Researchers at Chalmers University of Technology in Gothenburg, Sweden have developed an AI-powered fast-charging strategy that extends lithium-ion EV battery life by up to 22.9% — all without adding a single meaningful second to charging time. Published in the prestigious IEEE Transactions on Transportation Electrification, the breakthrough uses reinforcement learning to adapt charging current in real time to each battery's unique chemistry and condition, addressing one of the most persistent friction points in electric vehicle adoption.
The Problem: Why Fast Charging Kills Batteries
Electric vehicles have removed many of the maintenance headaches associated with combustion engines — no oil changes, fuel filters, or spark plugs. But they have introduced a new and pressing concern: the long-term health of the battery pack. Lithium-ion batteries degrade naturally over time, and frequent fast charging accelerates that degradation significantly.
The most damaging consequence is a phenomenon known as lithium plating — where lithium ions, under the stress of high-powered charging, fail to intercalate correctly into the anode and instead deposit as metallic lithium on the electrode surface. This progressively reduces capacity and, in severe cases, can cause a short circuit. Crucially, the risk intensifies as a battery ages — yet today's standard charging protocols apply the same current and voltage to a brand-new pack and a five-year-old one alike. The system simply does not adapt.
A 2024 study by Geotab estimates average EV battery degradation at around 1.8% per year, supporting a projected lifespan of at least 20 years or 200,000 miles for many vehicles. Some estimates place Tesla battery longevity at 300,000 to 500,000 miles depending on use and charging habits. The Chalmers breakthrough could add meaningful headroom on top of those already impressive figures.
The Solution: Reinforcement Learning That Learns Every Charge
The Chalmers team embedded a reinforcement learning model directly into the battery management system. Unlike conventional machine learning, reinforcement learning improves through continuous trial and error — identifying optimal outcomes by refining its approach with each new charging session. The AI takes two real-time inputs: the battery's instantaneous state of charge and its accumulated state of health, then dynamically adjusts the charging current accordingly.
As a battery ages, the system continuously recalibrates voltage to minimise stress on the anode, cathode, and electrolyte. The result, as measured in the IEEE study, is a battery lifespan extended to 703 equivalent full cycles to 80% of original capacity — a 22.9% improvement over the standard baseline. For drivers who regularly rely on fast charging, this translates to an estimated 70,000 to over 100,000 additional miles of useful battery life.
Critically, charging time is affected by only a matter of seconds — a negligible trade-off that preserves the fast-charging convenience modern EV drivers expect. The model was trained on a digital twin of a commonly available EV cell type, simulating the parameters that influence both charging speed and long-term battery health.
"Our study shows that smart adaptation of the current during charging, taking into account the changing electrochemical state of the battery, can maximise both the performance and the life of the battery."— Changfu Zou, Professor, Department of Electrical Engineering, Chalmers University of Technology
Why This Matters for the EV Industry
Battery longevity is one of the most consequential factors shaping consumer confidence in electric vehicles. Range anxiety has dominated the EV conversation for years, but battery degradation anxiety is an equally real barrier — particularly for fleet operators, taxi companies, and long-distance drivers who rely on fast charging daily and cannot afford accelerated wear.
Current EV batteries are estimated to last between 8 and 15 years depending on usage patterns. A near-23% extension in cycle life — achievable without any reduction in charging speed — would represent a material improvement in both the economics and sustainability of EV ownership. Fewer battery replacements means reduced costs for drivers, lower demand for raw materials such as lithium, cobalt, and nickel, and a meaningful reduction in the environmental footprint of electric transport at scale.
The researchers have expressed ambition beyond passenger vehicles. The technology holds particular relevance for heavy industrial vehicles, commercial fleets, and public transport — sectors where fast charging is not a convenience but an operational necessity, and where battery replacement carries significant financial and logistical weight.
"We show that it is possible to charge more or less as fast as today, but with significantly less long-term degradation of the battery."— Meng Yuan, Assistant Professor, Victoria University of Wellington & Former Researcher, Chalmers University of Technology