Spanish utility Iberdrola has commissioned the 225 MW / 210 GWh Valdecanas pumped hydro energy storage (PHES) project in Spain’s Extremadura region. For context, Spain’s average daily electricity generation is around 650 GWh, and 210 GWh of storage depth would cost $80 billion if built out in the form of utility-scale lithium-ion batteries.
Instead, the Valdecanas project’s battery component will be just 15 MW / 7.5 MWh, and the investment cost has been so small, due to using pre-existing reservoirs and only adding three 83 MW turbines plus pumping systems, that it’s gone unmentioned. We estimate this project’s capex at no more than $300 million.
The company has stated that its Conso II pumped hydro project in the Sil basin, Ourense, will have 1.8 GW capacity at a $1.6 billion price tag, uniting two reservoirs which have a 690-meter height differential. This facility is expected to generate 4,000 GWh a year, meaning it will be discharging about 25% of the time, but its storage depth will be a mammoth 58 GWh. That project will be constructed by 2031, but the company has a whole fleet of such projects operational (6 GW) or under development across Spain and Portugal.
Of course the comparison between 210 GWh of pumped hydro reservoir depth and $80 billion of batteries is completely unrealistic, since it would take 1,000 hours, which is to say six weeks, of consistent use in one direction to charge or discharge the reservoir using 225 MW of turbines – a timeframe which is only relevant to seasonal variation in energy supply.
But Spain does have a lot of seasonal variation in supply and demand, perhaps more than most countries. Spain’s nuclear fleet is set for phaseout from 2027 to 2035, it’s far enough from the equator that its solar output halves in midwinter compared to midsummer, and its considerable hydropower is also seasonal in terms of rainfall.
So Spain, which currently has considerable electricity price variation including negative pricing hours, has a clear solution to this problem – upgrading its existing hydropower, already built out to some 20 GW of capacity in a hilly country, with pumped hydro capability in order to integrate the nation’s rapidly growing renewable fleet.
To change topic somewhat, if pumped hydro can be cost-effectively developed in flat terrain (rather than being geographically restricted), which will depend on the nature of the ground being excavated, then pumped hydro will be able to provide full reliability to data centers, enabling them to cost-effectively run off renewable energy all year round.
That would enable data center hyperscalers to pursue a fully renewable, reliable and even cost-effective energy strategy, unlike their current strategy which relies on natural gas for reliability in the short term, followed by small modular reactors (SMRs) in the longer term – which will likely be expensive, and don’t even exist yet. Batteries, at least not today’s 4-hour duration lithium BESS, don’t provide the storage duration and depth to make a renewable energy portfolio fully reliable 24/7 and 365 days a year. Pumped hydro is needed to achieve that.
SMRs are guaranteed to happen, backed by China’s ambitions as well as the West’s Big Tech companies – but it will take 3-5 years for the first to appear. Don’t be surprised if pumped hydro also makes inroads into data center hyperscalers’ plans by then.