Five years ago the anxiety in the room was cobalt. The dominant cathode chemistry was NMC, most of the world's cobalt came from the Democratic Republic of Congo with its environmental, child labour and political corruption problems, and nobody was sure about nickel either.
Then, to an extent Adair Turner admits he did not see coming, a wave of innovation produced lithium iron phosphate, which needs neither. Cobalt and nickel prices, having risen three or fourfold between roughly 2019 and 2022, are back where they started. More than half of all batteries now produced in the world are LFP.
A complete change in five years, is how he puts it. And it is the second time he has watched this happen. Lithium prices rose about fourfold between 2020 and 2022 and then completely reversed, while battery energy storage was booming and global EV sales were running ahead of what anyone predicted five years ago.
Turner became the first chair of the UK's Climate Change Committee in 2008, a long enough baseline to be unsentimental about forecasts. Lithium-ion is down 90% or more in price since then. His conclusion: "every time we think there's a constraint, it turns out that it isn't an important one."
He is not blanket-optimistic. He is specifically worried about the local environmental effects of the large nickel developments in Indonesia, and accepts future pinch points are possible. But he expects supply to be answered first by new resources, then by less harmful processing, then by recycling, which within ten years he thinks will supply a significant proportion of the lithium, nickel and cobalt going into batteries. He also keeps the comparison in view: the environmental impact is two or three orders of magnitude below that of the fossil fuel system being replaced.
The economics stop batteries doing long duration, not the physics
A cell has a maximum C rate. It has no minimum. Nothing stops anyone charging a battery slowly over a week and running it down over the following week. The reason nobody does is that batteries have historically cost enough per kilowatt hour storable that they only pay if cycled many times a year, at least in the region of 365, filled by day and drained by night.
So the question is where the cost goes. In 2010, in today's money, a pack was 1,300 to 1,400 dollars per kilowatt hour. It is now under 100. Turner's threshold is 10. Give him a battery at ten dollars a kilowatt hour and he would use it for multi-day and even hundred-hour storage. For now the working assumption is that batteries handle two, four, eight, perhaps twelve hours.
That is enough to settle the diurnal problem outright, which is why the answer is regional. India, Indonesia and Africa face a challenge almost entirely about getting solar generated by day into an air conditioner running at night, with little seasonal variation on top. In the sunbelt, solar plus batteries becomes close to a complete solution.
Northwest Europe's hardest case is a different shape. A February anticyclone over the North Sea can suppress wind not for an hour or a day but for two or three weeks. It might happen once or twice in a winter and nobody knows when. If it is a cold one, it lands exactly when demand peaks, and it will land harder once gas boilers have been replaced by heat pumps.
Between the two ends sit compressed air, possibly liquid air, possibly vanadium redox flow, and pumped hydro wherever the elevation allows. Heat storage interests him too: warm bricks or stones with surplus electricity, then release the heat directly or use it to drive turbines.
For the seasonal gap, the Dunkelflaute, his answer is hydrogen made by electrolysis from surplus autumn wind and stored in salt caverns. UK analysis suggests plenty of cavern capacity at low cost, and the capital cost per kilowatt hour storable is low enough that cycling once or twice a year does not matter. The penalty is conversion: electrolysis losses plus combustion losses in a gas turbine come to something like 50%, so the electricity coming out is by definition twice as expensive as the electricity that went in.
Ninety terawatt hours parked on the street
Turner's fleet arithmetic is worth doing slowly. Assume 1.5 billion cars by 2050, all electric, averaging a 60 kilowatt hour pack. That is 90 terawatt hours of storage sitting on streets and in garages, against daily electricity demand of around 120 terawatt hours. Privately owned cars, as distinct from taxis and fleets, are in use about 5% of the hours in a year.
If that capacity could be used, it would cover the entire short-duration balancing requirement without a single additional stationary installation.
He is honest about the gap between the arithmetic and the reality. He first heard the vehicle-to-grid idea 10 or 15 years ago, and the sophisticated version, where a tariff lets the supplier push electricity in and pull it out under software control, has not arrived. The first step is duller and far more achievable: get a growing fleet to charge at a sensible time of day. Time-of-day pricing is the lever.
His example comes from Ausgrid, an Australian distribution company and an ETC member. There are parts of Australia where electricity is now free between 11am and 3pm, because there is so much rooftop solar that the grid does not want to pay for it, and customers are being told to charge the car in the window when electricity is being thrown away.
On the vehicles themselves he is dismissive of the standard objections. When British sceptics tell him batteries do not work in cold climates or over long distances, he asks them to explain Norway, where 95% of new sales are electric and mostly pure battery rather than plug-in hybrid. China is above 50% of new passenger car sales and he suspects 90% by 2030. Price parity with an equivalent combustion car is two or three years away, and already reached in China. Energy density is climbing 6 to 8% a year, which is what has put 350 and 400 mile ranges on the market.
Containers, brownfield land and the old coal connection
The cost story on stationary storage has a hardware explanation that gets less attention than the cell price.
Until about four years ago, stationary storage cost more per kilowatt hour than EV storage, which was odd given that EVs carry the harder packaging constraint. The reason was standardisation: automotive had large OEMs placing large standardised orders, and the stationary market did not. Containerisation closed the gap. Packs go into 20-foot equivalent units, moved around like Lego bricks, sold either as a DC block or as an AC block with the inverter integrated, and battery management system costs have come down with them.
Permitting is the part containerisation does not fix. Turner expects storage to draw less opposition than pylons, because it can be kept low, put on brownfield land, even sunk below sight lines, and it produces no effluent and no noise. Every installation still needs a grid connection, though, and wires attract objections.
Which produces the neatest suggestion in the conversation. Every closed coal plant is a brownfield site with a massive grid connection already attached. Put the battery installation there and the asset is reused rather than rebuilt. The other route is downward: storage in homes and on commercial sites, paired with solar, balances locally and reduces how much new grid gets built at all.
How far the UK got without any storage, and what has to come next
Here is the fact that should recalibrate anyone planning a build-out. UK electricity had a carbon intensity of 500 grams per kilowatt hour in 2010. In 2024 it was 125, a 75% reduction, achieved with very little battery investment and almost no long-duration storage.
That came from closing coal, using gas as the fossil fleet, and running that fleet in inverse to wind output, while wind and solar grew past 30% and towards 40% of supply. Turner treats this as the important message for India and China: a flexible fossil fleet takes you a long way, and while coal is inherently less flexible than gas, new technologies can make coal flexible too.
The trick stops working at the end. The UK is committed to zero or near-zero carbon electricity by the early 2030s, which means periods with no wind, no solar and no gas fleet running either. That is where the batteries, compressed air, pumped hydro and hydrogen have to appear, and the markets have to pay for them.
He declines to be prescriptive about how, and describes three structures that work. Quantitative mandates, where developing wind or solar obliges you to install a set amount of storage. India's round-the-clock renewables contracts, where the developer commits to delivering electricity for 80% of hours a year, and increasingly to specific half-hours nominated 24 hours ahead, which hands the wind-solar-battery optimisation to the developer along with the risk of buying very expensive wholesale power if they get it wrong. Those auctions are already showing wind, solar and batteries beating new coal in India across the year. And capacity markets, with one condition he is firm about: neutrality between paying a gas turbine to exist, paying a battery to exist and contracting demand-side flexibility. Some, he says, have historically favoured gas.
Who pays is mostly settled in his view: private developers, pension funds, insurers, the long-term infrastructure investors. State or infrastructure bank money belongs in first-of-a-kind complexity, salt caverns and hydrogen pipelines, where several unrelated parties have to move at once.
Sodium is an insurance policy, and a bet that settles in 2070
The sodium-ion case has been squeezed by its rival's success. Turner has been told sodium at mass scale could reach 20 to 25 dollars per kilowatt hour at cell level. People were saying that three years ago, when a lithium-ion cell was 80 dollars, and part of the reason it was 80 was that lithium cost four times what it does now. The cheapest lithium-ion cells are now 35 dollars. So sodium's advantage narrows, and its function becomes insurance: if lithium supply ever does bind and the price runs, sodium is immediately in the money.
Iron-air interests him at the opposite end of the duration curve. It runs on rusting and de-rusting, oxidation releasing energy and reduction absorbing it, with a large round-trip inefficiency and a very low capital cost, because the material is cheap and the process is straightforward. On cycles run a few times a year, low capital cost beats efficiency, which makes iron-air a possible alternative to hydrogen.
Then the prediction he offers as a wager. Batteries today are good for road transport and the very short end of aviation, a ten-seater going 500 kilometres, nothing resembling a jumbo crossing the Atlantic. But the theoretical energy density of lithium-air, or lithium-oxygen, is sufficient for the long-haul case, and density has been compounding for years. His bet is that by 2070 there will be long-distance battery aviation.
He attaches two conditions. Nothing in aviation policy before 2050 can rely on it, so decarbonising flight in the meantime still means sustainable aviation fuel and a liquid hydrocarbon.
The other is about settlement: "I will be dead then. So you ain't going to get your money back."
This piece draws on the full conversation, which is available with a complete transcript on the episode page.