Long-Duration Storage Needs a Procurement Model, Not Just Better Technology

Topic: Date: Reads: 158

Long-duration energy storage has an appealing story. It promises to cover the hours and days when solar, wind and short-duration batteries are not enough. It could reduce curtailment, support reliability and make high-renewable power systems more resilient. The difficulty is that a useful story is not the same as a bankable market. Long-duration storage does not only need better technology. It needs procurement models that pay for the specific reliability services it provides.

Most battery markets developed around shorter services: frequency response, reserve capacity, intraday shifting and price arbitrage. These services fit lithium-ion batteries well because they require fast response and relatively limited discharge duration. Long-duration storage is different. Its value may appear during rare weather patterns, multi-day renewable shortfalls, seasonal stress or transmission outages. If the market only pays for daily spreads, technologies designed for longer discharge may look uneconomic even when the system needs them.

This creates a timing problem. Grid planners can see that deeper renewable penetration will increase the value of longer flexibility, but developers need revenue today. A project cannot be financed on the argument that it might be useful in 2032 unless contracts, capacity payments or regulated cost recovery bridge the gap. Without a procurement model, the industry risks a cycle of promising pilots that do not turn into repeatable deployment.

A good procurement model should define the service before choosing the technology. Does the system need eight-hour storage, multi-day backup, seasonal balancing, black-start capability, congestion relief or clean capacity during winter evenings? Each requirement points to different options, including flow batteries, compressed air, thermal storage, pumped hydro, hydrogen-based storage or hybrid renewable portfolios. Technology neutrality is useful only if the performance obligation is clear.

Risk allocation is equally important. If a storage project is paid only when scarcity prices spike, investors may demand high returns because revenue is uncertain. If a utility signs a long-term availability contract, customers may carry some cost even in years when the asset is rarely used. Neither model is automatically wrong. The question is whether the payment reflects the insurance value of reliability. Power systems pay for backup precisely because the worst hours matter more than average hours.

Long-duration storage should also be evaluated against alternatives. Transmission expansion, demand response, flexible thermal plants, interregional trade and overbuilding renewables can all reduce reliability risk. In some grids, the cheapest solution may be a portfolio rather than a single storage technology. This makes planning more complex, but it also protects consumers from paying for fashionable assets that do not solve the binding constraint.

Procurement should also recognize that different grids need different durations. A sunny region with steep evening ramps may need four to ten hours of storage before it needs multi-day assets. A windy region with seasonal weather patterns may need longer coverage during extended calm periods. An island grid may value resilience and fuel import reduction more than a large interconnected market does. Treating long-duration storage as one category can lead to poor procurement. The requirement should specify the problem: how many hours, how often, under what weather conditions and with what response time.

Demonstration projects still have value, but they should be designed to answer commercial questions. Does the technology perform after repeated cycling? How does it degrade? Can it be permitted at useful scale? What maintenance skills are required? Does it interact well with market dispatch systems? A pilot that only proves that energy can be stored and released is not enough. Investors need evidence that the asset can live inside a real grid business model. Public funding is most useful when it reduces uncertainty that private capital cannot reasonably price on its own.

For buyers, the central question should be whether the storage asset lowers total system risk. A project with a high levelized cost may still be valuable if it avoids emergency fuel purchases, blackouts or expensive network upgrades. Conversely, a cheap storage project may add little if it discharges during hours that are already well served. Long-duration storage should be judged by avoided system costs, not by technology enthusiasm or by a single headline price.

This is where regulators can create confidence by procuring in tranches. Instead of betting everything on one technology or one long contract, they can buy defined reliability services in stages, observe performance and adjust future solicitations. That approach gives new technologies a route to scale while protecting customers from paying too much before the system need is proven.

The market should also avoid pretending that every reliability need must be solved by a new asset. Better forecasting, regional coordination and flexible demand can reduce the amount of storage required. That makes the storage that is built more valuable, because it is targeted at the hours and risks that cheaper measures cannot cover.

The storage debate needs to mature from technology enthusiasm to service design. Long-duration storage can be critical in cleaner grids, but only if markets specify what they need and pay for it in a durable way. Better chemistry will help. Better procurement may matter even more.

Sources reviewed