By Kyle Proffitt
June 16, 2025 | While EV, grid storage, and consumer electronics may be the most common battery applications in the news, there are some challenging and exciting edge cases. At the 2025 International Battery Seminar and Exhibit earlier this year, SpaceX principal engineer Ray Barsa spoke about the very unique conditions encountered in low earth orbit (LEO) affecting battery performance.
Satellites from SpaceX
SpaceX has developed the Starlink service, based on more than 7,000 satellites delivering internet to about 10 million people, and Barsa says the reach is truly global, including access—even in Antarctica—that enables things like FaceTime calls. The numerous satellites communicate with each other using lasers, creating a global mesh through which data can find routes to and from ground stations at any time. Because the speed of light is faster in a vacuum than through fiber, the latency can actually best terrestrial internet solutions, especially over long distances. Unlike geostationary satellites such as those associated with satellite TV, the LEO satellites circle the earth every 90 minutes. The tradeoffs are that an individual satellite serves a smaller area, requiring the mesh and complicated information transfer, but proximity to ground reduces latency and renders it a more viable internet solution.
Space Batteries
Barsa explained that their satellites need a good bit of power, both for running the lasers and for maneuvering. They use Hall effect argon thrusters to achieve and maintain orbit, avoid space junk, and de-orbit at end of life; these thrusters use electricity to ionize argon gas, and then a magnetic field accelerates the ions and releases them to create thrust. Barsa shared that each 500-kg satellite is powered by a battery pack with energy density over 230 Wh/kg, which is recharged using solar arrays during solar exposure. In total, they have over 80 MWh of orbiting batteries, a figure suggesting about 11 kWh/satellite.
Barsa then talked their unique challenges and opportunities. On the one hand, they don’t have to worry about fast charging. He says the satellites run about 15 partial depth of discharge cycles per day, with no rest. He commented on Jeff Dahn discussing cell orientation as an important parameter affecting salt imbalance, but said that in microgravity, there isn’t necessarily a meaningful measure of cell orientation. The vacuum of space can also cause cell swelling and leakage. Additionally, “many plastics, foams, some common adhesives just don’t work in a vacuum… they disappear, they outgas,” Barsa said. Other strange things happen too. He showed what he suggested may be the only image of arcing occurring during thermal runaway. “What we’re seeing here is plasma generated from a secondary metal vapor arc… if that arc is able to liberate enough metal, that metal gets ionized and spreads through the pack as plasma, and you’re toast.”
Satellites also face a unique challenge of micrometeorite and other space debris, “things like grains of sand traveling at Mach 25,” Barsa said, which can impact the batteries. In low earth orbit, you also deal with atomic oxygen—unpaired, highly reactive single atoms of O, which like to strip away materials—and UV exposure, which darkens materials. He gave the example of a white paint designed for radiative properties getting darker from UV, at the same time as fresh material becomes exposed by atomic oxygen stripping the surface layer.
Charging and Discharging
The satellites experience roughly 1 hour in the sun and 30 mins in the earth’s shadow for each orbit, translating to about C/2 charging. Longevity is critical, though. “We want to get the equivalent of 5,000 full cycles out of off-the-shelf cells from tier 1 OEMs,” Barsa explained. He went on to report that with cells from several years ago, cycling at a modest depth of discharge and C/2 rate, “we’re getting 90% retention up to 2,000 full cycles.” The batteries get no rest, but the cycling profile is very predictable. “You don’t have an aggressive driver of a satellite,” Barsa explained. He shared that their max depth of discharge is about 50%, but this number has increased as they’ve optimized various parameters. They also avoid higher voltages, at least until it’s time to de-orbit. Prior to this, he said they always keep some reserve energy to maneuver if something goes wrong.
“One thing that SpaceX really holds important in the design here is that we don’t have ‘bricked’ satellites on orbit… what that means for the battery is SOC [state of charge] estimation is critical,” Barsa said. For this reason, LFP chemistry is challenging. He shared that they are using NMC chemistry 2170 cells.
Temperature Considerations
Barsa said that the temperature of battery packs is an important parameter for optimization. “We’ve emphasized a lot of thermal radiant improvements in the pack, and we’re able to hold less than 5 °C across all the cells in the pack,” he said. He explained that because they get passive cooling via radiation, “you’re able to tailor your surfaces to either absorb or project heat however you want, and if you know your C-rates and you know your heat generation, you can adjust emissivity using absorptivities to get really nice thermals in the pack.” This also translates to wins for low-resistance cells, as less ohmic heating occurs, and more insulation can be used.
But What If They Burn?
Barsa said that the critical runaway thresholds that work for EVs don’t hold up well in space, because the heat only dissipates by radiation and can thus accumulate more easily. SpaceX is also in a complicated position because they design their satellites to burn completely at end of life; they never get them back to run further analysis, and there is some conflict in designing an object to not burn—until the time is right and that’s exactly what you want it to do.
Barsa said that SpaceX performs extensive testing, looking at the potential for sidewall rupture of cells and thermal runaway events in the pack. However, he said, they “have never even seen a single cell runaway in orbit.” Still, they’ve run simulations of what would happen in a runaway event, showing at most that it would “increase micrometeroid and orbital debris at our altitude about 20%, for about a month, depending on the solar cycle.” He explained that the decreased altitude they use means materials demise more quickly as they fall back toward earth and noted they are below both the International Space Station and the Chinese Space Station.
Unlike other battery applications, SpaceX has little incentive to save money on the batteries. “The cost of the cells in the pack is about 1%… we want the smallest pack that can achieve the mission,” Barsa explained. He said their takt time (effectively their demand) is 1-2 seconds per cell, enabling them to try many different processes; they’re fine with a new technology that isn’t quite gigafactory-ready.
