SWITCHED ON
The daily technology series nobody asked for but everyone needed
The Storage Bottleneck: Grid-Scale Batteries and the Real Constraint on a Renewable Future
Solar and wind are now the cheapest new electricity in most of the world. That fact alone has not decarbonised the grid, because cheap generation that switches off when the sun sets or the wind stops is only half the equation. Storage is the unglamorous, unsolved half.
Every episode of this series that touched climate technology arrived, eventually, at the same conclusion: generation is no longer the bottleneck. Storage is. A grid running on abundant, cheap solar and wind that has nowhere to put the excess energy generated at midday and no reserves to draw on at night is not a decarbonised grid. It is a grid waiting for a technology that does not yet exist at the scale required.
Yesterday we looked at the technology of an aging world — Japan's care robotics as the global leading-edge case study, what functional assistive devices actually deliver versus companion robots' more limited evidence base, the loneliness problem technology cannot solve, and the honest conclusion that robotics cannot substitute for the pension and healthcare policy reckoning that demographic change actually requires. Today we are closing a loop that has run underneath this entire season and much of Season One: energy storage, the unglamorous technology that determines whether cheap renewable generation actually translates into a reliable, decarbonised electricity grid, or remains a partial answer that still requires fossil fuel backup for the hours when the sun is not shining and the wind is not blowing.
01 — Why Generation Was the Easy Part
The cost of solar and wind electricity generation has fallen so dramatically over the past fifteen years — solar costs have fallen by roughly ninety percent, wind by roughly seventy percent — that both are now the cheapest source of new electricity generation in most parts of the world, a genuinely remarkable achievement in industrial cost reduction that deserves the recognition it has received. What this achievement obscures is that generation cost is only one half of what a functioning electricity grid requires. The other half is matching supply to demand continuously, in real time, twenty-four hours a day, because unlike almost any other commodity, electricity that is generated and not immediately used or stored is simply lost.
Conventional grids, built around fossil fuel and nuclear generation that can be dispatched on demand — turned up or down to match consumption patterns — did not need to solve this matching problem through storage, because the generation itself could be controlled to follow demand. Solar and wind generation follow weather and daylight, not demand. A grid with a high proportion of renewable generation needs either storage to bank excess generation for use when renewable output is low, or backup dispatchable generation (typically natural gas) to fill the gaps — and the second option, while currently the more common practical solution, defeats much of the climate purpose of the renewable transition in the first place.
California, which has achieved some of the highest renewable penetration of any major grid, has repeatedly experienced the "duck curve" phenomenon — a dramatic oversupply of solar generation in the middle of the day, sometimes so severe that electricity prices go negative, followed by a steep evening demand ramp as solar output falls just as people return home and turn on appliances. This is the storage problem made visible on a chart.
02 — Lithium-Ion at Grid Scale: What It Does and Doesn't Solve
Lithium-ion battery storage, the same fundamental technology covered in our electric vehicle episode, has become the dominant grid-scale storage technology deployed to date, benefiting from the massive cost reductions driven by EV manufacturing scale. Grid-scale battery storage capacity has grown extraordinarily rapidly — global installations have roughly tripled over the past three years, with major deployments including Tesla's Moss Landing facility in California (one of the largest battery storage installations in the world) and comparable large-scale projects across Australia, the UK, and China.
Lithium-ion grid storage excels at short-duration applications — smoothing out fluctuations over minutes to a few hours, providing frequency regulation services that keep grid voltage and frequency stable, and shifting solar generation from midday oversupply to the evening demand peak, directly addressing the duck curve problem. What lithium-ion is poorly suited to, given current costs, is long-duration storage — banking energy across multiple days, or across seasons, to handle extended periods of low renewable output (a windless week, a cloudy stretch, or the seasonal reduction in solar output during winter in temperate climates). The cost of lithium-ion storage scales roughly linearly with the duration of storage required, meaning that storing enough energy to cover several days, let alone weeks or a full season, becomes prohibitively expensive using current battery chemistry at anything approaching the scale a national grid would require.
03 — The Long-Duration Storage Race
The technology gap for long-duration storage — durations of eight hours to multiple days, and eventually seasonal storage — has motivated a wave of alternative storage chemistry and mechanical storage approaches, several of which have moved from research to early commercial deployment over the past few years.
Iron-air batteries, developed by companies including Form Energy, use a reversible rusting reaction — iron rusting to absorb energy during charging, and un-rusting (converting rust back to iron while releasing oxygen) during discharge — using abundant, cheap materials rather than lithium and cobalt. The technology has significantly lower energy density than lithium-ion, meaning iron-air batteries take up considerably more physical space for the same energy storage, but the low cost of iron as a raw material makes it potentially far cheaper per unit of storage for long-duration applications where space is less constrained than cost. Form Energy's first commercial-scale iron-air deployment, a 100-hour-duration storage facility in Minnesota, began construction in 2023, representing one of the first serious commercial tests of the technology at meaningful scale.
Pumped hydro storage — using excess electricity to pump water uphill into a reservoir, then releasing it through turbines to generate electricity when needed — remains, despite receiving far less attention than battery technology, the dominant form of grid-scale energy storage currently deployed worldwide by a wide margin, accounting for the large majority of all grid storage capacity globally. It is mature, reliable, and can be built at durations from hours to days, but is geographically constrained to locations with suitable topography (elevation difference and water availability), meaning it cannot be deployed everywhere a grid needs storage. Several countries, including China and the US, have significant pumped hydro expansion underway, alongside emerging variations including underground and closed-loop pumped hydro systems designed to reduce the geographic constraints of traditional reservoir-based systems.
04 — Thermal, Chemical, and Mechanical Alternatives
Beyond batteries and pumped hydro, a range of alternative storage approaches are being developed and deployed at pilot to early-commercial scale, each targeting different points on the cost-duration-location trade-off that no single storage technology fully solves.
Thermal storage — using excess electricity to heat materials like molten salt, rocks, or even sand to very high temperatures, then extracting that heat later either directly for industrial processes or converting it back to electricity — has the advantage of using cheap, abundant materials, though the round-trip efficiency (the proportion of energy recovered relative to what was stored) is typically lower than battery storage. Companies including Malta Inc. and various molten salt storage developers have deployed pilot-scale systems, with the technology particularly well suited to providing storage for industrial heat applications directly, avoiding the efficiency losses of converting stored thermal energy back to electricity.
Green hydrogen, which we examined critically in Season One's climate technology episode, has a genuine role in long-duration and seasonal storage specifically — despite its poor round-trip efficiency for general electricity storage, its ability to be stored for extended periods (weeks to months) in existing gas storage infrastructure gives it a niche advantage for the specific problem of seasonal storage that battery technology, even long-duration variants, struggles to address economically. Compressed air energy storage, using excess electricity to compress air into underground caverns or storage vessels and releasing it through turbines when needed, has a smaller deployment footprint globally but offers another long-duration option with a proven, if limited, commercial track record extending back decades.
05 — Why This Determines Everything Else
The storage bottleneck is not merely a technical curiosity. It is the load-bearing constraint on the pace of the entire climate technology transition covered across this series — the nuclear renaissance we opened Season Two with is motivated substantially by the recognition that storage technology may not solve the full dispatchability gap fast enough, making dispatchable low-carbon generation (nuclear) valuable as a complement to storage rather than a competitor to it. The grid infrastructure investment gap covered in our very first climate episode of Season One is, in significant part, an investment gap in the transmission and storage infrastructure required to make renewable generation's abundant but variable output usable.
The honest assessment, consistent with the pattern this series has documented repeatedly: the underlying physics and chemistry of long-duration storage are genuinely difficult problems, considerably harder than the generation cost reduction that solar and wind have already achieved, and the technologies being developed to solve them — iron-air, advanced pumped hydro, thermal storage, and others — are real, promising, and not yet deployed at anything close to the scale required for the fully renewable, fully storage-backed grid that the climate math ultimately requires. Getting from where we are now to that grid is not primarily a question of whether the technology will eventually exist. Several credible technology pathways already do. It is a question of whether the capital investment, manufacturing scale-up, and grid integration can happen fast enough to matter for the climate timeline the physics is imposing on us regardless of how quickly we move.
Tomorrow we close out Season Two's thirty-episode run with something we have not done since the Season One finale — a direct, personal reflection on what an additional fifteen episodes, spanning nuclear reactors to grid batteries, geopolitics to elder care, taught us about the shape of the next decade. See you then, for the last time this season.
Switched On is a daily technology series covering the ideas, systems, and arguments shaping the digital world. Opinionated. Witty. Occasionally wrong. Always worth the argument.
