Three years and tens of thousands of miles into ownership, the average electric vehicle still delivers roughly 97 percent of the range it had when it rolled off the lot. And across a tracked fleet of more than 15,000 EVs through model year 2023, barely 0.3 percent of battery packs have needed a failure-driven replacement. Those findings, drawn from federal data cited by the U.S. Environmental Protection Agency and corroborated by fleet-tracking firm Recurrent, land squarely against one of the most stubborn fears in the car market: that an EV’s battery will crater long before the loan is paid off.
As of mid-2026, that fear is not supported by the available evidence. But the evidence itself has limits, and understanding both sides of the ledger matters for anyone buying, insuring, or financing an electric car right now.
What the federal data actually shows
The EPA’s page addressing common EV battery myths anchors its case on a dataset of approximately 15,000 vehicles spanning multiple brands and model years through 2023. Two numbers stand out. First, 97.5 percent of the tracked vehicles were still operational at the time of assessment. Second, a model-year-by-model-year chart of battery replacements attributed to failure shows rates well below one percent, with the agency explicitly excluding replacements carried out under manufacturer recalls.
That exclusion is not a statistical trick. It reflects a real distinction. The Chevrolet Bolt recall is the clearest example. The National Highway Traffic Safety Administration issued recall and safety instructions for 2017 through 2019 Bolt models after a cell defect traced to supplier LG Energy Solution raised fire risks. General Motors replaced battery modules across the affected fleet. Counting those swaps alongside organic failures would inflate the replacement rate dramatically, obscuring the signal about how batteries perform under normal use.
Recurrent’s independent analysis reinforces the federal picture. The company monitors real-world battery health across a fleet that now exceeds 20,000 vehicles, using telematics data rather than lab conditions. Its published research shows that most EVs retain between 95 and 100 percent of their original range through the first three years, with degradation curves that flatten rather than steepen over time. Tesla models, which make up the largest share of the U.S. EV fleet, show particularly stable retention in Recurrent’s data, though results vary by battery chemistry and thermal management design.
Where the data runs thin
Three years of tracking can confirm that early catastrophic failure is rare. It cannot tell you what happens at 150,000 miles, or after a decade of daily DC fast charging in Phoenix summers. The oldest vehicles in the EPA-cited dataset are only a few years into service, and the raw per-model-year replacement counts and detailed range-retention curves from the roughly 15,000 vehicles have not been released publicly. Only the summary statistics are available, which limits independent verification of how different climates, charging habits, and vehicle brands compare against one another.
Warranty coverage complicates the picture further. Most major automakers, including Tesla, Hyundai, Ford, and GM, offer eight-year or 100,000-mile battery warranties. The current dataset largely captures vehicles still inside that protection window. Whether failure rates climb once warranties expire is an open question. A noticeable uptick between years four and seven, when high-mileage ride-share and fleet vehicles exit coverage, would reshape used-car pricing and insurance risk models quickly. So far, no primary institutional source, including Argonne National Laboratory’s GREET lifecycle model, has published longitudinal degradation data beyond the three-year horizon the EPA references.
Climate is another gap. Lithium-ion batteries degrade faster in sustained heat. Geotab, a fleet telematics company that has tracked EV battery health across thousands of commercial vehicles, has published data showing measurably steeper degradation in hot climates compared to temperate ones. But that data has not been integrated into the EPA’s public-facing statistics, leaving buyers in states like Arizona or Texas without a federally sourced benchmark for their specific conditions.
How to read the evidence without overcounting or undercounting
The EPA is a federal regulatory body with no commercial stake in EV sales, which makes its data a strong starting point. Recurrent and Geotab are private companies, but their datasets are large, their methodologies are published, and their findings align with the federal numbers rather than contradicting them. That convergence across independent sources is meaningful.
NHTSA’s recall documentation serves a different purpose entirely. It records safety defects and regulatory responses, not degradation. Treating the Bolt recall as evidence that EV batteries are unreliable would be a category error, comparable to citing Takata airbag recalls as proof that all airbags fail. The recall record is most useful as context for understanding why the EPA draws a line between failure-driven and recall-driven replacements, and why that line matters when comparing models.
Anecdotal reports from owners, forum posts about range loss, and social media complaints can flag emerging problems before they surface in federal datasets. But without controlled baselines, they are hard to interpret. A driver who reports losing 20 miles of range in January may be describing normal temperature effects on lithium-ion chemistry, not permanent degradation. Individual accounts are best treated as prompts for further investigation, not as standalone evidence. Readers can check EPA-rated ranges for specific models through the Department of Energy’s vehicle comparison tool, which provides a baseline against which any future degradation reports can be measured.
What this means if you are buying, selling, or insuring an EV right now
For anyone shopping for a new or used electric vehicle in 2026, the practical first step is straightforward: check the car’s original EPA-rated range, compare it against any available battery health report from the dealer or a third-party diagnostic tool like Recurrent’s free VIN lookup, and confirm whether the manufacturer warranty still applies. If the vehicle falls within the three-year window covered by federal data, the odds of needing a battery replacement are extremely low based on current evidence.
Beyond that window, buyers should factor in the absence of long-term federal tracking and price accordingly, especially for vehicles with high mileage, sparse service records, or a history of frequent DC fast charging. Replacement costs for a full battery pack still range from roughly $5,000 to $15,000 or more depending on the vehicle, though prices have dropped as cell costs have fallen and more third-party repair shops have entered the market.
Insurers and lenders are already adjusting to the early data. Low failure-driven replacement rates support more favorable residual-value assumptions for late-model EVs, which can translate into lower lease payments and more competitive loan terms. At the same time, the lack of long-horizon degradation data encourages conservative modeling for vehicles older than five years, particularly in hot climates or for models that have already seen software-limited range reductions. Those cautious assumptions may ease as more cohorts of vehicles accumulate mileage without corresponding spikes in failures.
Encouraging numbers, incomplete picture
The data available as of mid-2026 paints a clear short-term picture: EV batteries are not failing at anything close to the rates that early critics predicted, and most packs remain functional well into their first ownership cycle. The 97 percent range-retention figure and the sub-one-percent failure-driven replacement rate are the strongest quantitative evidence the market has, and they come from credible, independent sources.
The long-term outlook is still being written. Until federal datasets extend beyond the current three-year horizon and account for climate variation, charging behavior, and post-warranty aging, buyers and regulators will be working with partial information. The prudent read is that early failure is genuinely rare, that the technology is performing better than most consumers assume, and that the full lifespan of today’s EV batteries has yet to be measured in the real world. For policymakers weighing tighter emissions standards and for regulators deciding how much degradation transparency to require from automakers, closing that data gap is not optional. It is the next step.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
