How are batteries improving so quickly?

Posted: March 17, 2026

In 2010—the year the first mass-market electric vehicle went on sale—the average battery pack cost $1,474 per kilowatt-hour. Since then, prices have fallen 93%, to $108/kWh, and are projected to keep falling.^[1] Real-world performance, meanwhile, has increased roughly fivefold: the Nissan Leaf, that 2010 trailblazer, had an advertised range of 100 miles (160 km); the latest EV models claim over 500 miles in range.

Although these improvements are gigantic, few are likely to be surprised by them. We live in a world of Moore’s Law and surging solar panel deployment —it’s practically a given that incremental, year-on-year improvements will occur, compounding over time to yield a collapse in price and an explosion in performance.

Yet, at the risk of stating the obvious, prices don’t fall of their own accord; nor does performance automatically improve.

So what’s happened—and happening—to the battery industry that’s prompting such rapid and sustained improvements? The answer can be separated into three parts: market dynamics, the role of software in developing battery chemistries, and improvements to manufacturing processes.

Market dynamics: competition and overcapacity

A simplistic explanation for the improvements in battery quality can be expressed in just a few words, courtesy of BloombergNEF’s Evelina Stoikou: “Cutthroat competition is making batteries cheaper every year.” Indeed, according to BNEF battery prices fell 8% in 2025—despite a rise in the price of metals.

Competition is most intense in China, where the government’s industrial policy has considered the battery industry of strategic importance ever since the publication of the 10^th Five-Year Plan, in 2000.

Over the intervening quarter-century, government support helped Chinese companies like CATL and BYD to dominate the battery market, but it has also resulted in overinvestment. The world deployed roughly 1.6 TWh’s worth of batteries in 2025; China alone produced 1.98 TWh.^[2]

The country’s tremendous manufacturing capacity means battery prices are markedly cheaper in China than in the U.S. (+44%) and Europe (+56%). Chinese battery companies also have lower margins than overseas companies.

In response to this state of over-competition—or “involution,” as the Chinese Communist Party calls it—in 2024, the Chinese government introduced regulations designed to reduce the capital spending that is driving overcapacity. These regulations include more stringent battery safety requirements and a minimum spend on R&D by battery companies. The latter stipulation appears unnecessary: Chinese firms are already spending roughly double the mandated 3% of revenue on research and development.

AI is accelerating battery chemistry development

Capital expenditure may be driving falling prices, but it’s the R&D spending of battery companies that is behind improved performance.

How a battery performs is largely a function of the compositional mixes of its cathode, anode and electrolyte. Collectively, these are referred to as the chemistry of the battery. Innumerable tweaks to chemistries have resulted in the current prevalence of the lithium iron phosphate (LFP) chemistry. Thanks to their low cost and high stability, LFP batteries now account for “nearly 50% of EV battery sales and dominate the stationary storage market,” according to the Volta Foundation.^[3]

LFP’s dominance will not necessarily continue. For all their benefits, LFP batteries are less energy dense than nickel/cobalt alternatives; they also perform worse in cold conditions.

More to the point, recent developments in software and AI promise to accelerate the discovery of new or improved chemistries. Much of the current AI-powered research centers on better understanding the micro-structures of battery materials and optimizing manufacturing processes accordingly.

The U.K. startup Polaron has a platform for analyzing micro-structures and generating new material designs. It outlined the logic of a focus on micro-structures in a February  press release : "Processing determines structure, and structure determines performance. The arrangement of grains, pores, phases, and defects inside a material governs properties such as strength, lifetime, and failure."

Polaron is using microstructural images to predict how materials will behave, thereby "supporting cleaner, more efficient manufacturing at scale."

AI platforms like Polaron’s are present across the R&D labs of major battery players and tech companies. Results are already striking. In 2024, for example, LG Energy claimed it had used AI to compress the cell design process, which usually takes weeks, into a single day.

In August 2025, researchers at IBM published a paper outlining how they trained a “chemical foundation model” to identify a new, more conductive electrolyte formulation. The team has since partnered with Mercedes Benz’s research arm.  “Collaborating on the research and development process will allow us to validate the AI-workflow, the use of quantum computers, and other tools for the discovery of sustainable battery materials,” says IBM.

With AI-powered material design now in full swing, it’s quite possible that the huge performance improvements of the last couple of decades will come to seem minor.

Manufacturing improvements and dry electrode coating

Alongside improvements in chemistry, battery manufacturing efficiency has also improved dramatically. Some of the gains, as enumerated in Brian Potter’s Construction Physics, were obtained from larger, faster machines, ever-thinner battery casings and new double-sided electrode coating machines.

The industry is also exploring how to combine CT scanning and AI—which are already used together in the R&D phase of battery production—into a next-gen quality-control method.^[4]

As with chemistries, though, there remains obvious room for improvement in manufacturing processes—most notably in the production of battery cathodes and anodes.

At present, electrodes of all chemistries are produced by mixing active materials, conductive additives, and solvents into a liquid slurry. After the slurry is spread onto the current-collecting foil, the solvent must be evaporated, which involves passing the slurry through drying ovens.

Dry electrode coating eschews solvents altogether, instead combining electrode powders to form a film that can be laminated onto the foil. Dry electrode coating would not require drying ovens, thereby shrinking the footprint and cost of the production process. By foregoing the use of highly toxic solvents, dry coating could also make the permitting of new battery factories much more straightforward, as Benchmark Minerals notes.

On February 1, Elon Musk claimed that Tesla had made the dry electrode process “work at scale”—although it’s unclear from Musk’s statement exactly how close Tesla is to deploying the technology in its battery factories. LG Energy, another leading battery manufacturer, is aiming to have dry electrode coating in its factories by 2028.

If and when dry electrode coating is achieved, it would represent a major manufacturing advance.

What better batteries enable: EVs, BESS, and robotics

Advances in battery technology are having, and will have, major impacts on other industries.

Despite slowing momentum and increasingly unfavorable policy environments, EV sales continue to grow.^[5] The ultra-stable and low-cost LFP chemistry is being used in battery electric stationary storage plants, which are crucial for grid stability in the age of renewable generation. And batteries look set to play a crucial role in the emerging robotics industry.

Further reading: Much of the information in this piece comes from the Volta Foundation’s annual Battery Report—an excellent and completely free resource that readers are encouraged to consult for themselves.

^{[1] According to BloombergNEF.
[2] Deployments: International Energy Agency. Production: Volta Foundation, Battery Report, 2025, p. 86.
[3] ibid., p. 6.
[4] ibid., p. 111. 
[5] ibid., p. 6.}