Energy Storage Technologies Compared: Batteries, Pumped Hydro and More

In most of the world, wind and solar are now the cheapest new sources of electricity. They generate when the weather allows, though, not when people need power. Storage fills that gap, over timescales from milliseconds (frequency control) to months (seasonal shifts). No single technology covers all of them at a reasonable cost, so the real question is which technology fits which duration. This guide compares the main options on cost per kW and per kWh, round-trip efficiency, duration, lifetime and maturity.

Key metrics

• Power capacity (MW): the maximum charge/discharge rate.
• Energy capacity (MWh): how much is stored.
• Duration (h) = energy / power.
• Round-trip efficiency (RTE): energy out / energy in.
• Cycle life and calendar life: how many cycles and years pass before capacity fades.
• Levelised cost of storage (LCOS): lifetime cost per MWh discharged.

The comparison table

Cost ranges are indicative and swing widely with project size and region. Sources include NREL ATB, PNNL and BNEF.

Lithium-ion batteries: the workhorse

In about a decade, lithium-ion went from phones to power grids. BloombergNEF’s survey found that average pack prices fell about 20% in 2024 to around $115/kWh. Stationary storage packs were cheaper still, and Chinese LFP cells traded well below $60/kWh. A complete grid system (inverters, containers, controls and grid connection) costs more, but turnkey prices of $150–250/kWh are now common for large 4-hour projects outside the US, and lower in China.

Deployment is climbing fast. The IEA reports that battery storage additions in the power sector roughly doubled in 2023, and each year far more battery capacity is added than new pumped hydro. California, Texas, China, Australia, the UK and Chile lead.

LFP (lithium iron phosphate) now dominates stationary storage. Compared with NMC it is cheaper, uses no cobalt or nickel, lasts more cycles and is more thermally stable. Its lower energy density hardly matters for a container sitting on the ground.

What batteries do on the grid:

• Frequency response (sub-second), the first and most lucrative market.
• Energy arbitrage: charging on the midday solar surplus and discharging into the evening peak.
• Capacity / peak shaving, taking over from gas peakers.
• Congestion relief, which defers grid upgrades.
• Black start and grid-forming services. Newer inverters can stabilise weak grids.

Worked example: a 100 MW / 400 MWh battery

• Installed cost: $220/kWh → $88 million. Life 20 years, WACC 7% → CRF 0.0944 → $8.3 m/yr; O&M $2.5 m/yr.
• One full cycle per day, 330 days, 90% depth, ~90% average capacity retention: 400 × 330 × 0.9 × 0.9 ≈ 107,000 MWh discharged per year.
• Capital + O&M per MWh: $10.8 m / 107,000 ≈ $101/MWh.
• Charging at $30/MWh with 88% RTE: $34/MWh.
• LCOS ≈ $135/MWh. The project makes money if the average spread between discharge and charge prices, plus ancillary revenue, is larger than that.

In solar-heavy markets such as California, Australia and Spain, midday-to-evening spreads of $100–200/MWh are increasingly common. That is why batteries are going up there at scale.

Home batteries

A typical home battery stores 5–15 kWh and delivers 3–10 kW. It raises self-consumption of rooftop solar, can supply backup power in some setups, and can arbitrage time-of-use tariffs.

With an installed cost of €5,000–7,000, simple payback in this example is roughly 12–17 years. Subsidies, dynamic tariffs, high retail prices or low export rates shorten it. Model your own case in the solar battery calculator together with the solar panel calculator.

Pumped hydro: the long-lived giant

Pumped storage hydropower (PSH) accounts for roughly 190 GW of capacity worldwide (IHA) and still holds most of the world’s stored energy. When power is cheap, water is pumped from a lower reservoir to an upper one. When power is needed, it runs back down through turbines.

PSH needs the right topography, takes 5–10 years to build and costs billions, but a plant lasts a century. China is building most of the new capacity, with over 50 GW running and much more under construction. “Closed-loop” off-river sites, which the Australian National University’s global atlas has mapped in large numbers, have smaller environmental impacts. Size a scheme in the pumped hydro storage calculator.

Long-duration storage (8+ hours)

Once renewable shares pass 60–70%, grids need storage that can carry them through nights, windless days and weather fronts. The main candidates:

• Flow batteries keep power (the stack) separate from energy (tanks of electrolyte), so extra hours are cheap to add. Vanadium is the most mature; zinc-bromine, iron and organic chemistries are still developing. China has commissioned several vanadium projects of 100+ MWh.
• Iron-air batteries store energy for ~100 hours by “rusting” and “un-rusting” iron, with very low materials cost. Form Energy began its first utility deployments in the US in 2024–2025.
• Compressed air in salt caverns: Huntorf (Germany, 1978) and McIntosh (Alabama, 1991), with new adiabatic plants in China and Canada.
• Gravity storage using blocks or mine shafts, still mostly at demonstration stage.
• Thermal storage in sand, rocks or molten salt, used either directly for heat (very efficient) or to make power again (less efficient).

Seasonal storage: hydrogen and heat

Moving summer solar into winter means storing huge amounts of energy cheaply. Here cost per kWh is what counts and efficiency matters less. Hydrogen in salt caverns costs in the order of $1–2 per kWh of storage capacity or less, far below batteries, but 55–70% of the energy is lost in the round trip. Germany and the Netherlands are developing cavern hydrogen storage. Seasonal heat storage already pays its way in district heating, through large insulated water pits in Danish towns such as Vojens or through borehole thermal storage. See Green Hydrogen Explained and the green hydrogen calculator.

Matching technology to the job

Storage also competes with other sources of flexibility: demand response, interconnectors, smart EV charging and flexible generation. A water heater, or a heat pump with a buffer tank, makes a cheap “thermal battery”, and an EV holds 50–100 kWh.

Storage in practice: what large grids are learning

• California. According to the state grid operator, CAISO, grid-scale battery capacity went from a few hundred MW in 2020 to well over 10 GW by 2024–2025. On many summer evenings, batteries are now the largest single source of supply for a few hours. They move surplus midday solar into the evening peak and cut gas-fired generation.
• Texas (ERCOT). The market has no capacity payments, yet batteries grew quickly on the back of volatile prices and ancillary services. They have helped avoid emergency conditions during summer peaks more than once.
• Australia. The Hornsdale Power Reserve in South Australia (commissioned 2017 at 100 MW, later expanded to 150 MW) showed how fast-responding batteries bring down the cost of frequency control. Australia now has some of the highest rooftop solar penetration anywhere and a large pipeline of 2–4 hour batteries to replace retiring coal plants.
• United Kingdom. Batteries first took share in fast frequency response and moved into energy trading once that market was saturated. Many other countries have followed the same pattern.
• China is now the largest market for new grid storage. Provincial rules require new solar and wind farms to add storage, and standalone LFP projects are growing fast.

The same lesson keeps coming up. The first batteries in a market earn high revenue from ancillary services, but those markets are small and fill up quickly. Over the long run the business case rests on energy arbitrage and capacity value, both of which grow as solar and wind shares rise.

Safety and sustainability

Battery fires are rare but serious. LFP is much less prone to thermal runaway than NMC, and modern containers come with gas detection, fire suppression and spacing standards (NFPA 855, IEC 62933). Worries over materials (lithium, cobalt, nickel, graphite) are easing with LFP and sodium-ion, and recycling capacity is expanding under the EU Battery Regulation. For pumped hydro, the impacts come from land and water use.

Bottom line

Storage is no longer the bottleneck it was ten years ago. In sunny markets, lithium-ion batteries are now cheap enough to shift solar into the evening peak, and they are being built faster than any other grid technology. Pumped hydro is still the backbone of long-duration storage, while flow batteries, iron-air, compressed air and thermal storage compete for the 8–100 hour market. Seasonal balancing will probably rely on a mix of hydro, interconnection, hydrogen and firm low-carbon power. See how storage fits into each country’s grid on the world energy map, and run household numbers in the solar battery calculator.

Figures in this guide are typical ranges drawn from the sources above; your own numbers depend on location, prices and equipment. Use the linked calculators for site-specific estimates.