Do Tesla Batteries Last Forever?



Tesla batteries aren’t forever in the same way that diamonds are, but they’re still pretty awesome.

Forever is a shorter stretch of time than some of us might have imagined. In a new video post (above), the folks over at Teslanomics make the provocative and controversial claim that Tesla batteries — those lithium ion cells which power the all-electric Models S, 3, and X — will “last forever (basically).” Never has a statement been more in need of a cautionary YMMV (your mileage may vary). You see, in the eyes of the youngish presenter, forever, in terms of battery life, is about 23 to 25 years.

Tesla Battery Degradation Chart

That’s how long he suggests it will take the average driver to reduce the useful capacity of their Tesla’s battery to 80 percent of its original ability. In terms of most people’s car ownership experience, indeed, 20-some years is long past the time a trade-in will occur. For others, including ultra high-mileage drivers and used-car buyers, batteries aren’t as forever as diamonds. Even those from Tesla.

Still, Teslanomics does back up its claim with some amount of proof, mostly collected from a number of different owners who’ve been contributing their data to a shared Google document. After many deep discharge cycles, a graph compiled from their results shows a surprisingly small amount of degradation: about 8 percent after 240,000 kilometers (149,129 miles).  Extrapolating from that finding, and assuming performance of the cells will continue to degrade in a linear fashion, than eighty percent battery capacity would be reached at around 780,000 kms (484,669.53 miles). Not forever, but pretty darn impressive.teslanomics

Now, this doesn’t mean all lithium ion batteries are created, or managed, equally. Even the original Tesla Roadster batteries didn’t perform this well (although, they don’t seem too shabby, either) and the issue prompted Nissan to settle a lawsuit involving early owners of the Leaf and replace other’s packs. Luckily, the technology and our knowledge of how to better prolong cell life is improving, and this appears to a far smaller issue today. Perhaps the day isn’t far off when battery are, truly forever.

Source: Teslanomics

Categories: Battery Tech, Tesla, Videos

Tags: ,

Leave a Reply

47 Comments on "Do Tesla Batteries Last Forever?"

newest oldest most voted

“a surprisingly small amount of degradation: about 92 percent”
Surely you must mean 8%?

You’re right. I was going to word that differently, but slipped up. Fixed.

And please fix the “kms”, SI-units and SI-derived symbols of units are never in plural.

It is 1 km and 780 000 km, just like it is 1 kWh and 70 kWh. Never ever kms/kWhs/kWhr/kWhrs/kgs etc.

Thank you. 🙂

s implies the unit of seconds

You know, I’ve using an “s” to pluralize SI-derived units my entire life, and no editor has ever pointed that out.
Thanks, going forward will not do.

The proper response here was “You’re right. I was going to word that differently, but slipped up. Fixed. And stop calling me Shirley.”

Extrapolation of a fitted curve to such data out to much larger values (like 3000 cycles) has quite a lot of uncertainty. Note that the fitted curve is above ALL the data points on the right side of the graph. I wouldn’t put much faith in it.

GM did lots of testing on the original Volt batteries, I never saw any extensive testing data for Tesla.

Yes, of course. Fearlessly defying Betteridge’s Law of Headlines**, Tesla batteries are fully capable of ignoring entropy and common sense, and will last until the heat death of the universe.

Or… perhaps not. 😉

**”Any headline that ends in a question mark can be answered by the word no.”

Yeah, I hate such hyperbole. Nothing lasts forever.

“In terms of most people’s car ownership experience, indeed, 20-some years is long past the time a trade-in will occur. For others, including ultra high-mileage drivers and used-car buyers, batteries aren’t as forever as diamonds. Even those from Tesla.” Given that the oldest Model S’s are only 5 years old, this debate seems rather premature to me. Sure, we EV advocates know all about the claims that electric motors last forever, but quite clearly Tesla drivetrains do not! The Model S’s aluminum body may also be expected to last longer than an ordinary gasmobile’s steel body, but again nobody is going to keep a car just because the body hasn’t rusted if everything else is falling apart and the car needs constant repairs. The truth is that we don’t know how well Tesla’s cars will hold up past the normal ~14 years that a car is used before being scrapped. Tesla’s well-publicized problems with reliability might indicate that those who are hoping to drive their Tesla “forever” will be sadly disappointed. Contrariwise, we may all be pleasantly surprised to find that, unlike most gasmobiles, a Tesla car’s maintenance/repair fees don’t increase as the car ages nearly as fast as a… Read more »

“Tesla’s well-publicized problems with reliability might indicate that those who are hoping to drive their Tesla “forever” will be sadly disappointed.”

Well, I think most of their reliability problems have been design flaws that can often happen with a new car design and especially a new car company. Thinks like the malfunctioning door handles, the milling sound from the motors, some leaks, etc.

Presumably the newer cars have those design flaws fixed and may last a very long time. Well, those complex door handles will probably always be an issue…and thus I appreciate the more simple Model 3 door handles.

In theory, EVs have FAR fewer moving parts and thus should last much longer than a typical ICE car.

I’d really like a Model S variant with the model 3 handles, or ones that work like them. Was never sold on that particular feature.

I love the fact that they are completely flush with the door and thus are aerodynamic.

But the pop-out mechanism? Meh. KISS principle. Get rid of that little mechanical gimmick that is just something else that can break down.

I fully agree, give me those more simple Model 3 door handles that are flush.

Yet incredibly, Elon stated the Model 3 is composed of 10,000 parts. Still a high number, even if few of them move.

Sounds like it….but Toyota says a typical car has 30,000 parts!

“A single car has about 30,000 parts, counting every part down to the smallest screws. Some of these parts are made at Toyota, but we also have lots of suppliers that make many of these parts. The 30,000 or so parts use different raw materials and different manufacturing processes.”

I doubt these are equal comparisons, but if so, Tesla really slashed the number of parts!

Entropy bows to no man.

They can last a long time. I suspect that is because because they are so big and they are so shallowly cycled.

But they do seem to last as long as the car. And after that, I suspect many people will buy them & use them as home batteries.

A 300 mile battery at 80% is still a 240 mile battery. Not exactly dead…Probably pretty obsolete by the time it gets there though.

degradation from 80% and down to 0% is fast. 240 mile will very soon be 0 mile

You’re confusing battery “calendar life” with the normal process of losing capacity by cycling them hundreds or thousands of times.

If a battery hits the “calendar life” limit, which means it short-circuits, then yes it will go to 0 capacity in short order. But that doesn’t mean it’s going to drop off a cliff after 80% due to normal aging from cycling. That’s just not how it happens.

In stationary storage, li-ion batteries are typically used until they fall to 50% capacity before being replaced.

No. From personal experience I can tell you he is right. Aging accelerates as the cell gets older. There is also another affect which is loss of usable capacity not due to loss of actual capacity. Due to the cell impedance rising over time you end up with both a loss in useful output described by the amount lost to the square of the current draw multiplied by the impedance. That is, if the cell would otherwise be producing 4.0V at 1A and the cell impedance rises to 0.1 Ohm, then the cell will actually produce 3.9V at 1A. That reduction in power means a loss in useful output that the cell is still producing (it isn’t a loss in output) but you can’t utilize it. Also due to increased impedance BMS will cut off cell output earlier. The BMS will have a voltage cutoff where it will cease to draw from the cell so as to prevent damage to it. The lost in voltage at the output (equal to the impedance times the current) makes the apparent cell voltage lower. So the BMS will cut off (consider the pack flat) sooner. This reduction starts off slowly, as the tail… Read more »

Bjork —

It depends on what is causing the battery to fail. Some failure modes are more linear, and will take roughly the same amount of time to lose another 20%. Other failure modes would indeed act like what you describe and drop from 80% to 0 quite spectacularly.

The hope is that this type of much quicker failure would happen within the 8yr/100K mile battery warranty. The types of failures that you described are exactly the kind of failures the warranty exists to cover.

Some of their battery packs will indeed fail this way and get replaced. Some small percent of outliers will also likely fail between 100K and 250K while the car is still on the road. That’s a shame, it is like some high mileage cars having to get a rebuilt engine. It happens. But based on these numbers, the vast majority of battery packs should still be in the cars and still working when they are inevitably sent to the junk yard.

Yeah, although we don’t know for sure if it is 80% or 60%, or outside that range. It’s hard to test and it varies with cell type. Jeff Dahn says he’s not sure how long they will last, but he is willing to bet that his measurements will tell us if the cells are improving.

Of course, this is the big problem with reusing used batteries. For a do-it-yourself project it makes sense to try, but for utilities I think they have enough other costs that they will want to start with new batteries.

Comparing the Tesla battery performance with the degradation of my 2011 Nissan Leaf battery, which is down 8 to 10% after only 25,000 miles and you have a compelling reason to buy a Tesla.

It’ more a Argument for big batteries than for Tesla. If the battery in the leaf would be twice as big you would have experienced less than half the degration for two reasons.

Battety double as big AS before means half the cycles = half the degration. Aditionally a bigger battery has a lower c-rate when charged from the same charger. And lower d-rate during driving. Both means less stress for the battery, resulting in better battery health. Aditionally you don’t drive near battery empty so offen which also helps battery health.

A leaf with 180 mile range will easily hold over 90% of original capacity at 80.000 miles driven. That is without any battery changes. So no active cooling required. The bigger the battery is, the less it’s needed from my point of view.

Well, you’re half right. A larger battery pack, all else being equal, will certainly last longer than a smaller one.

But if you charge or discharge a li-ion battery when it’s hot, undesirable chemical reactions occur which prematurely age each cell in the pack, so that happens whether the pack is large or small. It’s not the size, it’s the temperature.

Technical details here:

If you leave a Li-Ion battery fully charged while hot it also degrades relatively quickly (compared to other conditions).

If you can, try to discharge your pack to 85% full (or less) before leaving your EV out in a hot sun for days.

Last I checked, Nissan tells their dealers to not charge unsold cars to over 80% until just before delivery to the customer. Otherwise they’ll degrade on the lot (especially before the lizard pack).

That they can last a certain numbers of cycles and that they will run for a long distance is old news.

What will be interesting to see is how they will degrade depending on age and the conditions they have aged in.
Something we will not get a proper answer to for a number of years to come. And by then the battery technology has changed so much that it might be fairly irrelevant anyway.

Yeah, there’s a Catch-22 there: Since battery cell manufacturers keep changing the chemistry, by the time we know what happens to a particular battery chemistry due to aging, that chemistry probably isn’t an production any longer; everybody’s using a new and improved chemistry.

Of course battery makers and testing labs do “accelerated aging”, but that’s an artificial process that may not fully replicate what happens in normal aging.

And isn’t that what Nissan did, yet hot climates caused then so much grief?

Of course. All you need to make a lithium cell last is to keep it in its optimum 1.) temperature and 2.)charge level range.

Tesla has both:
1.) thermal protection, and
2.) 0% and 100% state of charge set at >5% and <90% (sometimes much less) of what most integrators would set their systems at.

The data indicates that degradation so far hasn’t been a big issue for the majority of packs. But it also shows quite a lot of variance. There are a few data points at about 18% degradation with only 150 cycles or so, and you’d be foolish not to expect those packs to keep degrading much faster than average. So, contrary to the upbeat Teslanomics conclusion, the data actually show some people will almost certainly get very serious degradation issues long before the car is otherwise worn out. From the chart it’s not easy to say anything about the percentage of packs that degrade at a rate more than, say, 50% higher than average. But it seems to me there’s a significant risk for the buyer here (a 5% risk for something as serious as this is definitely significant!), and I think Tesla should create a proper battery warranty that promises something concrete with respect to capacity. Today they simply guarantee it won’t fail completely, and nobody knows at what capacity loss Tesla considers the pack to be faulty and thus replaceable under warranty. Since Tesla is best positioned to know the full data set here they ought to simply bake… Read more »

Huh? I’ve been all over the forums for years and no one has had 18% degradation and talked about it.

Occasionally peoplehave algorithm errors from failing to charge the range of the battery. My original 240 only charge to 232 2 days ago… I know that I will probably get closer to 236 if I drain it all the way down and charge again.

40k miles, 2 years. About 2% but good look like 3% but that is an algorithm issue.

There are 2 big issues of course. Age and use. Age is actually a bigger deal generally not mileage.

Fully agree with your statement. What a customer is interested in is a guaranteed quality, not the average performance.
A kitchen appliance might have the life of 5 years, but often last 10+ years. If some machines break at 2-3 years, but most last 10 years, this is not a really good result for quality.

20 years or 400,000 miles is pretty much the life time of any car. I suspect that the batteries won’t last that long though. After 10 years Li batteries normally degrade anyway – even if they aren’t cycled.

What would be interesting to see is a plot of average temperature (outdoor and pack) vs battery degradation.

We know heat is the Li-ion killer; ‘m curious what the window is.

Also one thing I was never clear on and am curious if anyone knows — if the electrolyte freezes does that permanently affect the range? For example I know early i-Mievs did not have pack heaters; I was always curious if that affected the durability of the pack if it was stored in sub-freezing temps. . . .

From the hobby side we see degradation from time as well as cycles. I think we have not had enough time yet to see true time related degradation. It will be interesting to see how the cells look in 10 or 20 years. As someone else noted the chemistry has changed over time. I have hobby cells that are 1-2 years old that still cycle equal to new cells. But they don’t cycle as well as when they were new. So the chemistry has actually become worse.
We abuse our hobby cells by high charge rates, 4 to 8 C, maximum cell voltage 4.22 per cell, and higher temperatures for low IR. Cells typically get replaced every 6 months and 100 ish cycles for less than $100, so life is less of a priority than voltage during discharge.

The last time we looked at this data, it showed that one out of twenty Tesla owners have had their battery replaced. Is that number still accurate?

There probably were only 20 or 40 participants in the survey, and the two guys with the replaced batteries most likely found the forums and joined in the survey because they googled “Tesla battery degradation”. So even if that number is correct you can expect there to be a huge confimration bias.

No. Not anymore. Pack replacement was very common in the first model year or perhaps a bit more. But it’s very rare now.

So I can’t imagine you would consider it common across a spectrum of Model S (or X) owners.

I feel like i need to do a public study on this. These cells only last about 300 full cycles at 0.5C and faster rates.

Plug-In America has been running a longitudinal owner survey with similar results:

My single data point: 2013 P85, 108K miles, am at 94% of original capacity, and I do not baby my battery pack.

Why are some above 100%? If you would start the scale (100%) at the very top results, the lower ones would be around 85% , which would be a more realistic degragation…

Sometimes S’s are delivered with a bit more than rated. They usually lose that excess very quickly. You can call it degradation but the fact is that the original range varies by 1% or so. There are multiple factors here but it bears a review. 1. All batteries have a quick initial degradation on the order of 1-2% 2. There are 2 components of degradation – time and use. Time is probably the bigger one. Use degradation does not equal time degradation. 3. The algorithm for measuring range (and therefore degradation) is imperfect and incompletely understood. Keeping pack in the middle of charge range is healthy for the pack but will cause the range to drift downward as calculated. Pack needs to be rebalanced to reset calculations. 4. Because of number 3, all range data that isn’t appropriately controlled is garbage. And this leads many to garbage out conclusions. This data is unlikely appropriately controlled. 5. Related to number 4, the only measurement that counts is full charge and run to zero on controlled circuit which no one does with any regularity. 6. Temperature effects range calculations so that a warm battery has more range. There is no evidence that… Read more »

Does anyone know what is the most miles a Model S has ever driven with the original battery still in place?

It was around 300k. There were algorithm problems at that point and it was replaced. It apparently was about 12% degraded but the software was getting confused. Tesla said it needed to tweak to software but in the meantime, they gave that person a new battery under warranty.

That is the last I heard and that was 6 or more months ago. Now one could say Tesla was lying and the battery was dead.

After 300k miles it makes no difference if the battery died or not…it accomplished its mission and beyond.