A forecaster expects prices to keep falling
BloombergNEF, a market research firm whose annual battery price survey this site has already cited twice, states plainly that it expects average battery pack prices to fall again in 2026. It names 2 forces pulling in opposite directions. Rising raw material prices push costs up. The continued spread of a cheaper battery chemistry called lithium iron phosphate, or LFP, pushes costs down. BloombergNEF forecasts that the second force wins again. That call follows an 8% fall in the worldwide average battery pack price over 2025, to a record 108 dollars a kilowatt hour, a unit that measures how much energy a battery can store. No source read for this article gives a dollar figure for 2026, only the direction.
Factories outside China are catching up, at a higher cost
Global lithium ion battery manufacturing capacity passed 4 terawatt hours by the end of 2025, according to the International Energy Agency, the intergovernmental body that tracks world energy trends. That capacity grew about 30% in a single year. Over 80% of it still sits in China. The European Union and the United States each hold only 6% to 7% of world capacity, but each grew its own capacity by about 50% in the same year, faster than the growth inside China itself, just over 25%. The International Energy Agency states that building that capacity in Europe and the United States costs as much as 50% more than in China, before counting any public subsidy.
International Energy Agency figures. The growth inside China is stated as just over 25%, and the European Union and United States growth is stated as about 50% for both regions jointly. All 4 figures are approximate.
Show the numbers
| Global | 30 |
| China | 25 |
| European Union | 50 |
| United States | 50 |
The material makers behind the cells are losing money
That cost gap sits on top of a second problem, inside the supply chain that feeds every battery maker. The International Energy Agency states that many producers of cathode active material, the compound inside a battery cell that determines how much energy it stores, are operating at a loss while still expanding their own capacity. The agency says that combination raises the risk of the industry consolidating, meaning weaker producers close or get bought and fewer companies are left to set the price. The agency does not say prices will rise or by how much. It names a structural risk facing prices right now, not a forecast of its own.
Demand for raw materials is growing faster than supply
The same agency, in a separate report on critical minerals, states that demand for lithium alone grew about 25% a year on average over the past 2 years. Total battery demand grew over 35% in 2025 alone, passing 1.5 terawatt hours for the first time. Both figures run far ahead of the roughly 10% a year growth in demand for energy minerals generally. The energy sector now drives around 75% of that overall mineral demand growth, up from 70% in 2024. Demand growing this much faster than the minerals used to make it is the reason a raw material price rise was plausible in the first place.
International Energy Agency figures. The minerals and lithium rows are average annual growth rates over recent years. The battery demand row is a single year, 2025 alone, not an annual average, so it is not directly comparable to the other 2 bars.
Show the numbers
| Key energy minerals overall | 10 |
| Lithium | 25 |
| Battery demand, 2025 alone | 35 |
Cobalt is the clearest sign of the pressure
Cobalt, a metal used in some battery chemistries, priced at 56,290 dollars a tonne on 28 August 2026, according to Trading Economics, a live commodity data service, a rise of 69% over the year before. The International Energy Agency had already named the same metal as a risk in a report published months earlier. Lithium, the mineral behind the price spike this site has already reported, had not eased either, as of the most current reading available. Cobalt and lithium together are the raw material pressure BloombergNEF names as the reason its own forecast could be wrong.
The about a year earlier figure is this article own arithmetic, 56,290 divided by 1 plus 68.86%, and is not a number Trading Economics states directly. The 28 August 2026 figure is read straight from the source.
Show the numbers
| About 1 year earlier | 33,335 |
| 28 August 2026 | 56,290 |