Energy Storage
Storage does not make energy. It moves energy in time, at a cost in efficiency, and the question that decides which technology fits is always how long it has to hold the charge.
~95%
of the world’s installed storage capacity is still pumped hydro
4 hrs
the duration nearly every new utility-scale battery is sized to deliver
What does storage do for a grid?
A grid without storage has to produce every kilowatt-hour at the instant it is consumed. Storage breaks that constraint. It absorbs energy when supply exceeds demand and returns it later, minus losses, which lets a system built around inflexible or weather-driven generation serve a demand curve that does not match it.
That single capability shows up as several distinct services, and they are sold separately. Moving energy from a cheap hour to an expensive one is arbitrage. Standing ready to inject power within seconds is a reliability service. Deferring a substation upgrade by shaving the local peak is a wires service. A battery can provide all three, but only one at a time, which is why storage economics depend as much on market design as on the hardware.
Seconds
Frequency response
Fast injection or absorption that holds the system at 60 Hz when a plant trips.
Hours
Shifting and peak shaving
Charge at midday, discharge into the evening ramp. The core use for today’s batteries.
Days
Riding through weather
Covering a still, overcast stretch. Expensive with lithium, which is why other chemistries are being tried.
Seasons
Summer into winter
Only chemical fuels and very large reservoirs hold energy this long at plausible cost.
The technologies, by duration
Every storage system has two costs that scale separately: the cost of power, meaning how fast it can charge and discharge, and the cost of energy, meaning how much it can hold. Batteries are cheap on power and expensive on energy, so they are built short. Pumped hydro and hydrogen are the reverse, which is what makes them candidates for long duration.
| Technology | How it stores energy | Typical duration | Round-trip efficiency | Limiting factor |
|---|---|---|---|---|
| Pumped hydro | Water lifted to an upper reservoir | 8–20+ hours | 70–85% | Needs two reservoirs with height between them; permitting runs a decade |
| Lithium-ion | Ions moved between electrodes | 1–8 hours | 85–92% | Cost of added energy capacity; degradation with cycling |
| Flow batteries | Charged liquid electrolyte in tanks | 4–12 hours | 65–80% | Lower efficiency and larger footprint; energy scales cheaply with tank size |
| Thermal | Heat held in salt, rock or water | 6–24 hours | 40–60% to power, higher as heat | Converting heat back to electricity is lossy; best used as heat |
| Compressed air | Air pressurized in caverns | 8–24 hours | 50–70% | Suitable geology; heat management on compression |
| Hydrogen and derived fuels | Electrolysis into a storable molecule | Weeks to seasons | 25–45% | Poor round trip; justified only where nothing else holds energy that long |
| Flywheels and capacitors | Rotating mass or electric field | Seconds to minutes | 85–95% | Almost no energy capacity; used purely for power quality |
Storage always loses something
Round-trip efficiency is the honest accounting of a storage asset. Put 100 kilowatt-hours into a lithium battery and roughly 90 come back. Put the same energy through electrolysis, compression, and a fuel cell and perhaps 35 come back. Neither number is disqualifying on its own, because what matters is the value of the energy at each end. Cheap curtailed midday power converted at 35% can still beat expensive gas at the January peak.
100 kWh in, how much comes back
Lithium-ion
~90 kWh
Pumped hydro
~78 kWh
Compressed air
~60 kWh
Hydrogen round trip
~35 kWh