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

The Numbers Behind The Need

The scale of the challenge is substantial. According to Cervicorn Consulting and Market Techie, the global data-center battery market was valued at roughly $3.38 billion in 2025 and is projected to reach nearly $6 billion by 2035, driven largely by the AI infrastructure boom.

Lithium-ion battery shipments dedicated to AI data-center energy storage are projected to surge from 12 gigawatt-hours in 2025 to 272 gigawatt-hours by 2030. At the same time, more than 17 million electric vehicles were sold globally in 2024, creating a rapidly expanding pipeline of battery packs that will eventually be retired from transportation use.

Re-Teck occupies a unique position in this emerging market. The company receives retired battery packs from automakers including Tesla, Lucid, and BMW while also providing decommissioning services for major AI data center operators, including Microsoft. That gives it access to both the battery supply and potential customers.

Unlike batteries commonly deployed alongside renewable-energy projects, lithium-ion batteries are designed to deliver rapid bursts of power. That makes them particularly useful for AI facilities, where electricity demand can fluctuate sharply. A large battery farm assembled from repurposed EV packs can absorb those swings while reducing stress on the grid.

China already requires wind and solar facilities to be paired with energy-storage systems that absorb excess generation and release it when needed. Re-Teck’s proposal applies a similar concept to the demand side: battery farms located adjacent to data centers that charge during periods of low demand and discharge during peak consumption.

The idea is not merely theoretical.

In the summer of 2025, Redwood Materials—the battery recycling company founded by Tesla co-founder JB Straubel—deployed what it described as the largest second-life battery storage system in North America. Located at its Sparks, Nevada campus, the project consists of a 12-megawatt, 63-megawatt-hour microgrid assembled from hundreds of repurposed EV battery packs.

The system powers a modular data center operated by Crusoe Energy, which houses roughly 2,000 GPUs and other advanced computing systems. Since entering service, the project has reportedly maintained 99.2% operational availability. By early 2026, Crusoe announced plans to expand the facility from four to 24 modular data-center units, increasing power demand to approximately 20 megawatts.

“Where needed, we incorporate onsite power generation and battery energy storage systems to reduce pressure during peak periods, and we cover the cost of grid infrastructure upgrades to avoid shifting those burdens onto local ratepayers,” write Chris Dolan and Jamie McGrath of Crusoe in a blog post.

Data Center Demands

The concept is promising, but skepticism is understandable. Data-center operators spend enormous sums guaranteeing uptime, and any new component added to the power chain receives intense scrutiny. Repurposed batteries vary in age, chemistry, and degradation rates, making performance less predictable than that of newly manufactured systems.

Safety is another concern. Thermal runaway risks must be carefully managed through rigorous testing, monitoring systems, and battery-management software. Certification standards and insurance requirements also remain hurdles for widespread adoption.

Re-Teck acknowledges it is still building the case studies and cost analyses needed to move conversations from curiosity to contracts. But economics is moving in its favor. Bloomberg New Energy Finance found that new lithium-ion battery pack prices for stationary storage fell to roughly $70 per kilowatt-hour in 2025, a decline of about 45% from the previous year. Repurposed battery packs sourced directly from automakers can be significantly cheaper.

But the significance of battery reuse extends beyond electricity costs. It also touches one of the most overlooked vulnerabilities in the AI race: supply-chain dependence.

Consider bismuth, a metal increasingly used in electronics, soldering, and liquid-cooling systems. As AI processors become more powerful, traditional air cooling is often insufficient, making advanced cooling technologies increasingly important. China controls roughly 90% of the global bismuth supply. If that material becomes a geopolitical pressure point, AI infrastructure development could feel the impact.

Bismuth is only one example. Copper, rare earth elements, and advanced semiconductors present similar vulnerabilities.

Roughly half of all U.S. copper scrap is shipped overseas for processing, much of it to China. Meanwhile, China controls approximately 85% of global rare-earth processing capacity. As geopolitical competition intensifies, such concentrations pose risks to industries dependent on uninterrupted access to critical materials.

The case for domestic recycling—whether of EV batteries, server hardware, or the materials embedded within them—is therefore more than an environmental argument. It is increasingly a national security argument. Every battery pack that is repurposed rather than dismantled and exported represents a modest step toward greater supply-chain resilience.

There is an irony in all of this. The electric-vehicle revolution was largely promoted as a response to the environmental challenges posed by fossil fuels. Now, the batteries from that revolution are finding a second life, helping power the infrastructure behind artificial intelligence.

“If you have to buy new batteries to do this, it will never be practical,” says Li of Re-Teck. “Data center operators are drawing energy from the grid 24/7, peak and off-peak. They know that if they have an alternative, like a storage unit, they can supplement peak-hour energy with off-peak energy. But nobody has invested in building this energy storage farm because batteries are too expensive.”

The electricity challenge facing AI will not be solved by any single technology. Nuclear power, renewable generation, demand-response programs, and grid modernization will all play important roles. But repurposed EV batteries offer something rare in today’s infrastructure debate: a solution that exists now, scales with a growing waste stream, and reduces dependence on vulnerable supply chains.

In the race to power artificial intelligence, that combination is worth watching.