The AI craze in Silicon Valley is hitting a reality "ceiling" — there's not enough electricity. With the surge of generative AI, large models have an insatiable appetite for power, while the aging power grid system and lagging power infrastructure in North America are clearly unprepared for this computing storm. Industry forecasts suggest that this "power shortage crisis" is not temporary but rather showing signs of becoming a long-term trend.

This dilemma has opened up a trillion-dollar opportunity for the energy storage market. According to the latest data, the U.S. energy storage market is on the verge of a boom, with new installed capacity expected to reach 52.5 GWh by 2025. In addition to established players like California and Texas, emerging states such as Arizona are also starting to stockpile electricity aggressively.

The current North American energy storage market is driven by three main factors: one is energy arbitrage through buying low and selling high, the second is the capacity market that ensures grid stability, and the third, which is currently the hottest — the massive power demand from data centers. For the AI servers that run non-stop, energy storage systems are no longer optional "backup," but essential "lifesavers" to prevent outages.

Although the U.S. has introduced the "big and beautiful" Act (IRA) to extend subsidies and try to "exclude outsiders" by increasing domestic content, Chinese energy storage companies continue to be a headache for competitors. Companies like Sungrow and CATL, despite tariff and policy challenges, remain indispensable "hard currency" in the North American market due to their extreme cost-effectiveness and mature supply chains.

Industry estimates suggest that by 2027, U.S. energy storage capacity will break through the 110 GWh threshold. The next five years will be a critical window period in determining the global energy landscape. In this race between AI and power, whoever holds the "energy reservoir" of energy storage will secure a ticket to the future.