Every week somebody tells us their Tesla battery is “shot.” They looked at the range number, did some math, got a scary result, and decided the car is on borrowed time. Most of the time the battery is fine. What’s off is the story the car is telling about it.

We got certified to work on Teslas in 2022. Since then we’ve tested a lot of packs and pulled a lot of readings, on cars with 8,000 miles and cars with 180,000. Here’s what we’ve actually seen, checked against the best public data we could find, minus the forum folklore.

Tesla battery degradation is smaller than you think

A lot of shoppers assume a Tesla loses something like 20 percent of its battery in the first 100,000 miles. When we pull the actual CAC readings (more on CAC below), real loss over that stretch is more like 9 to 10 percent. The big data sets point the same way. Tesla’s own Impact Report puts average Model S and Model X capacity loss at about 12 percent after 200,000 miles (Electrek). Recurrent, which tracks more than 30,000 connected EVs, sees a Model 3 holding about 93 percent of its original range at five years (Recurrent).

The loss isn’t a straight line, either. Most of it happens early, in roughly the first 10,000 to 20,000 miles, then the curve flattens out and gets boring (Recurrent). Boring is good.

It does vary by model.

  • Model S and Model X with the older battery systems: in our shop, about 8 to 10 percent real loss after 120,000 miles or more. That tracks with Tesla’s curve of about 12 percent at 200,000 miles (InsideEVs) and with long-running owner survey data showing about 5 percent loss early, then a slow crawl after that (Steinbuch). One caveat: not every pack is equal, and some early 90 kWh packs faded faster than the 85s (Electrek).
  • Model 3 and Model Y: a little more aggressive. On the cars we’ve tested, we see closer to 15 to 20 percent after 120,000 to 130,000 miles. To be straight with you, that’s on the cautious side of the published numbers. Tesla says Long Range 3s and Ys average about 15 percent loss at 200,000 miles (InsideEVs), and Recurrent’s highest-mileage Model 3s, at 200,000 miles and up, mostly still show 85 to 91 percent of their original range (Recurrent). So plan around our number and there’s a good chance you’ll be pleasantly surprised. Either way, it’s a usable car with plenty of life left.

Tesla oversold the range on day one

Here’s the part nobody wants to hear. A lot of people think something is wrong with their car because they aren’t getting the advertised range. But the car often never got that range in real driving, even brand new.

Car and Driver runs every EV it tests on a 75 mph highway loop. EVs overall average about 13 percent short of their window sticker, and Teslas have come up short by the widest margin of any brand, about 26 percent on average (Car and Driver). The magazine’s numbers went into an SAE paper making the same point (Car and Driver). Car and Driver also found the Tesla range readout is basically the battery percentage multiplied by the EPA number, so it doesn’t adjust for how you’re actually driving (Car and Driver). In 2023, Reuters reported that Tesla had tuned its dashboard to show rosy range estimates near a full charge, and even set up a team to cancel range-complaint service appointments (Reuters, Ars Technica). South Korea fined Tesla about $2.2 million that same year for overstating range and not disclosing how much it drops in the cold (Teslarati).

Newer cars are closer. A 2024 Model 3 Long Range hit about 85 percent of its EPA number in that same highway test, up from about 75 percent on older ones (Car and Driver). But if you compare a used car’s range to its old brochure, you’re blaming the battery for a promise the battery never made.

The BMS is making an educated guess

The battery management system (BMS) is the software that tracks what your pack can hold. It doesn’t put the battery on a scale. According to Tesla’s own service training material, as shared by owners, the BMS tracks capacity in a value called CAC, short for Calculated Amp-hour Capacity. It figures CAC by counting the current going in and out of the pack and comparing that to voltage readings taken when the car has been sitting still (Tesla Motors Club). The same material says the BMS only saves a new CAC when it decides the math was accurate, and a car that’s constantly charged and driven without resting may not update it at all.

So the number on your screen depends on how the car has been charged, driven, and parked. That’s why it wanders. We’ve seen readings that look off on nearly new cars. Tesla’s own Battery Health test works by draining and refilling the pack to recalibrate the BMS, and owners often see their range number jump afterward even though the battery didn’t change (Not a Tesla App). Same pack, new estimate. Some of the early drop owners see in the first few months is the BMS settling in, too. Tesla tells LFP owners to charge to 100 percent at least once a week largely so the BMS can keep its estimate honest (InsideEVs). And in 2024 Tesla changed its range math again to factor in battery age (Electrek). Recurrent puts it plainly: range shown is not the same thing as battery degradation (Recurrent).

Why CAC matters more than the battery percentage

If you remember one thing from this article, make it this one.

The battery percentage is a relative number. Tesla’s own material says it straight: 100 percent is always 100 percent, but what 100 percent can actually hold changes as the battery ages (Tesla Motors Club). A brand-new pack and a tired pack both say 100 percent when they’re full. The percentage can’t tell you which one you’re looking at.

CAC is the closer-to-raw number. It’s measured in amp-hours, and it’s tracked for each brick, meaning each group of cells wired together inside the pack. The weakest brick sets the limit for the whole battery. The range on your screen is built on top of it: weakest-brick CAC, times the brick count, times average voltage, times your percentage, divided by a fixed efficiency figure based on the EPA test (Tesla Motors Club). Every layer on top of CAC is another place for the number to get fuzzy.

That’s why two cars can look identical on the screen and be different underneath. Two cars at the same percentage can have very different CAC. Even two cars showing similar range can tell different stories once you dig in. One might have bricks that all match. The other might have one weak brick dragging the pack down, or a CAC that hasn’t updated in months because the car never sits long enough. Independent pack specialists watch the spread between bricks for exactly this reason. A growing gap between the strongest and weakest brick is how the BMS ends up throwing some of its worst fault codes (EV Inventions).

CAC is still calculated, not weighed on a lab bench. But it’s the closest thing to the truth the car keeps, and it’s what a shop actually looks at when we put a scan tool on a pack. When we test a Tesla, we’re not reading the range on the dash. We’re reading CAC brick by brick, checking the balance, and running a proper health cycle when it matters. That’s how you tell a healthy pack with a gloomy estimate from a pack that actually has a problem.

Your charging rituals matter far less than the forums claim

Charge to 80 percent. Never Supercharge. Don’t let it sit full. Precondition at exactly the right time. People treat these like religion. In what we’ve tested, results land in about the same place no matter how the car was treated. The babied car and the road-trip Supercharger car end up close together.

The data mostly backs that up, with some fine print. Recurrent looked at about 13,000 Teslas and found no statistically significant difference in range loss between cars that fast-charge more than 70 percent of the time and cars that rarely do, though most of those cars are still young (Recurrent). Our favorite real-world example: a former taxi Model 3 with about 217,500 miles, a car that racked up tons of fast-charging sessions, still tested at 88.5 percent of its original capacity in a side-by-side shown by the YouTube channel RSEV (InsideEVs). That’s right in line with the 200,000-mile-plus Model 3s in Recurrent’s data. On the other side, fleet tracker Geotab’s study of more than 22,000 EVs found heavy use of high-power fast charging can roughly double the yearly loss rate, from about 1.5 to 3 percent, and that hot climates add a bit (Geotab). Lab research has shown for years that heat and sitting at a high charge speed up aging on the cell types Tesla uses (Journal of The Electrochemical Society).

Our take: habits matter at the margins, not enough to keep you up at night. Drive it and enjoy it. If you want the cheap insurance, don’t leave a non-LFP car parked at 100 percent for weeks in a Texas summer.

When packs do fail, it’s usually not degradation

This is the part that matters most if you’re buying. Tesla batteries last a lot longer than most people think, and real pack failures are rare. Recurrent puts battery replacements, outside of big recalls, at about 2 percent for 2017 to 2021 EVs and 0.3 percent for 2022 and newer (Recurrent). But when one does go, it usually doesn’t fade out slowly. It goes all at once. The BMS sees a fault, opens the contactors, and the pack goes offline (EV Inventions). From the driver’s seat there’s often little or no warning.

What usually causes it? In our experience, water getting in or wiring that’s been compromised, not age or miles. The independent repair world sees the same pattern. On older Model S packs, the big failure modes are corroded components on the voltage sensing boards (made worse by moisture), broken sense wires inside the pack, and rusted fuse covers that let water in. Thousands of packs in rust-belt states have died that way (EV Inventions). Tesla itself is replacing the flood port seals on 2021 to 2022 Model S and Model X packs to better keep water out (Not a Tesla App). And in 2025 Tesla recalled some new Model 3 and Model Y cars because a battery contactor could suddenly open from a poor coil connection and cut drive power (NHTSA). That’s a connection problem, not a worn-out battery.

That’s why a real look at the car’s history and condition, plus a proper pack scan, tells you far more than a range screenshot ever will.

What this means if you’re shopping

  • Don’t walk away from a good Tesla because the range estimate looks low. Ask what the CAC and brick balance look like.
  • Don’t overpay for a low-mile car because you’re scared of degradation. The curve is flatter than the internet says.
  • Do care about how the car has lived: water exposure, flood history, accident history, and any sign of pack or wiring damage.

That’s the work we do on every Tesla we find for a client. If you want someone who has actually tested these batteries to find you a good one, start with our Match Maker.

Check us out at www.motorphilia.com, and keep us in mind for any car needs, especially Tesla!