Old EV Batteries Could Help Solve AI’s Exploding Power Problem

Old electric vehicle batteries, typically discarded after powering cars for a decade, are finding a second life as giant storage units for AI data centers. This unexpected solution is emerging as artificial intelligence's insatiable appetite for electricity threatens to overwhelm power grids worldwide.

Companies like B2U Storage Solutions are repurposing used EV battery packs into stationary energy storage systems that can absorb excess power from renewable sources and discharge it when AI data centers need a surge. The innovation arrives as AI computing power demand doubles every 100 days, with data centers projected to consume up to 8% of U.S. electricity by 2030, according to the Electric Power Research Institute.

The idea isn't new—second-life batteries have been used for grid storage for years—but the scale and urgency are unprecedented. AI training and inference workloads are intensely energy-hungry, and traditional backup diesel generators or lithium-ion battery farms are expensive and environmentally taxing. Used EV batteries, which retain 70–80% of their original capacity, offer a cheaper, greener alternative.

B2U Storage Solutions operates a 5.9-megawatt facility in Lancaster, California, using 1,300 second-hand EV battery packs from Honda and Nissan. The system stores solar power during the day and feeds it to the grid during peak hours, smoothing demand curves. Other players include Connected Energy in the UK, which partners with Volvo and Renault, and Smartville in California, which reuses GM and Tesla batteries. A study by the National Renewable Energy Laboratory (NREL) found that repurposing used EV batteries could save 30–50% compared to buying new stationary storage.

“We're turning a waste stream into a revenue stream,” says B2U CEO Freeman Hall. “AI's power problem is real, but it doesn't have to mean building more power plants. It can mean using the batteries already on the road.” The dual benefit—reducing e-waste and providing low-cost storage—is attracting attention from major cloud providers like Microsoft and Amazon, which have set ambitious carbon-neutrality targets.

Yet challenges remain. Uniform standards for testing and certifying used battery health are lacking. Degradation rates vary wildly depending on the original vehicle's usage. And the supply of retired EV batteries is only now beginning to trickle in—most EVs on the road today are still in their first life. By 2030, however, an estimated 200,000 tonnes of EV batteries will be retired annually, creating a massive potential feedstock.

If the logistics work out, second-life batteries could transform AI's energy footprint. Data centers already use enormous amounts of electricity; the flexibility to charge from solar during the day and discharge during high-demand evening hours could slash both costs and carbon emissions. Researchers at MIT and Stanford are building AI models to predict battery degradation, which would help grid operators trust these repurposed systems.

The next milestones to watch: Volkswagen's planned second-life battery facility in Germany, expected to energize parts of Tesla's European charging network; California's grid operator integrating repurposed batteries into its capacity market; and the release of IEEE standards for second-life battery testing in 2027. If battery reuse scales as projected, the same lithium-ion packs that once propelled SUVs could soon be powering the neural networks that run the global economy.