Battery or pumped storage? Procure the duration before choosing the technology

Begin with the residual-load problem

An industrial buyer may need a short evening peak covered, a utility may need overnight energy, and a system operator may need rapid reserve. Calling all three “storage” does not make their requirements identical. Define the delivery profile, charging window, reliability standard and connection point before selecting a technology. Otherwise, the comparison can reward a product that does not solve the actual problem.

CEA’s 2026 planning case envisages both batteries and pumped storage by 2035-36, with different power and energy ratings.[S1] That portfolio approach is a useful starting point: system needs are heterogeneous. The official figures are planning outcomes rather than a project pipeline, and they do not establish which technology wins at a particular site.

Equal MW does not mean equal energy

The eight-hour product needs four times as much delivered energy as the two-hour product, despite the same 100 MW rating. Its charging window must also be physically feasible. Charging 1,000 MW h in ten hours requires 100 MW average input; in six hours it requires approximately 167 MW before other constraints. Connection capacity and hourly energy availability can therefore invalidate an apparently attractive duration offer.

For batteries, confirm whether duration is guaranteed at beginning or end of life and whether augmentation is included. For pumped storage, confirm usable water storage, head, operating restrictions and the guaranteed pumping/generation envelope. Neither an energy nameplate nor reservoir volume alone proves net delivered MW h under the contractual duty cycle.

Separate operating fit from execution risk

The matrix is not a scorecard with preset winners. Weight its rows according to the buyer’s need. A short critical peak with limited land may lead to a different solution from repeated overnight energy at a connection-ready hydro site. Execution risk should be priced through probability-weighted delays and contingency costs, rather than buried in a qualitative footnote.

A technology with lower expected operating cost can still be the wrong near-term choice if its delivery date cannot meet the requirement. Conversely, speedy deployment should not justify ignoring replacement costs or long-term performance. A staged portfolio may be preferable if different assets become ready at different times.

Compare equivalent cash flows

For a storage-service contract, convert capacity payments into cost per net discharged kWh using realistic cycles and duration. Add charging energy, losses, network charges and excluded items. For an owned asset, use discounted capital, operating, augmentation and terminal cash flows against discounted net output. Do not compare an energy-inclusive pumped-storage quote with an energy-exclusive battery capacity payment.

Government’s November 2026 VGF announcement covers both BESS and PSP categories.[S3] Eligibility and disbursement timing must be verified for each project. Support for a particular procurer or asset does not imply that every competing offer has the same subsidy basis. Create a clear pre-support and post-support comparison, and show who retains the benefit.

Availability and use must also be separated. A long-duration asset reserved for rare contingencies can have a low annual energy denominator while providing valuable reliability. In that case, compare the cost of equivalent dependable capacity and energy security, not only average LCOS. Conversely, an energy-shifting project should demonstrate repeated economically useful dispatch.

Use triggers to choose a portfolio

Screen a 100 MW need under several residual-load profiles: two-hour spikes, four-hour evening deficits and extended overnight gaps. Test adverse charging prices, lower cycling and delayed commissioning for each candidate. Assign the actual contractual cost of non-delivery, not an arbitrary national average. Rank solutions by the cost of meeting the same obligation and the risks remaining with the buyer.

A strong procurement request asks bidders for a machine-readable hourly dispatch and cost model, guaranteed performance at the specified boundary and a documented readiness pack. It also asks how the offer changes if duration or annual cycles change. That exposes the cost of optionality and can make a mixed portfolio or staggered commitment more attractive than one all-purpose project.

Before final approval, run one joint downside case rather than only separate sensitivities. Combine a delayed connection, expensive charging hours and lower utilisation. These risks can occur together. The comparison should reveal the additional funding requirement and the replacement-supply exposure during the delay, so the procurement committee understands both investment cost and business continuity.