Iron batteries that store power for 100 hours are starting to ship

Iron-air cells are now shipping from a factory in Weirton, West Virginia. The first 150-megawatt-hour system goes to Minnesota's Great River Energy in 2027. Xcel Energy has signed for 30 gigawatt-hours more to back a new Google data center, phased in between 2028 and 2031. None of this matters if the chemistry does not work as advertised. The chemistry is, in plain terms: iron rusts.

A Form Energy cell takes in oxygen from the air on discharge, oxidizes iron metal into rust, and releases electrons along the way. Charging runs current backward and reduces the rust back to metallic iron. The cell breathes, the iron rusts and un-rusts, and the round trip stores energy. It is a battery in the same sense a lithium-ion pack is a battery, but it is built for a different job.

The job it is built for is duration. Form rates its commercial design at 100 hours of discharge, against a baseline where most large industrial lithium-ion systems are sized for four to six hours. That gap is the point. A 4-hour lithium pack handles the evening peak after the sun drops. A 100-hour iron pack is meant to dispatch through a multi-day wind lull, or a week of cloudy winter in a renewable-heavy region. The two chemistries are not competing for the same hour. They cover different curves on the same load chart.

Iron is also cheap, abundant, and politically easier to source than the lithium, nickel, and cobalt in incumbent cells. Form's pitch is that the chemistry scales to commodity inputs and a domestic supply chain in a way that the lithium-ion buildout does not. Whether that pitch survives contact with full-scale production is the open question.

The first commercial answer arrives in 2027, when Great River Energy's 150 MWh system in Minnesota is scheduled to come online. That site is small in the scheme of the grid but real: a working iron-air installation feeding an actual cooperative, not a lab demo. Production cells from the Weirton factory will be the first ones any utility actually meters revenue against.

The Xcel-Google deal would put 30 GWh of Form's iron-air storage behind a single data center in Minnesota, with first phases online in 2028 and the full build running through 2031. At that energy capacity, Form says, it would be the largest battery project in the world. The deal is signed, but it is also phased over years, and a megawatt-hour that does not yet exist is not a megawatt-hour on the grid.

Several open questions will decide whether this becomes a category or a cautionary tale. Form's 100-hour figure is a manufacturer spec; independent test data on round-trip efficiency, capacity fade, and cost per delivered megawatt-hour is not yet public. Iron-air cells are also bulky and heavy compared with lithium-ion, so they are unlikely to show up in vehicles or short-duration frequency regulation. The chemistry's edge is cost and duration per unit of energy stored, not power density.

The data-center deal shows that hyperscalers are starting to plan for the grid they want, not the grid they have. Google is not buying a 30 GWh iron pack because iron is romantic. It is buying a long-duration buffer so that the wind and solar behind the same interconnection agreement can firm up enough to run a large compute load. If that model works at one site, it will be replicated. If it does not, the same iron-air chemistry will still have a place in seasonal storage, microgrids, and the long tail of renewable-heavy systems that need to ride out a calm week.

Form's first answer lands in Minnesota in 2027. The 30 GWh test follows in 2028. The category either compounds from there, or it does not.