Report 017 · Energy Storage
Google's new "world's largest battery" hands back only about half the energy you put into it. On a lithium spec sheet that would be a disqualifying flaw. Here it's the whole point, because the battery isn't doing lithium's job.
In February 2026, Google and utility Xcel Energy announced plans for a battery in Pine Island, Minnesota that the coverage called the largest ever announced by energy capacity: 300 megawatts of power, 30 gigawatt-hours of storage, able to run for a hundred hours straight. That is a genuinely huge machine, and the headlines treated its size as the story.
The number the headlines skipped is the one an engineer looks for first. This battery, Form Energy's iron-air chemistry, has a round-trip efficiency of roughly 40 to 50 percent. Put ten megawatt-hours in, get four or five back. A lithium-ion battery returns 85 to 90 percent. If you judged iron-air by the standard we use for every other battery, it would look broken. It isn't. It's built for a different job, and the gap between those two facts is worth understanding.
Iron-air does something almost quaint: it stores energy by rusting iron, on purpose, and then un-rusting it. Form describes the principle as "reversible rusting." When the battery discharges, iron metal reacts with oxygen from the air and turns to iron oxide, plain rust, releasing electrons as it goes. To charge it back up, you run current the other way, which strips the oxygen back off and returns the rust to metallic iron. The active ingredients are, in Form's own words, "low-cost iron, water, and air." There is no lithium, no cobalt, no nickel, none of the constrained supply chains that price and politics fight over.
That chemistry buys two things. The first is duration: the system can "cost-effectively store and discharge energy for up to 100 hours," which is more than a day of continuous output, where a typical lithium grid battery is built to run for two to four. The second is cost. Form claims the technology stores energy "at less than 1/10th the cost of lithium-ion." Cheap, abundant materials and a container that's mostly iron and water will do that.
What that chemistry does not buy is efficiency. Rusting and un-rusting iron is a lossy round trip, and the roughly half of your energy that doesn't come back out is the price of admission. Independent analysis puts iron-air's round-trip efficiency near 40 to 50 percent against lithium's 85 to 90, and Form's own product materials, tellingly, quote the duration and the cost but not the efficiency.
I don't read that as hiding the ball, exactly. I read it as a company that knows efficiency is the wrong question for its product, and would rather not invite the wrong comparison. But if you're going to understand why anyone builds this, you have to hold the ugly number in view: a big fraction of the electricity you store here is simply lost. The interesting part is why that can still be the right choice.
Efficiency matters when you cycle a battery constantly. A lithium pack that charges cheap at midday and discharges into the evening peak, every single day, lives or dies on how much it gives back each trip; a 15 percent loss compounds into real money over thousands of cycles. That's daily energy arbitrage, and it's lithium's job. Judged there, a 50 percent battery would be a disaster.
Iron-air is not for that. As one analyst put it bluntly, "It's not about daily arbitrage. It's about the Dunkelflaute," the German grid term for a dark, still stretch when solar and wind both go quiet for days at a time. Form framed the Minnesota deal the same way: Xcel and Google "were looking for a solution capable of storing energy for multiple days at a time in order to support reliable, around-the-clock power on a grid with growing renewable penetration." You are not cycling this battery daily. You are filling it slowly over good stretches and holding a multi-day reserve for the rare bad one. It might do a handful of deep discharges a year.
When the battery sits mostly full and only earns its keep during a once-a-season lull, efficiency stops being the metric that matters. What matters is whether you could afford to build enough of it to cover a hundred hours at all, and there, lithium is the one that fails: at lithium prices, a 100-hour lithium battery is absurd. So the honest comparison isn't "iron-air loses half, lithium loses a tenth, lithium wins." It's "iron-air can cover a four-day gap for a tenth the material cost, and lithium can't be built at that duration for any sane price." Different tools. The Georgia Power project makes the split concrete: a 15-megawatt system rated at 1,500 megawatt-hours, a hundred hours of runtime, deliberately sized for endurance rather than for a quick, efficient daily punch.
The part that fascinates me is what a 100-hour asset does to the control software. A four-hour lithium battery is a sprinter you tell to run hard every day; you optimize each cycle for efficiency and squeeze the arbitrage. A 100-hour iron-air battery is a reservoir you might tap three times a year, and the decision that matters is when, whether the weather ahead justifies spending down a reserve you can only refill slowly. Those are different problems, and a cycling strategy tuned for one is wrong for the other. I help design the AI battery-cycling systems for a veteran-owned (HUBZone) energy-storage integrator, the software that decides how a pack charges, discharges, and gets watched for drift, so this distinction isn't abstract to me: mixing up the two jobs is how you either burn a reserve you'll need or waste a sprinter idling. (Disclosure: I help design that company's AI; I don't own it and earn nothing from this link. Full policy here.)
"World's largest battery" is a real and impressive fact, and "only 50 percent efficient" is also a real fact, and the two are not in tension once you know what the battery is for. Iron-air isn't a worse lithium battery; it's a different device that happens to share the word "battery," built to hold cheap energy for days rather than to shuttle it efficiently every night. The mistake the coverage invites is judging it on lithium's scorecard. The right question for any storage project isn't "how efficient is it," it's "what job is it doing," and for the multi-day gaps that a renewable grid has to survive, a battery that rusts on purpose and gives back half is not a compromise. It's the tool that fits.
U.S. Army combat veteran (Counter-IED / Electronic Warfare), peer-reviewed researcher in microwave spectroscopy, and founder & CEO of Shroombiosis. Consults on laboratory operations, AI, and supplement formulation.