Material Insights · Understanding the Material World

Storing energy: from fuel stocks to batteries: Choosing a store: efficiency, duration, cost and minerals

Understanding Energy / Module 10: Storing energy: from fuel stocks to batteries Part 4 of 4

Choosing a store: efficiency, duration, cost and minerals

No single storage technology does every job. The right choice depends on how long energy must be held, how often the store cycles, how much loss is acceptable and what materials it needs. Putting the options side by side shows why batteries dominate short durations, why long-duration storage is still unsettled, and why minerals have become a matter of state policy.

4.1 Duration and cycling decide the right store

A useful way to sort storage is by the job it does. Fast response over seconds to minutes favours flywheels, supercapacitors and batteries. Daily shifting over two to eight hours, such as moving midday solar into the evening, is the sweet spot for lithium-ion batteries and pumped hydro. Storage over several days, to ride through a week of poor weather, is contested between pumped hydro, flow batteries and other new battery types, compressed air and thermal stores. Seasonal storage is the domain of fuels: gas in caverns, hydrogen, and water held behind dams.

The economics follow a simple logic. A store's cost has two parts: the cost of its power equipment (pumps, turbines, inverters, electrolysers) and the cost of its energy capacity (cells, reservoirs, caverns, tanks). Batteries have a high cost per kWh of capacity but are efficient and quick to build, so they win when they cycle every day. Hydrogen in caverns has a very low cost per kWh of capacity but expensive conversion equipment and high losses, so it suits stores that are used rarely but for a long time. Pumped hydro sits between them.

How often a store cycles matters as much as what it costs to build. A battery cycled 350 times a year spreads its capital over 350 discharges; a seasonal store cycled once a year must recover all its capital from a single discharge. That is why purpose-built seasonal electricity storage is so expensive, and why existing reservoirs and fuel stocks are so valuable.

Round-trip efficiency matters most for stores that cycle often, because their losses repeat every day. For stores used rarely, low capacity cost matters more than efficiency. Demand response and vehicle-to-grid add a further, often cheaper, option: shifting when energy is used rather than storing it. The tile "Demand response and vehicle-to-grid" covers that option.

Data Electricity stores compared
StoreRound-trip efficiencyTypical durationMain limits
Lithium-ion battery (LFP)about 85 to 88%1 to 8 hoursCapacity fade, minerals, fire safety
Pumped hydroabout 75 to 80%6 hours to daysSuitable sites, long build times
Compressed air, older diabatic plantsabout 42 to 54%hours to a dayNeeds caverns; burns gas
Compressed air, adiabatic (target)about 60 to 70%hours to a dayStill being proven at scale
Liquid airabout 50 to 60%hours to daysLow efficiency without waste heat
Flow batteryabout 60 to 70%4 to 12 hoursBulky, higher cost
Flywheelabout 85 to 90%seconds to minutesHolds very little energy
Heat store, used as heatover 90% over short periodshours to seasonsCannot return electricity cheaply
Hydrogen, power to powerabout 25 to 40%weeks to seasonsLarge losses, costly equipment

Indicative ranges: PNNL assessments (lithium-ion 86 to 88%, vanadium flow 67.5 to 70%, hydrogen about 35%), Smil (pumped hydro), the Huntorf and McIntosh plants, and a 2025 review of liquid air. Field results can be lower: a utility vanadium flow battery tested by PNNL returned 54 to 63% including its auxiliary loads.

Source: Compiled from PNNL Energy Storage Grand Challenge methodology documents (lithium-ion, redox flow, hydrogen), Shafiee et al. (2017), She et al. (2025) and Smil, https://www.pnnl.gov/projects/esgc-cost-performance; see module sources

4.2 Storage hype meets arithmetic

Smil's long-running argument about energy transitions is that new technologies are routinely expected to scale faster than they do. Storage offers examples on both sides. Batteries have outrun many forecasts: their prices fell faster and deployment rose faster than most energy agencies expected a decade ago. Other ideas, from flywheel buses to a hydrogen economy, have been promised for decades and remain small.

Three popular claims deserve care. The first is that batteries will soon let a grid run on wind and solar alone. They can shift daily surpluses, and in sunny regions with many batteries they already supply a large share of evening demand. But covering a week of low wind in winter with batteries would need many times today's battery fleet in those regions, charged and discharged only rarely, which is very expensive. Smil makes the point with a large city that relied wholly on wind and solar: it would need storage measured in tens of GWh just to ride through a few calm, cloudy days.

The second is that electric vehicles will act as a giant grid battery. Car batteries are large in total, but owners need their cars charged when they want to drive, extra cycles wear the batteries, and the systems for coordinated vehicle-to-grid supply are still being built. Smart charging, which simply shifts when cars charge, delivers much of the benefit with less complication.

The third is that storage is clean by definition. A store is only as clean as the electricity that charges it, and every cycle loses some energy. A battery charged from a coal-heavy grid at night and discharged at noon, when solar is plentiful, would raise emissions rather than cut them.

None of this argues against storage. It argues for matching each store to its job, counting losses honestly, and keeping the cheap storage the system already has, notably hydro reservoirs and fuel stocks, while new forms are built.

  • Smil, Energy Myths and Realities: Bringing Science to the Energy Policy Debate (2010) · via Smil Mastery Course, M9 'Energy myths and realities'
  • Smil, Invention and Innovation: A Brief History of Hype and Failure (2023) · via Smil Mastery Course, M9 'Energy myths and realities'
  • Smil, Numbers Don't Lie: 71 Stories to Help Us Understand the Modern World (2020) · via Smil Mastery Course, M7 'Grids, storage and the real cost of electricity'
  • Smil, Energy Transitions: Global and National Perspectives (2017) · via Smil Mastery Course, M7 'Grids, storage and the real cost of electricity'
  • US Department of Energy, The history of the electric car (2014) · web

4.3 Critical minerals tie storage to a few countries

Batteries are built from a small set of materials: lithium, graphite, nickel, cobalt, manganese, iron, phosphate, copper and aluminium. Smil warns that the energy transition is material-hungry and will raise, not end, the world's demand for mining. Each GWh of batteries needs hundreds of tonnes of processed minerals, and grid storage is now one of the largest buyers.

Ore is fairly widely spread; processing is not. Australia, China, Chile, Zimbabwe and Argentina mine most of the world's lithium, the Democratic Republic of the Congo most of its cobalt, and Indonesia most of its nickel. But refining and manufacturing are concentrated. China refined about 79% of the world's cobalt in 2025, makes nearly all battery-grade graphite anode material and most lithium chemicals, and produces the great majority of cathode powder. The IEA's Global Critical Minerals Outlook 2025 found that the average share of the top three refining countries for key energy minerals had risen to about 86% in 2024, from around 82% in 2020, and that China was the leading refiner for 19 of the 20 strategic minerals it examined, with an average share of around 70%. China has used export controls on graphite and on battery and cathode technology, imposing and at times suspending them.

Prices swing hard. Battery-grade lithium carbonate averaged about USD 11,700 a tonne in the United States in 2021 and about USD 63,700 in 2022, then fell to about USD 9,000 in 2025 before rising again into 2026. Such swings bankrupt mines in slumps and delay new supply.

Responses include new mines and refineries outside China, recycling, and chemistries that use more abundant materials: LFP avoids nickel and cobalt, and sodium-ion avoids lithium. India has no commercial lithium mining or refining yet. Its production-linked incentive scheme offers Rs 18,100 crore for 50 GWh of domestic advanced cell manufacturing, and the state-owned KABIL has taken lithium blocks in Argentina. Building the refining, cathode and anode steps at home is the harder part.

See it in the table

H₂O WaterCu CopperAl AluminiumNi NickelMn ManganeseLi LithiumCo CobaltGr GraphiteH2 HydrogenVRF Flow batteriesV2G Demand response and vehicle-to-gridElz Electrolysers

Check your understanding

Why can hydrogen in salt caverns suit seasonal storage despite its low efficiency?

A seasonal store cycles perhaps once a year, so cheap energy capacity matters more than efficiency.

Which stage of the battery supply chain is most geographically concentrated?

Ore comes from several countries, but refining, cathodes and anodes are dominated by China.

A battery is charged from a coal-heavy grid at night and discharged at noon when solar is plentiful. What is the likely effect on emissions?

Storage is only as clean as its charging power, and each cycle loses some energy.
← Part 3: Mechanical, thermal and chemical storesMark read · Wrap-up →