General Motors just made its biggest play yet for the energy business, and AI’s growing appetite for electricity is the reason why. At a San Francisco event, the automaker rolled out a package of vehicle-to-grid features, a new commercial battery storage strategy, and a charging tool meant to make public charging less of a headache, according to The Verge AI. The pitch is simple: as AI data centers strain the grid, the hundreds of thousands of EV batteries sitting idle in American driveways could help carry the load.
What stands out here is the framing. GM isn’t selling this purely as a green story. It’s positioning EVs as backup power plants on wheels, a way to grab a slice of the multibillion-dollar energy storage market it’s been chasing for nearly four years.
What GM Actually Announced
- Vehicle-to-grid (V2G) goes live. GM is pushing a firmware update to existing vehicle-to-home customers, letting their cars send stored energy back to the grid during peak demand. Owners who already have the equipment get the update automatically, no new hardware required.
- A massive installed base. The Verge AI reports there are already over 250,000 bidirectional-capable Chevy, Cadillac, and GMC EVs on US roads. Combined, GM says their batteries hold enough juice to theoretically power 120,000 homes for a full week.
- Sodium-ion batteries for the grid. Partnering with California’s Peak Energy, GM is developing sodium-ion chemistry for industrial-scale storage. Sodium is cheaper to source, more stable, and handles cold weather better than lithium. GM says it fits commercial storage well but not EVs, since it prioritizes longevity and cost over range.
- Energy Pass for charging. A new app feature lets GM EV owners find, start, and pay for charging across third-party networks, including Tesla, Electrify America, and IONNA, without separate accounts for each. EVgo and ChargePoint are planned next.
Why This Matters
EVs are unique in their ability to push power back to the grid, not just pull from it. Most are built with bidirectional charging, which turns a high-capacity battery into a backup cell for a home or the broader grid. With data center demand climbing, GM is betting utilities will want access to that distributed storage even as EV sales cool.
“We see a future where electric vehicles, batteries that power them, and the country’s power grids work together,” GM chief product officer Sterling Anderson said in prepared remarks, per The Verge AI. There’s a potential payday for owners too. Anderson argued V2G can lower aggregate energy costs and “create a potential financial return for the consumer” while making the grid more reliable.
The Real-World Tests
GM is already running this theory in two states:
- Northern California: A partnership with PG&E to build a localized fleet of 52,000 EVs for grid-balancing, targeted to be operational by 2030.
- Michigan: Work with DTE Energy to stress-test bidirectional charging using 30 of GM’s own employees’ homes.
The company is also teaming with Redwood Materials to build storage from US-made batteries and “second-life” EV packs. For its own future vehicles, GM is betting on lithium manganese-rich (LMR) batteries to close the gap with China, and it’s adopting Tesla’s NACS charging standard.
The Catch
Flipping on V2G isn’t as easy as a switch. GM Energy VP Wade Sheffer published an open letter urging regulators to formalize V2G infrastructure, pointing to International Energy Agency findings that call it the technology with the largest hourly flexibility to limit future grid costs. His asks: the auto industry needs to work with government to educate the public, and utilities have to simplify the enrollment process so customers can actually opt in.
That’s the honest tension in this announcement. The hardware is largely in place, but the regulatory and administrative plumbing isn’t. GM launched its GM Energy spinoff back in 2022, and this is the most concrete step yet toward making parked EVs a grid asset. Whether utilities and regulators move fast enough to match it is the open question. For the full breakdown, check the original report at The Verge AI.