Energy Storage Technologies Compared: Battery, Hydrogen, CAES, Flywheel, Sand Battery, and Proton Battery

Battery vs Hydrogen Energy Storage — The Core Long-Duration Debate

Choosing an energy storage technology is not about finding the single “best” option. It is about solving a three-body problem of efficiency, duration, and cost. No technology wins on all three dimensions. The real question is which trade-offs your project can afford.

Round-Trip Efficiency and Cost — Where Batteries Dominate

If efficiency is your primary metric, the answer is decisive. Lithium-ion batteries achieve a round-trip efficiency (RTE) of 85–95%, with grid-scale systems consistently delivering around 90%. Only 5–10% of input electricity is lost across a full charge-discharge cycle.

Hydrogen energy storage tells a very different story. Every step in the hydrogen pathway — electrolysis, compression, storage, and fuel cell reconversion — introduces losses. Electrolysis alone converts electricity to hydrogen at 70–80% efficiency. Compression and storage shave off another 10–15%. The fuel cell or turbine that converts hydrogen back to electricity operates at 50–60% efficiency. Multiply these together and the full-chain RTE lands at just 30–47%. For every 100 kWh you put into a hydrogen storage system, you get back roughly 30–47 kWh. A battery returns 85–95 kWh.

Think of it as water passing through buckets with holes. The battery system is a single bucket that leaks 5–15%. The hydrogen system runs the same water through three buckets in sequence, each leaking a portion. What reaches the end is roughly one-third of what you started with.

Cost follows a similar pattern, with one important caveat about duration. Lithium-ion batteries currently achieve a Levelized Cost of Storage (LCOS) of €120–180 per MWh for short- to medium-duration applications, according to a 2025 techno-economic assessment published in multiple peer-reviewed journals. Hydrogen systems remain above €250 per MWh across most analyses. But hydrogen’s cost curve is projected to fall by up to 60% by 2040 as electrolyzer manufacturing scales, according to industry forecasts. And here is the critical nuance: battery costs scale almost linearly with storage duration. Double the hours, roughly double the per-kWh cost. Hydrogen storage costs, by contrast, are dominated by the power conversion equipment (electrolyzer + fuel cell), not the storage medium itself. For very long durations, hydrogen’s per-kWh economics begin to improve relative to batteries.

Application Scenarios — The Duration Dictates the Choice

The efficiency-versus-duration trade-off plays out differently at each timescale. Your choice between battery and hydrogen is ultimately a choice about time.

For short-duration needs — seconds to roughly six hours — batteries are unmatched. Grid frequency regulation, daily solar peak shaving, and commercial tariff arbitrage all fall squarely in battery territory. High RTE means minimal energy waste, and the sub-20-millisecond response time satisfies the most demanding grid services.

For medium-duration needs of 6 to 24 hours, the landscape shifts. Batteries can still serve these applications, but the linear cost scaling becomes painful. Extending a battery from 4-hour to 8-hour duration nearly doubles the per-kWh capital cost. Hydrogen starts to become economically interesting in this window, particularly when the stored energy will be used infrequently or seasonally.

For long-duration and seasonal storage spanning days to months, hydrogen is currently the only commercially viable electrochemical option. A 2025 microgrid study found that properly sized hydrogen systems can achieve 0% Loss of Load Probability, whereas battery-only configurations left a 4.34% unmet load. The research community has converged on a clear insight: batteries handle the daily rhythm of renewable energy, and hydrogen covers the seasonal gaps. They are complementary, not competing.

Environmental Impact and Safety Considerations

Both technologies avoid substantial CO₂ emissions compared to fossil fuel equivalents — roughly 8.5 tonnes of CO₂ per year avoided in a typical building-scale application. On the manufacturing side, batteries carry a heavier carbon footprint (approximately 80–200 kg CO₂ per kWh of capacity, depending on cathode chemistry), while hydrogen equipment emissions depend heavily on whether the electrolyzer runs on green or grey electricity. LFP (lithium iron phosphate) batteries sit at the lower end of that range and are the dominant chemistry in stationary storage today.

On safety, the comparison is less about “which is safer” and more about “which risks can you manage.” Batteries carry thermal runaway risk, addressed through standards like IEC 62619 and UL 9540A. Hydrogen carries leakage and explosion risk, managed through ISO 22734 and proper ventilation design. Both are manageable with modern engineering. Neither should disqualify a technology from project consideration.

Compressed Air Energy Storage — The Grid-Scale Mechanical Alternative

Compressed air energy storage (CAES) takes a fundamentally different approach from electrochemical storage. It uses surplus electricity to compress air into underground caverns or pressurized tanks. When power is needed, that compressed air is released through a turbine to regenerate electricity. It is, in essence, a giant mechanical spring.

The numbers that matter: CAES achieves 45–60% round-trip efficiency. While this trails lithium-ion batteries, the capital cost comparison tells a more interesting story. BloombergNEF’s 2024 Long-Duration Energy Storage Cost Survey pegged CAES at an average global CAPEX of $293 per kWh — slightly below lithium-ion’s $304 per kWh for 4-hour systems. More importantly, CAES costs drop with each additional hour of storage duration, while battery costs scale linearly. For storage durations exceeding 8 hours, CAES becomes the more cost-effective choice in most markets.

The catch? Geography. Traditional CAES requires suitable underground geology — salt caverns, depleted gas fields, or deep saline aquifers. This limits deployment to regions with the right subsurface conditions. China leads global deployment with gigawatt-hour-scale projects already operational, including a 100 MW advanced CAES facility in Zhangjiakou. Outside China, commercialization remains early-stage, with companies like Canada’s Hydrostor developing 500 MW-class projects using purpose-built underground caverns. Advanced adiabatic CAES designs, which capture and reuse the heat generated during compression, promise to push RTE toward 65–70%. These remain largely pre-commercial.

For grid-scale projects with 8- to 24-hour storage requirements and suitable geology, CAES deserves a seat at the evaluation table alongside batteries and hydrogen.

Flywheel Energy Storage — Power Quality and Rapid Response

Among the five technologies covered in this article, the flywheel is the most invisible to the general public. Yet it occupies an irreplaceable niche in power quality and industrial UPS applications.

A flywheel energy storage system (FESS) works on an elegantly simple principle: an electric motor/generator spins a massive rotor to extremely high speeds inside a vacuum enclosure. When the grid needs power, the rotor’s kinetic energy is converted back to electricity. Because the only “wear” is on magnetic bearings that float the rotor without physical contact, a flywheel can endure over one million charge-discharge cycles with virtually no degradation. Compare that to the 5,000–10,000 cycle life of a typical lithium-ion battery, and the difference in longevity philosophy becomes clear.

The trade-off is equally clear: flywheels carry high initial capital cost, rapid self-discharge (making them unsuitable for storage beyond minutes to an hour), and lower energy density than batteries. What they lack in staying power, they make up in speed. Response times under 5 milliseconds make flywheels the go-to solution for power disturbances that batteries cannot catch — voltage sags, frequency deviations, and the split-second gap between grid failure and generator startup.

Real-world deployment data confirms this niche positioning. At a European tire manufacturing plant, a flywheel system covered 73% of peak-shaving demand, compared to 80% for lithium-ion batteries. The flywheel required zero cell replacements over a decade of operation. Beacon Power’s 20 MW flywheel frequency regulation plant in New York State has demonstrated the technology at utility scale. The emerging best practice is hybrid deployment: flywheels handle millisecond-level power quality events, while batteries handle minutes-to-hours of backup duration. Each protects the other’s weaknesses.

Sand Battery Thermal Energy Storage — Low-Cost Heat and Beyond

If the flywheel is the most invisible energy storage technology, the sand battery is the most counterintuitive. The idea is almost too simple: heat ordinary sand to 500–800°C using cheap renewable electricity, store it in a well-insulated steel silo, and extract the heat later as hot water, steam, or (in advanced configurations) electricity.

The construction is straightforward. A typical sand battery consists of a steel silo. Polar Night Energy’s Pornainen installation in Finland measures 13 meters tall and 15 meters in diameter, holding approximately 2,000 tons of soapstone — a natural rock with excellent heat retention. Resistive heating elements embedded in the sand bed raise the temperature during charging. Heat transfer pipes run through the storage medium; when energy is needed, air or water circulates through these pipes to extract the stored heat. Heavy insulation — mineral wool, ceramic fiber blankets — minimizes standby losses. Magaldi’s fluidized-bed MGTES system reports thermal losses under 2% per day.

The economics are compelling because the storage medium is nearly free. Sand costs essentially nothing compared to lithium, cobalt, or nickel. But the efficiency picture depends entirely on how you use the stored energy. If the output is heat — for district heating networks, industrial process steam, or building heating — the effective efficiency exceeds 90%. If you must convert the heat back to electricity through a Stirling engine or steam turbine, the round-trip efficiency drops to 4–32%, though research prototypes suggest 31.6% may be achievable with optimized designs.

Commercial deployment is accelerating. Polar Night Energy’s 8 MWh pilot in Kankaanpää, Finland, has operated since 2022. The 100 MWh Pornainen facility, commissioned in 2025, delivers 1 MW of heat output. In Italy, Magaldi’s MGTES system began producing industrial steam (120–400°C) from a fluidized sand bed in late 2024. For projects where the end use is heat rather than electricity, sand batteries represent a category-killer on cost. No other technology can compete with free storage media and decades-long system lifetimes.

Thermal management connects every energy storage technology at the engineering level. Whether you store heat in sand or manage the temperature of lithium-ion cells in a battery cabinet, precise thermal control determines both efficiency and lifespan. A BESS installation with inadequate cooling can lose 20–30% of its rated cycle life simply because operating temperatures drift outside the optimal 15–35°C window. The industry has learned to solve this: active liquid cooling in modern battery energy storage systems now maintains cell temperature within ±1°C across the entire rack. That precision directly translates to longer warranty periods and better return on investment.

Proton Battery Technology — The Next-Generation Contender

Proton batteries are not a lithium-ion replacement — at least not for the next decade. But their underlying physics make them arguably the most intriguing long-term candidate in energy storage research. The fundamental advantage sits right on the periodic table: a proton (H⁺) is roughly 100,000 times smaller than a lithium ion (Li⁺). This size difference enables a unique charge transport mechanism that no lithium-based system can replicate.

How Proton Batteries Work — The Grotthuss Advantage

A proton battery, as formally defined in a landmark 2024 review by researchers at RMIT University (Rezaei Niya, Heidari, and Andrews, Journal of Physics D: Applied Physics), is a rechargeable battery based on proton transfer and reversible electrochemical hydrogen storage. Protons are released from a source material — usually water — transported through a proton-conducting electrolyte, and stored in a porous electrode.

The defining mechanism is the Grotthuss proton hopping effect. Instead of a lithium ion physically diffusing through an electrolyte — a slow process, like one person pushing through a crowded corridor — protons hop along chains of hydrogen-bonded water molecules. Each proton jumps from one water molecule to the next, relay-style, without any single proton traversing the full distance. The result: proton conductivity in aqueous electrolytes can reach approximately 1 S/cm in 4M sulfuric acid, roughly ten times the lithium-ion conductivity in typical organic electrolytes. When hydrogen bonds shorten to about 2.5 Å, conductivity can spike to 1,389 mS/cm. Lithium systems cannot approach these numbers.

The practical implications are compelling. Fast charging (limited by transport kinetics, not by the Grotthuss mechanism), inherent safety (water-based electrolytes do not catch fire, and no hydrogen gas is produced in normal operation), and resource abundance (water is the proton source, not a mined mineral) make proton batteries a theoretically elegant solution for large-scale stationary storage.

Current Performance and Key Challenges

Laboratory results are promising but deserve clear-eyed context. The highest reported performance comes from RMIT’s activated-carbon-based design: 598 mAh/g specific capacity and 245 Wh/kg energy density. This approaches commercial lithium-ion territory (typically 150–250 Wh/kg for LFP packs). Other notable designs include a molybdenum oxide anode paired with a manganese dioxide cathode (210 mAh/g, 177 Wh/kg) and an organic quinone-based anode (208 mAh/g, 133 Wh/kg).

But every reported proton battery to date exists at the milligram scale in laboratory coin cells. The technology readiness level sits at TRL 3–4 — laboratory validation, nowhere near pilot production. Five critical challenges stand between today’s research papers and tomorrow’s commercial products.

First, cathode materials. There is a genuine scarcity of high-voltage, high-capacity cathode materials compatible with acidic aqueous electrolytes. Metal oxides, Prussian blue analogues, and organic materials are the three major research directions. None has yet demonstrated the combination of capacity, voltage, and stability that commercial deployment demands.

Second, acidic corrosion. The electrolytes that enable high proton conductivity — typically sulfuric or phosphoric acid solutions — aggressively attack most metallic current collectors and many electrode materials.

Third, self-discharge. Some reported designs lose 19–100% of stored capacity per day. That rate makes any practical application impossible without fundamental breakthroughs in electrode-electrolyte interface engineering.

Fourth, volumetric energy density. Even if gravimetric energy density approaches lithium-ion levels, the aqueous electrolyte and porous carbon electrodes produce physically larger cells — a disadvantage for space-constrained applications.

Fifth, definitional confusion. As the RMIT review documented, many papers published before 2020 under the “proton battery” label are more accurately described as zinc-based batteries with incidental proton involvement. The field is still consolidating its own identity.

Future Outlook — Where Proton Batteries Fit

The likely timeline, based on the historical trajectory of lithium-ion batteries (laboratory discovery in the 1970s, first commercial product in 1991), suggests proton batteries are currently at roughly the 1980-equivalent stage. Expect fundamental research to dominate through at least 2035, prototype scale-up attempts in the 2035–2040 window, and possible commercial entry after 2040 — if the cathode and corrosion challenges are solved.

When they do arrive, proton batteries will compete in large-scale stationary storage, where weight and volume are secondary to cost, safety, and resource availability. They will not compete in electric vehicles or portable electronics, where volumetric density is paramount. The most likely role is as a complement to lithium-ion and hydrogen — filling a middle ground where you need something safer than lithium-ion but more efficient than hydrogen.

Cross-Technology Comparison — Five Technologies at a Glance

The following table applies consistent metrics across all five technologies discussed in this article. Use it as a quick-reference scoring card. Just remember: the most important dimension — your project’s specific requirements — sits outside any table.

*Projected. All proton battery data from laboratory-scale devices.

Reading across the rows, one pattern becomes obvious: no single technology wins on every metric. Lithium-ion leads on efficiency and maturity but falters on long-duration cost scaling. Hydrogen leads on seasonal capability but loses two-thirds of input energy in conversion. Flywheels last forever but cannot store energy beyond minutes. Each technology is optimal somewhere — and suboptimal everywhere else.

More importantly for project developers, the industry is moving decisively toward hybrid architectures. A solar-plus-storage project in 2026 is increasingly likely to pair lithium-ion batteries for daily cycling with a long-duration technology — hydrogen or CAES — for multi-day backup, rather than attempting to stretch a single technology across incompatible requirements.

System integration becomes the critical success factor at this point. A project that stitches together components from separate battery, hydrogen, and power conversion vendors takes on integration risk at every interface — communication protocols, thermal management coordination, warranty boundary disputes. Working with manufacturers who offer integrated solutions spanning multiple technologies from a single engineering team avoids this fragmentation. BENY, a manufacturer with deployed projects across 70+ countries, offers integrated PV, battery energy storage, and EV charging solutions designed and tested as a unified system. One documented deployment in Thailand combines 240 kW MPPT solar, 100 kW / 241 kWh BESS, and 240 kW DC fast charging for an electric truck fleet (BENY integrated energy storage and EV charging solutions). When your project spans multiple energy storage technologies, integration architecture matters as much as component specifications.

Choosing the Right Energy Storage — A Project Decision Framework

Before diving into duration-based recommendations, anchor your evaluation on three questions. How long do you need to store energy? Is your budget driven by upfront capital cost or total lifecycle cost? Do you have space or geographic constraints that rule out certain technologies? Answer these three, and the technology shortlist largely writes itself.

Short-Duration Applications (Seconds to 6 Hours) — Batteries Are the Default

For grid frequency regulation, daily solar peak shaving, data center UPS, and commercial electricity arbitrage, lithium-ion batteries are the default choice — for good reason. Their combination of high efficiency, mature supply chains, and competitive pricing in the 1- to 4-hour range is unmatched. The decision is less about “which technology” and more about “which battery chemistry and supplier.”

Flywheels earn consideration in one specific sub-scenario: when you need sub-10-millisecond response for power quality. Think data centers, semiconductor fabs, and industrial processes where a voltage sag lasting milliseconds can scrap an entire production batch. In these cases, a hybrid flywheel-plus-battery configuration provides the best of both worlds. Flywheels catch the microseconds, batteries cover the minutes.

Key selection criteria: cycle life rating at your expected depth of discharge, warranty terms tied to throughput (MWh) rather than calendar years, and local service availability.

Medium-Duration Applications (6 to 24 Hours) — The Technology Battlefield

Six to 24 hours is the most contested duration range in energy storage. Your choice depends heavily on site-specific factors.

Lithium-ion batteries can stretch to 8 hours, but the economics deteriorate quickly. Extending from 4 to 8 hours roughly doubles the per-kWh capital cost because you add more cells without adding more power conversion equipment. CAES, by contrast, sees its unit costs fall with each additional hour of storage — the underground cavern (the expensive part) is already built. For durations beyond 8 hours and projects with suitable geology, CAES becomes the cost leader.

Sand batteries enter the conversation when the end use is heat. If your facility needs industrial steam, district heating, or hot water — and you have access to cheap renewable electricity for charging — sand batteries offer the lowest per-kWh storage cost of any technology in this comparison. The storage medium is sand.

Decision framework: CapEx-constrained and duration over 8 hours → evaluate CAES. Heat is the primary output → evaluate sand batteries. RTE is the primary metric and duration is under 8 hours → batteries remain the answer. Geographic restrictions (no suitable geology) → eliminate CAES.

Long-Duration and Seasonal Storage (Days to Months) — Hydrogen and Thermal Lead

This is energy storage’s final frontier. It is the one that determines whether a 100% renewable grid is feasible. During a Northern European winter, solar output can drop to 30–40% of summer levels for months at a time. No battery, no flywheel, and no CAES system can bridge that gap economically. Hydrogen can.

The concept is straightforward: during seasons of renewable surplus, excess electricity produces green hydrogen via electrolysis. That hydrogen is stored in salt caverns — at roughly $1 per kWh, the cheapest large-scale energy storage medium available — and converted back to electricity during renewable droughts. The LDES Council projects global long-duration energy storage deployment of 1.5–2.5 TW by 2040, equivalent to 85–140 TWh of storage capacity.

Sand batteries address the thermal side of the same seasonal challenge. In Finland, where district heating networks serve over half of all buildings, storing summer solar energy as heat in sand silos for winter use is not a research concept. It is an operational reality at the 100 MWh Pornainen facility.

A practical architecture for grid-scale renewable integration is emerging: batteries for intra-day balancing, CAES for multi-day gaps, and hydrogen for seasonal reserves. Think of it as a water management system: batteries are the rooftop tanks (small, fast, used daily), CAES is the municipal reservoir (larger, slower, used weekly), and hydrogen is the strategic reserve (massive, rarely tapped, essential when needed). Each layer serves a different timescale. All three together make a 100% renewable grid technically credible.

References

1. Rezaei Niya, Heidari, Andrews. “The role of proton battery technologies in future global energy storage.” Journal of Physics D: Applied Physics, 2024. https://iopscience.iop.org/article/10.1088/1361-6463/ad809c
2. BloombergNEF. “Long-Duration Energy Storage Cost Survey.” 2024. Via PV Magazine Australia: https://www.pv-magazine-australia.com/2024/06/04/
3. DOAJ. “Comparative Techno-Economic and Life Cycle Assessment of Stationary Energy Storage Systems.” 2025. https://doaj.org/article/20c51070a1df470a93bfefb58060cede
4. Polar Night Energy / PV Magazine India. “Finnish startup launches its first industrial-scale sand-based heat storage system.” 2026. https://www.pv-magazine-india.com/2026/02/03/
5. Global Energy Association. “Finnish scientists create sand battery for energy storage.” 2026. https://globalenergyprize.org/en/2026/05/08/
6. Cleantechnica. “New Compressed Air Energy Storage Systems Vs. Li-ion Batteries.” 2024. https://cleantechnica.com/2024/06/03/
7. ScienceDirect. “4E performance evaluation of renewable microgrids: Comparing hydrogen and battery storage.” 2025. https://www.sciencedirect.com/science/article/abs/pii/S0196890425002341
8. BENY. “Integrated Energy Storage and EV Charging Solutions.” https://www.beny.com/by-solution/
9. BENY. “Solar-Storage-EV Charging Thailand Project.” https://www.beny.com/new/beny-solar-storage-ev-charging-thailand/
10. BENY. Contact. https://www.beny.com/contact-us/