Recurrent Auto's EV Battery Health Study: Real Degradation Data Across Major Models

Recurrent Auto has published the most comprehensive real-world battery degradation dataset accessible to EV consumers. The Seattle-based company aggregates continuous battery health readings from owner-authorized API connections and OBD-II readers across major EV brands — tracking estimated state of health (SoH) over time as vehicles accumulate miles, charge cycles, and climate exposure. Their annual battery health report synthesizes findings across the vehicles in their dataset, producing the closest thing to a longitudinal clinical trial that exists for EV battery aging under real ownership conditions.

The headline finding across their dataset is reassuring for most prospective EV buyers. Median battery capacity retention after five years of ownership is approximately 88–91% of original rated capacity for liquid-cooled vehicles. Tesla Model 3 and Model Y Long Range variants retain a median of 89–91% at five years, with higher retention among owners who limit daily charging to 80% state of charge and vehicles in moderate-climate states. The standout on longevity is the Tesla Model 3 Standard Range LFP (2021+), which tracks at approximately 94% median retention at five years — consistent with CATL's 6,000-cycle cell rating and LFP's fundamentally flat degradation curve. Recurrent's early data on this cohort at higher mileage continues to outperform NMC-based vehicles at equivalent age.

The most powerful predictor of degradation in Recurrent's data is not mileage, charging behavior, or model year — it is thermal architecture. Air-cooled vehicles, primarily the Nissan Leaf through the 2022 model year, show dramatically worse retention curves than liquid-cooled peers. Leaf owners in California, Arizona, Texas, and Florida report median retention of 72–76% at five years, compared to 85–88% for the same cohort in Minnesota and Wisconsin. The data quantifies what the community has observed qualitatively for years: passive air cooling is structurally inadequate for high-cycle residential EV use in warm climates, and the degradation penalty accumulates continuously, not suddenly. A Leaf in Phoenix with 50,000 miles is not comparable to a Leaf in Seattle with 50,000 miles, and neither is comparable to any liquid-cooled EV of similar vintage.

Recurrent's data on DC fast charging frequency offers more nuance than conventional guidance suggests. For liquid-cooled vehicles with active thermal management of the charging process — Teslas, Ioniq 5/6, BMW i4 — there is no statistically significant correlation between DCFC frequency and accelerated degradation at realistic usage rates (defined as DCFC comprising fewer than 80% of total charging sessions). For a Nissan Leaf in a hot climate, the correlation is present and meaningful: every additional DCFC session represents a thermal stress event without active cooling to dissipate it. The practical guidance differs by vehicle architecture, not by chemistry alone.

For used EV buyers, Recurrent's vehicle history reports — available through their website and integrated into several used-car marketplaces — translate this population-level data into a single-vehicle assessment. A report provides the specific vehicle's battery health reading relative to its model cohort, flags anomalous degradation, and includes a confidence interval based on data density for that vehicle. At $19–39 per report, the cost is trivial relative to the decision it informs: a 15% battery health difference in a used Model 3 represents 30–45 miles of reduced highway range and potentially $3,000–5,000 in resale value difference at current market prices. For private-party EV purchases in particular, a Recurrent report is the single most useful piece of pre-purchase due diligence available.