Four Hours or a Hundred? Grid Batteries Find Their Own Lanes

It might sound like an exaggeration to say Google bet on a giant lump of iron instead of a lithium battery — but that is roughly what happened. In February 2026, Google and the utility Xcel Energy announced a record-setting grid battery for a Minnesota data center: 300 megawatts of capacity and 30 gigawatt-hours of total storage, built not from lithium-ion cells but from an "iron-air" battery that runs on little more than iron and air. As solar and wind take on a larger share of the grid, utilities increasingly need to store electricity not just for a few hours but for several days at a stretch. Why would a technology that is, on paper, less efficient than lithium-ion be the one chosen for that job? The answer says something about how battery technology itself is starting to specialize by task.

A Four-Hour Battery and a Hundred-Hour Battery Aren't the Same Job

The battery Google chose comes from Form Energy, and it is being built in Pine Island, Minnesota. At 300 MW and 30 GWh, it can reportedly discharge continuously for up to 100 hours — among the largest grid storage systems ever attempted. Google is separately funding 1.4 gigawatts of wind power and 200 megawatts of solar to feed the system, while Xcel Energy manages the surrounding grid infrastructure.

Why 100 hours? Smoothing out an ordinary evening dip in solar output only requires a few hours of storage. But when cloudy, windless weather stretches on for several days — an increasingly recognized "weather event" for grid planners — a battery that only lasts a few hours leaves a gap that generation simply can't fill. For a facility like a data center that can't tolerate downtime, that multi-day gap is exactly the risk this kind of long-duration battery is meant to close.

About Half the Efficiency — and Still Chosen

The iron-air battery's biggest weakness is round-trip efficiency — how much of the electricity put in comes back out. Form Energy has not published a figure for its production systems. In March 2026 the trade outlet Energy-Storage.news cited roughly 40 to 50 percent from a private market report, and about 40 percent as the estimate of battery engineer Jim McDowall. The same article puts lithium-ion at 85 to 90 percent. If those estimates are right, an iron-air battery loses more than half the electricity it stores.

What makes it attractive anyway is the price of the electricity going in. When solar and wind output outpaces what the grid can absorb, utilities routinely curtail — throw away — that surplus power, which then costs next to nothing. Charging a low-efficiency battery with electricity that would otherwise be wasted makes the efficiency penalty far less costly in practice. Iron itself is also one of the most abundant metals on Earth. Xcel Energy, too, points to capital cost compared with the alternatives as a reason for choosing it — though no specific cost per kilowatt-hour has been disclosed.

Rusting and Un-Rusting Iron — Where the Real Chemistry Problem Lies

The basic mechanism is simple enough. During discharge, iron reacts with oxygen to form iron oxide — rust — releasing electrons as usable electricity in the process. Charging reverses the reaction, turning the rust back into iron. In principle, it's nothing more than iron slowly rusting and un-rusting, over and over.

The real barrier to commercialization lies elsewhere. A 2015 review of iron-air battery research (McKerracher et al.) points to two problems. One is the battery's "air electrode," which has to carry out both the oxygen-absorbing reaction during discharge and the oxygen-releasing reaction during charging, reliably, over a long life. The other is on the iron side, where an unwanted side reaction releases hydrogen gas during charging, wasting some of the electricity that goes in. A 2026 paper (Zhao et al.) likewise frames the challenge, as its title says, around improving the kinetics of the iron electrode's oxidation-reduction reaction — in plain terms, getting the iron to rust and un-rust as cleanly and repeatably as possible.

Neither Lithium Nor Iron — Sodium Enters as a Third Option

The specialization doesn't stop at a lithium-versus-iron split. In July 2026, Chinese battery giant CATL and the Dutch firm Alfen announced a plan to deploy a total of 5 gigawatt-hours of sodium-ion grid storage in Europe. Sodium is said to be roughly a thousand times more abundant than lithium and carries far less price-volatility risk. CATL's own figures put these sodium-ion systems at 15,000 charge cycles with a 25-to-30-year service life at 70 percent of their original capacity — well beyond the few thousand to roughly 10,000 cycles typically rated for lithium iron phosphate batteries.

In an electric vehicle, energy density — how much energy fits into a given weight and volume — matters above almost everything else. A stationary grid battery doesn't face that constraint; what matters instead is cycle life, safety, tolerance for a wide range of temperatures, and cost. On those terms, sodium-ion batteries appear to be encroaching on the several-hour storage niche lithium-ion has long owned, while the multi-day, hundred-hour tier still seems to belong to iron-air for now.

Whether iron-air can hold that position, though, has not yet been shown in practice. Form Energy's hardware is at the stage where a pilot system for the Minnesota cooperative Great River Energy (ground broken in 2024) has begun operating; nothing on the scale of 30 gigawatt-hours has yet run as designed. Lithium iron phosphate prices keep falling, and sodium-ion is now a live competitor too. Until measured efficiency and cost figures are published, the idea that multi-day storage belongs to iron-air remains an expectation rather than a result.

The story of grid batteries in 2026 looks less like one technology winning and more like a transition — from a single all-purpose battery to a toolbox where the right chemistry depends on the job at hand. A battery that squeezes a few hours of power out at high efficiency and a battery that accepts lower efficiency in exchange for several days of insurance are doing different jobs. What is clear is that batteries now have to be sorted by how many hours of electricity they hold, and what for.