Tesla battery degradation by year — what actually changes over time

Why "by year" is the wrong first question

Search for Tesla battery degradation and you'll find charts claiming to show exactly what percentage of capacity is lost in year one, year two, and so on. Treat those with caution. Degradation is driven far more by how a specific car is used and cared for — charging habits, climate, mileage, how often it sits near full or empty — than by the calendar. Two Teslas from the same model year, driven differently, can end up with meaningfully different capacity. A generic "year 3 = X% loss" chart describes an average across many cars, not what will happen to yours.

The general shape that does hold up

That said, there is a broadly consistent pattern across lithium-ion EV batteries, including Tesla's, that's worth understanding even without exact numbers:

• Degradation tends to be faster early on and slower later. Most packs lose a bit of capacity relatively quickly in the first period of use, then the rate slows and the capacity curve flattens out, continuing to decline only gradually for years afterward. This is a general characteristic of lithium-ion chemistry, not a Tesla-specific number.
• High-mileage driving and high-heat climates tend to accelerate loss somewhat, while moderate use and moderate climates tend to be gentler on the pack — again, as a general tendency, not a precise multiplier you can apply to any individual car.
• Charging habits matter at the margins. Keeping a pack at 100% for extended periods or regularly deep-cycling to near-empty is generally considered harder on long-term capacity than charging to a moderate daily level, though the practical effect for typical daily driving is modest.
• Fast charging (Supercharging) generates more heat than slow AC charging, and heat is a factor in long-term degradation — but for typical drivers who mix home charging with occasional Supercharger use, this is one input among several, not a dominant one.

None of this supports a specific "expect X% loss by year Y" claim, and any guide that gives you a precise number for your specific car without ever having seen your car's data is guessing.

What actually determines your car's number

Your own degradation curve depends on your climate, your annual mileage, your charging habits, and even manufacturing variation in the specific pack you got. The only reliable way to know where your car sits is to measure it — comparing your battery's current usable capacity against its original baseline, using your own charging history over time — not to read a percentage off someone else's chart.

How to actually track your own trajectory

Rather than trying to match your car to a generic year-by-year table, track the number that matters: your estimated usable capacity, and how it moves relative to your car's baseline, sampled from your own charging sessions over months. A single reading tells you almost nothing; a trend line built from many qualifying charges over time tells you whether your pack is degrading faster or slower than "typical," and whether that pace is stable, accelerating, or plateauing.

Amped's battery health tracker does exactly this: it estimates capacity from meaningful state-of-charge swings in your own home AC charging sessions, plots that estimate against a monthly baseline so you can see the actual trend for your specific car, and reports a degradation percentage measured against your own car's baseline rather than an industry average. Because it's built entirely from your car's own charging data, it never substitutes a generic year-by-year estimate for your real number — and it's upfront when there isn't yet enough charging history to be confident.

The takeaway

Ignore precise year-by-year percentage charts — they're averages dressed up as predictions. What matters is your own capacity trend, measured from your own charging history. Amped tracks exactly that. It's opening in stages — join the waitlist and we'll email you when your spot is ready, so you can see your actual trajectory instead of someone else's average.