If pumped hydro represents the large-infrastructure route of physical energy storage, compressed air, flywheels, and thermal storage represent the scenario-specific route. These technologies may not all become national-scale mainstream storage solutions, but in specific applications, they can provide value that batteries cannot fully replace.
Compressed air energy storage is suitable for long-duration, large-capacity applications with clear geological conditions. It uses electricity to compress air, then releases the compressed air to drive turbines or power systems when electricity is needed. Traditional compressed air projects rely on salt caverns or underground reservoirs, while newer advanced compressed air technologies aim to reduce dependence on gas combustion and improve efficiency and zero-carbon performance. A 2025 study found that advanced adiabatic compressed air storage could have economic potential for 10- to 100-hour storage, especially in areas with favorable geological storage conditions.
Flywheel storage follows a completely different path. It is not designed for long-duration discharge, but for high power, fast response, and frequent cycling. Flywheels store kinetic energy in high-speed rotating masses and are suitable for grid frequency regulation, rail transit energy recovery, data center UPS, industrial power quality management, and transient support in microgrids. Their advantages are fast response, long cycle life, and high power density. Their limitation is short storage duration, making them better suited for power-type storage rather than energy-type storage.
Thermal storage has a different market logic from electricity storage. Its core value is not necessarily power-to-power conversion, but direct service to industrial heat, steam, heating, and high-temperature processes. Reuters reported in 2026 that industry consumes about 37% of global energy, most of it as heat. Thermal storage is gaining attention because of industrial decarbonization needs and the use of low-cost materials such as molten salt, sand, concrete, and graphite. This means the potential market for thermal storage is not limited to power grids. It extends to steel, chemicals, food processing, paper, building materials, district heating, and concentrating solar power.
The shared opportunity for these specialized physical storage technologies is to avoid competing directly with lithium batteries in their strongest short-duration applications and instead find value zones that match their own characteristics. Compressed air targets long-duration power balancing. Flywheels target transient power support. Thermal storage targets industrial heat substitution and low-cost long-duration storage. Gravity storage targets special terrain and low-safety-risk applications.
In the future, physical storage will not follow a single technology path. Multiple technologies will coexist. The real determinant of commercialization speed is not laboratory performance, but whether a project can answer three questions: Can it secure stable revenue? Can it gain acceptance from grids or industrial users? Can it be cheaper, safer, and more reliable than alternatives over its full life cycle?
