Grid batteries are one of the genuine cost success stories of the energy transition. They are also solving a narrower problem than most coverage implies.

The success is real

According to the International Energy Agency, 108 GW of new battery storage capacity was deployed worldwide in 2025, 40% more than in 2024. Installed capacity is now eleven times higher than in 2021. About 80% of the new capacity was utility-scale, and lithium-iron-phosphate, a lithium chemistry that trades energy density for cost and cycle life, accounts for around 90% of deployments.

The prices behind that build-out have collapsed. BloombergNEF put the average lithium-ion pack price at $108 per kilowatt-hour in 2025, down 8% on the year and a record low. Packs for stationary storage fell furthest, down 45% to $70/kWh, the sharpest drop of any segment. LFP packs averaged $81/kWh against $128/kWh for nickel-manganese-cobalt. Chinese packs averaged $84/kWh, with North America 44% higher and Europe 56% higher.

The limit hiding inside those numbers

Here is the qualifier that matters. The IEA notes that battery storage durations are lengthening but that "most projects still cluster around two hours," with an increasing number able to run four hours or more.

Two hours is enough to do the job batteries are mainly bought for today: shifting solar output into the evening peak, and providing fast frequency response to stabilize the grid. Both are valuable, and both are short.

The economics of a battery split into two parts. Power capacity, the rate at which it charges and discharges, is governed by the inverters and connection equipment. Energy capacity, how long it can sustain that output, is governed by the cells themselves. Extending duration means buying proportionally more cells, and in a lithium-ion system the cells dominate cost. Doubling from four hours to eight roughly doubles the expensive part while the revenue from arbitrage grows more slowly, because the price spreads a battery captures get thinner the further out you go.

This is not an engineering barrier. A lithium-ion system can be built for any duration. It is an arithmetic barrier, and it is why nearly everything deployed clusters at the short end.

The gap it leaves is the one a high-renewables grid eventually has to fill: multi-day lulls in wind and solar, which no amount of two-hour storage addresses.

What is being tried, and how far along each is

The alternatives are at genuinely different stages, and coverage tends to flatten that distinction.

Flow batteries, typically vanadium, store energy in liquid electrolyte held in external tanks. Because the tanks are decoupled from the cell stack, adding duration means adding cheap tank volume rather than expensive cells, which is exactly the property lithium-ion lacks. They are commercially deployed, mostly in China, though at small scale against lithium-ion and with higher upfront costs.

Sodium-ion substitutes abundant sodium for lithium. It is less energy-dense, which matters little for a stationary installation where floor space is cheap, and it performs better in cold conditions. The IEA has tracked rapid growth from a small base. It is entering mass production, but as a cheaper competitor to lithium-ion at similar durations rather than a long-duration solution.

Iron-air batteries work by reversibly rusting iron, using materials that are close to free. The energy density is poor and the round-trip efficiency low, which for multi-day storage may be acceptable trade-offs. Form Energy is the leading developer and has a commercial pilot at Minnesota scale. This is a pilot, not a proven technology, and it should not be described as anything else.

Molten salt thermal storage stores heat rather than electricity, and is well established in concentrated solar power plants, where it is coupled to the plant rather than operating as a standalone grid battery.

The reason for caution

Long-duration storage has a long record of technologies that raised money, generated favorable coverage, and did not reach commercial scale.

Ambri, developing a liquid-metal battery out of MIT research, raised substantial venture funding and filed for bankruptcy after failing to solve sealing problems in its cells. Aquion, which made saltwater batteries, went bankrupt after failing to raise further funding. Several vanadium flow developers have also failed. The pattern, as Chemical & Engineering News has documented, is that these companies need either a utility willing to commit before the technology is proven or investors willing to fund a decade of losses, and both are scarce.

What is different now is demand-side. Utilities are writing long-duration storage into resource plans, regulators are designing markets that pay for capacity rather than only energy, and the AI data-center build-out has created large buyers who want firm power and are not especially price-sensitive. Form Energy's agreements are with that class of customer.

That improves the odds. It does not change the position today, which is that lithium-ion at short duration is cheap, proven and being deployed at record scale, while the multi-day problem remains unsolved at commercial scale by anything. Anyone reading a storage announcement should check three things: the chemistry, the duration in hours, and whether the project is a pilot or a commercial deployment. The third question is the one that most often goes unasked.