Battery storage is often discussed as if every project is solving the same grid problem. It is not. A two-hour lithium-ion system, an eight-hour iron-air project, a pumped-hydro plant, and a seasonal hydrogen store can all be called storage, but they compete in different parts of the grid's time problem. The useful way to think about storage is the duration stack : how fast a resource responds, how long it can sustain output, how often it cycles, and what grid need it is actually designed to meet. AI-generated image A storage portfolio needs different technologies for minutes, hours, days, and seasonal risk. Duration is not the same as capacity Storage projects are usually described with two numbers: power and energy. Power, measured in megawatts, is how hard the system can charge or discharge at a moment. Energy, measured in megawatt-hours, is how much it can deliver over time. Duration is the ratio between the two. A 100 MW / 400 MWh project can discharge at full power for about four hours. A 100 MW / 1,000 MWh project can run for about ten hours. That simple math hides a strategic question. What problem is the grid asking the storage system to solve? Some needs last seconds or minutes, such as frequency response and contingency support. Others last a few evening hours after solar output falls. Some stretch through a winter storm, a cloudy week, a transmission outage, or a seasonal mismatch between renewable supply and demand. A storage technology that is excellent for one duration can be a poor fit for another. Lithium-ion batteries dominate short-duration grid storage because they are efficient, modular, fast, and bankable. But if the job is to cover multiple days, the cost of adding more energy capacity can become the limiting factor. That is where flow batteries, iron-air systems, thermal storage, compressed air, pumped hydro, hydrogen, and other long-duration approaches try to earn a role. Sec Frequency and stability response 4h Solar shifting and peak capacity 10h+ Longer renewable gaps Days Weather and resilience events The short-duration layer The first layer of the stack is fast response. Batteries can inject or absorb power almost instantly, which makes them useful for frequency regulation, contingency response, ramp control, and other grid services. In markets that pay for those services, a storage project can earn revenue without always discharging for hours. Short-duration batteries also reduce renewable curtailment and shift solar into the evening peak. This is the familiar four-hour battery story: charge during low-price or high-renewable periods, then discharge when demand rises and solar falls. In many regions, that use case is still valuable and will keep growing. But the more four-hour batteries a region adds, the more the value can shift. Early projects may capture strong evening spreads. Later projects can compress those spreads by all discharging into the same window. The grid then starts needing either more flexible dispatch, longer duration, or complementary resources that can cover a wider range of conditions. The practical definition A duration stack is a portfolio view of storage. It matches technologies to grid needs by response speed, discharge length, cycle frequency, efficiency, location, cost structure, and reliability value. The long-duration layer Long-duration energy storage usually refers to systems that can deliver power for at least ten hours, though definitions vary by market and policy program. The important point is not the exact cutoff. It is that longer-duration systems target different events than a typical short-duration lithium-ion battery. An eight-to-ten-hour system may help cover a long evening peak, a cloudy day, or an extended renewable lull. A multi-day system may help during extreme weather, fuel supply disruption, or transmission constraints. Seasonal storage is another problem again: moving energy from one part of the year to another, which usually demands cheap energy capacity and tolerance for low cycling frequency. Different technologies make different tradeoffs. Flow batteries can decouple power and energy more easily than lithium-ion, but face scale-up, supply-chain, and bankability hurdles. Iron-air systems aim for very low energy-capacity cost over multi-day durations, but need field proof at utility scale. Pumped hydro is mature and massive where geography and permitting allow it. Hydrogen can store large amounts of energy, but round-trip efficiency, infrastructure, and end-use choices matter. Efficiency is only one part of the comparison. A daily cycling battery needs high round-trip efficiency because losses show up constantly. A seasonal or emergency resource may cycle rarely, so low energy-capacity cost, siting, fuel availability, and readiness can matter more than efficiency alone. The right metric changes with the duration layer. Duration layer Typical grid job Example technologies Seconds to minutes Frequency, voltage, contingency response Lithium-ion batteries, advanced inverters 2-6 hours Solar shifting, evening peaks, capacity Lithium-ion BESS 8-24 hours Longer renewable gaps and reliability events Flow batteries, thermal storage, pumped hydro, compressed air Multi-day to seasonal Weather risk, resilience, seasonal balancing Iron-air, hydrogen, large hydro reservoirs, other emerging LDES Why markets struggle with the stack Electricity markets are good at paying for energy that arrives now. They are less consistent at paying for capability that may be needed during rare but high-impact events. A multi-day storage project can look expensive if it is evaluated only on daily price arbitrage. Its value is clearer when the grid needs backup through a storm, drought, heat wave, fuel shortage, or week of weak renewable output. That creates a financing problem. Short-duration batteries can point to established markets, proven equipment, and repeatable revenue models. Long-duration projects often need capacity contracts, clean-firm procurement, resilience payments, regulated utility ownership, demonstration funding, or customer-specific offtake. Without those structures, the technology may be useful to the grid but difficult to finance. Planning models also matter. If a regulator or utility studies only average days, long-duration storage can appear unnecessary. If the study includes weather-driven renewable droughts, electrified heating peaks, generator outages, and transmission limits, the need can change sharply. The duration stack is therefore a modeling issue as much as a hardware issue. This is why procurement language needs to be precise. A solicitation for capacity, flexibility, resilience, resource adequacy, or clean energy shifting can produce very different projects. Buyers that specify only megawatts and megawatt-hours may miss whether the asset can actually respond to the event they are trying to cover. What buyers should ask A buyer should start with the grid problem, not the technology label. Is the resource needed for fast response, peak shaving, daily solar shifting, backup during outages, clean-firm capacity, or multi-day resilience? How many cycles per year will it run? How often must it sit ready? What degradation, fuel, water, siting, permitting, and safety constraints apply? The next question is deliverability. A storage asset that cannot interconnect, charge from clean energy, discharge when needed, or earn revenue for its reliability value may not solve the intended problem. The best duration-stack planning treats storage as part of a system that includes transmission, demand response, generation, forecasting, grid-forming controls, and market design. The CurrentCells takeaway There is no single storage winner because the grid does not have a single storage problem. Lithium-ion batteries will keep dominating fast and short-duration applications. Long-duration technologies will matter where renewable penetration, reliability risk, data-center load, electrificati