China EV/ESS battery From EVs to AI data centres: The next wave of energy storage demand

Research evidence excerpt

China EV/ESS battery From EVs to AI data centres: The next wave of energy storage demand

Equities ● Automobiles

11 June 2026

Exhibit 3. ESS application scenarios in AIDCs

Typical duration Current penetration /

Application Target requirement adoption

Back-up power Provides uninterrupted power during grid outages Minutes to hours High. Back-up power remains

and bridges the gap before diesel generators or the most established storage

alternative back-up systems start. Increasingly application in data centres and

important as AI training interruptions can result in is already widely deployed via

substantial economic losses. UPS systems.

Power quality AI training clusters generate millisecond-level power Milliseconds to Low and largely limited to

management spikes and load fluctuations (e.g., All-Reduce seconds (high- advanced AI clusters; expected

synchronization), which can cause voltage sag, GPU power, low-energy to increase as rack power

throttling or system shutdowns. Storage acts as a applications; often density rises and AI workloads

fast-response power buffer to stabilize power quality. BBUs or scale.

supercapacitors)

Grid interconnection Helps reduce peak demand, manage load profiles 2-4 hours typically Low but growing. Emerging as

& peak shaving and facilitate grid connection approvals in regions one of the primary drivers of

with transmission constraints. Also lowers demand new AIDC storage deployments,

charges and grid upgrade requirements. particularly in North America.

Renewable Enables higher penetration of solar and wind power 4-6 hours today; Very low; primarily used in pilot

integration (green by shifting renewable generation across time and potentially 6-8+ projects or renewable-powered

power firming) maintaining reliability.