As renewable generation scales up, storage becomes the critical enabler. But storage is not a single technology — it is a family of options with radically different efficiencies, cost profiles, and suitable applications. Round-trip efficiency (RTE) is the simplest metric for comparing them, and the differences across technologies are large enough to drive very different conclusions about cost and suitability.
What RTE measures
Round-trip efficiency is the ratio of electrical energy recovered from a storage system to the electrical energy put in:
Every percentage point of RTE lost is energy that became heat, sound, or chemical by-products during the charge-discharge cycle. At grid scale, these losses are significant — a system storing 100 GWh that is 75% efficient wastes 25 GWh per cycle, which at 1.25M in lost energy value per full cycle.
How major technologies compare
| Technology | Typical RTE | Primary loss mechanism |
|---|---|---|
| Lithium-ion batteries | 80–92% | Ohmic resistance, heat during charge/discharge |
| Vanadium redox flow | 70–80% | Pumping energy, shunt currents, membrane crossover |
| Flywheel | 80–90% | Bearing friction, aerodynamic drag, motor/generator losses |
| Pumped hydro | 70–85% | Turbine/pump efficiency, pipe friction, evaporation |
| Compressed air (CAES) | 40–60% | Heat of compression lost; heat of expansion must be supplied |
| Hydrogen (P2G2P) | 25–45% | Cascaded losses across electrolysis, compression, and fuel cell |
These are not manufacturer claims — they are system-level efficiencies including all auxiliary loads, parasitic consumption, and conversion losses at both the charging and discharging ends.
Where the losses come from
Lithium-ion loses energy primarily through internal resistance (heat during charge/discharge) and a small self-discharge rate over time. The dominant degradation mechanism — lithium plating at the anode during fast charging — reduces cycle life but has modest RTE impact. Modern lithium-ion at moderate charge rates achieves 90–92% RTE.
Compressed air energy storage (CAES) has inherently lower RTE because compression heats the air and that heat is typically vented and lost. When the compressed air is later expanded to drive a turbine, it cools dramatically (Joule-Thomson effect) and must be reheated using natural gas or a heat store. Adiabatic CAES — which stores the compression heat and uses it during expansion — can approach 70% RTE but has not yet been demonstrated at commercial scale.
Hydrogen's low RTE is a consequence of cascaded conversion steps. The full power-to-gas-to-power chain looks like:
Even with optimistic assumptions at each step, you recover less than half the electricity you invested. This is not a technology maturity problem — it reflects fundamental thermodynamic constraints on each conversion step.
The hydrogen nuance
The low RTE of hydrogen is often cited as a disqualifying argument for hydrogen as an energy storage medium. But this framing misses an important distinction.
Hydrogen's most compelling near-term role is not as an electricity storage medium — it is as a clean industrial feedstock to replace the roughly 100 million tonnes of grey hydrogen currently produced annually from steam methane reforming for ammonia, refining, and chemical production. Displacing grey hydrogen with green hydrogen (from electrolysis) eliminates a substantial emissions stream without needing to recover the hydrogen as electricity.
As an energy carrier for seasonal or long-duration electricity storage, hydrogen's poor RTE is a genuine disadvantage. As a zero-carbon molecule for industrial processes that already consume hydrogen, the RTE framing is irrelevant.
Why no technology wins on all dimensions
RTE is only one axis. Combining it with storage duration, cost per kWh of capacity, cycle life, and geographic constraints reveals why no single technology dominates:
- Lithium-ion is exceptional for durations of 1–8 hours and daily cycling, but prohibitively expensive at multi-day or seasonal timescales
- Pumped hydro provides the cheapest long-duration storage and dominates installed capacity globally (~90% of grid storage by energy), but is entirely site-constrained
- Vanadium flow has almost unlimited cycle life (the electrolyte does not degrade) and is well-suited to industrial sites with multi-day storage requirements, but upfront costs remain high
- Compressed air and hydrogen are the realistic options for very long-duration (weeks to months) or very large-scale storage where lithium-ion would be prohibitively expensive
The conclusion for grid planning: storage is a portfolio problem. Optimizing for a single metric picks the wrong technology for most applications.
The broader point
The renewable grid does not need one dominant storage technology — it needs layered flexibility: short-duration batteries for daily solar shifting and frequency regulation, medium-duration flow or pumped hydro for multi-day variability, and long-duration options for seasonal balancing. The RTE of each technology is one input to this matching problem, not the answer to it.
High-RTE storage is not always the right choice. An 80% efficient lithium-ion system that costs 150/kWh for a 10-day application will lose on total cost despite its efficiency advantage. What matters is RTE in the context of the value being delivered and the cost of the alternatives — including demand response, transmission, and the option of simply overbuilding generation.