Alibaba, Tencent and ByteDance bet on batteries as AI fuels power demand surge

China’s leading cloud and internet companies are making a strategic push into energy infrastructure, particularly large-scale energy storage, to cope with the surging power needs of generative AI workloads and meet government green electricity mandates. Policy requirements have made “clean data centers” a central priority: by end-2025, newly built national hub data centers are mandated to use at least 80% renewable power. At the same time, energy forecasts and industry research suggest that AI expansion could multiply data center electricity demand tenfold compared to 2022 levels, fueling demand for accompanying storage systems.

Alibaba has committed to combining renewable generation and storage systems to back its AI ambitions. At its 2025 Apsara Conference, the company unveiled a CNY 380 billion ($53 billion) investment plan in cloud and AI infrastructure over three years. As part of this, Alibaba is advancing “solar + storage” deployment in multiple locations. For example, in Zhangjiakou it has partnered with Mingyang to develop a source-grid-load-storage integration project: 200 MW of wind, paired with a 40 MW / 160 MWh electrochemical storage system. Its Hangzhou data center already hosts a 50 MWh liquid flow battery system to buffer intermittent renewables and secure power stability.

Facing rapidly growing AI workloads, Alibaba says its massive utilization of energy storage is just beginning because achieving high renewable penetration in its data centers is inseparable from large-scale storage. Its cloud division now strives to integrate its AI models, power systems, and infrastructure in a full-stack approach.

Tencent is one of the early adopters of microgrid design combining wind, solar and energy storage at its own data centers. Its Huailai project includes an on-site 11 MW PV array, 150 kW wind turbines and 1.25 MW battery storage, supplying 71% of the site’s power demand. At its Tianjin site, the company is also incorporating energy storage to complement renewables and respond to grid dispatch.

ByteDance (operator of TikTok/ Douyin) has recently announced procurement plans for at least 200 MWh of lithium-ion storage systems, to be complemented with wind and PV as part of microgrid solutions. The goal is to deploy these systems at data centers by the end of 2025. Its energy strategy is tightly aligned with AI infrastructure spending: ByteDance says that each 100 MW of new AI computing facilities may generate demand for 450–800 MWh of storage.

The moves by Alibaba, Tencent, ByteDance and other Chinese IT giants highlight a broader structural trend: the growth of data center energy demand is beginning to lift energy storage from niche grid applications into a central role in the AI era. As more capacity is built, the temporal mismatch between renewable generation and compute demand – especially at night or peak loads – will make energy storage indispensable.

BloombergNEF forecasted that from 2025 to 2027, China’s data center and 5G network power consumption share may rise from roughly 3% to 6% of total electricity use, driving demand for 30% of new data center projects to include storage. Meanwhile, global data center storage demand is projected to climb to the hundreds of gigawatt-hours by 2030.

In this environment, cloud and internet companies with deep pockets and AI ambitions are well positioned to shape storage markets through vertical integration: owning generation, grid interfaces, and battery systems. The winners will be those that succeed in merging computing, energy and storage into an optimized, dispatchable platform – turning digital infrastructure from passive load into intelligent energy assets.