that their goal in the AI and energy storage sector is to deeply integrate AI throughout the entire lifecycle of energy storage systems. “We aim to enhance the operational efficiency, safety, and economy of storage systems through AI technology.”
“Our strategic direction includes three key areas,” Cui explained. “First, we optimize energy management and dispatch of storage systems using AI to predict photovoltaic generation and user electricity demand, thus refining the charging and discharging strategies. Second, we employ AI for intelligent operations and real-time monitoring of equipment status, reducing failure rates and providing early warnings of potential issues. For instance, our AI model can predict the health status of storage batteries, enabling proactive maintenance and preventing system downtime.”
“AI big data can also enhance the value of energy storage systems in accommodating renewable energy and regulating power grid frequency,” Cui added, noting that AI systems can respond swiftly to frequency regulation needs, achieving millisecond-level power adjustments to bolster grid stability.
For example, the Risen Cloud energy management system from LONGi Green Energy can optimize power plant operations by analyzing operational data and user habits, ensuring optimal performance under various conditions. Additionally, the platform supports private deployment, allowing clients to maintain control over their data for privacy protection.
“We are incorporating technology across various scenarios, experimenting with popular model frameworks to assist in training and inference,” said Huo Jialong, Vice President of Lingchu Energy. In the smart operations sector, the company is developing an integrated platform for energy generation, grid, load, and storage, utilizing algorithmic models for power generation forecasting, load forecasting, and electricity trading.
According to Huo, the company is also building a cloud-based collaborative platform for intelligent operations in energy storage plants, focusing on data analysis of battery cells. “We are training specific AI models for safety diagnostics and lifespan analysis.”
The growing demand for AI computing power is beginning to have a notable impact on the energy storage sector, reflected in the order performance of companies. “The increasing demand for AI computing power is creating new market opportunities for the energy storage industry,” said Long Yan, a leader in a domestic energy storage company, to First Financial. As fields like data centers and cloud computing expand rapidly, the requirements for stable and reliable power supply are intensifying, showcasing the critical role of energy storage technologies in balancing supply and demand.
“Our orders indeed reflect this trend,” Cui stated, noting a significant increase in storage orders from data centers and computing hubs in recent years. Recently, the company provided a solar-storage integrated solution to a large cloud service provider, ensuring stable power supply while reducing operational costs through intelligent scheduling.
Huo also mentioned that recent orders have shown a clear positive response due to the emergence of Deepseek. “The cost of computing power has dramatically decreased, and the introduction of market-oriented pricing for renewable energy is pushing energy storage companies to transition from purely competing on price to offering comprehensive solutions, with AI technology at the core.”
Existing Conflicts
However, the relationship between AI and the renewable energy sector is not without its challenges. The rapid growth in electricity demand driven by AI and the stability of power supply are at odds, as are the high energy consumption of AI and the goals of green transformation. Training large AI models is leading to an explosive increase in electricity demand from data centers. Yet, the capacity for renewable energy absorption remains limited, and the development of AI adds further pressure on the grid in terms of power distribution and load capacity.
Sources within the industry indicate that if a data center suddenly requires extensive AI computational tasks, it could draw a large amount of electricity from the grid within seconds, potentially disrupting the stability of the local grid.
Nonetheless, Huo believes that the conflict between AI energy consumption and green transformation is temporary and localized. “Currently, many computing centers are being developed in the northwest, utilizing surplus solar and wind energy directly, which can alleviate some of the pressure. Additionally, green electricity from the northwest can be transferred to regions facing power shortages,” he explained. He further noted that, with the advancement of a national electricity market and the introduction of regional energy consumption regulations, this conflicting relationship can gradually be resolved.
Cui agrees that while AI large models require substantial electricity support, the intermittent and fluctuating nature of renewable energy does pose challenges to the grid. However, he believes this conflict is temporary and can be resolved through technological advancements that allow for synchronized development of both sectors. For instance, innovations could lead to the development of more efficient AI algorithms and hardware that reduce computational demands.
“Currently, the competition in the AI large model sector is no longer solely about raw computing power but rather about the efficiency of algorithms,” Huo noted, indicating that the primary issue with renewable energy absorption lies in the underutilization of solar and wind resources in the northwest, revealing challenges in power management and utilization.
Regarding the energy consumption issue, Cui mentioned that while AI training and inference indeed require significant computing power, efforts to optimize algorithms and hardware design are underway to gradually lower energy consumption. “Our company is adopting lightweight models and low-power chips to minimize computational requirements.”
“The opportunities presented by AI applications in the renewable energy sector far outweigh the challenges,” Huo stated, emphasizing that as large model technologies continue to evolve, the high energy consumption associated with AI computing power will be alleviated to some extent. Particularly after the introduction of DeepSeek, the cost of using AI has dropped significantly, subsequently reducing the demand for computing power.
Long Yan explained that companies can leverage AI to optimize the operational strategies of energy storage systems, charging during peak renewable output periods and discharging during low-demand times. This “renewable energy + energy storage” model can stabilize the fluctuations of renewable energy and ensure stable power supply.
Moreover, many AI applications in the renewable energy sector require high data security standards. An insider from an energy storage company remarked that for instance, in real-time electricity trading, any data tampering could lead to financial losses. “As security issues become increasingly exposed, there is a significant market opportunity for security companies. Businesses can follow up with effective security strategies and rely on market mechanisms to address security concerns.”
“Utilizing encryption technology and AI-driven security monitoring can effectively prevent cyber-attacks and data breaches,” Cui summarized, noting that these challenges are part of the technological development process and will gradually be resolved as technology advances and industry collaboration improves.
“AI technology in the renewable energy sector is diverse, and future directions will surely involve deeper integration of various AI technologies,” Huo remarked, suggesting that the core of competition will focus on specific applications and the ability to harness effective data. If the economic benefits are substantial, even complex challenges are worth pursuing; conversely, without sufficient economic viability, even simple technologies may lack significant value.
With AI’s powerful data analysis and intelligent decision-making capabilities, there is a growing consensus in the industry that the integrated solution of “AI + source-grid-load-storage” is the future trend, promising more accurate forecasts of generation, load management, and energy storage, thereby injecting new dynamism into the sector. Cui noted that AI technology can deeply integrate power sources, grids, loads, and storage systems, optimizing energy system operation strategies in real-time, dynamically adjusting storage charging and discharging, and predicting load demands to maximize energy utilization efficiency.
“Optimized source-grid-load-storage systems can also provide clean power to AI computing centers, reducing operational costs and carbon emissions from these facilities,” Cui added.
Huo envisions that the future national grid will evolve into a vast source-grid-load-storage platform, with AI serving as both a lubricant and accelerant for efficient operations of this system.
