Tesla CTO J.B. Straubel: 30% Increase In Battery Energy Density From Model S To Model 3 – Video



Tesla Model 3 to use new 21-70 cells

Tesla Model 3 to use new 21-70 cells

We still don’t know for sure what exactly the battery situation (21-70 cell is confirmed though) will be for the Tesla Model 3. Reports, information, and assumptions have been varied. Elon Musk has said that the car will have a range of over 215 miles per charge. At one point, it was also mentioned that the battery pack would be smaller than the smallest offered (60 kWh) in the Model S and the Model X.

Without revealing too much, Tesla CTO J.B. Straubel shed some light on the concept during a recent speech (full video of which below). He explained:

Tesla Model S

Tesla Model S

“One of the not very intuitive, maybe not common sense pieces of this is that as you increase the energy density of the battery, you are also reducing the cost. Because in really high volume, things tend to cost about the same as they weigh. You can essentially look at the raw materials that go into the battery and how much those weigh – and that tends to be what defines the cost.”

Energy density is basically how much energy a battery packs in, in relation to its weight. So, as the energy density of batteries improves, you can power more with less. Less weight + less materials = less money (for the same or greater capacity). Straubel also made it clear that there are no monumental “battery breakthroughs” occurring, nor does Tesla or others rely on this. Instead, there is a steady and constant improvement. Batteries improve at least 5% a year.

Looking at Tesla’s energy products provides a good example of the improvements. Both the Powerwall 2 and the Powerpack 2 battery packs weigh substantially less than earlier models per kWh, while providing double the capacity.

This will continue to be the case over time, so it only makes sense that the Model 3 will have an energy density increase over the Model S. J.B Straubel announced that the Model 3 will boast 30% more energy density than that of the original pack in the Model S. He said (quote from video below, via Electrek):

“These batteries are steadily improving every single year – maybe around 5% improvement in their energy density their ability to store energy in a given amount of mass. That’s probably one of the key metrics we worry about. And when we went from the Roadster to the Model S, they have improved by about 40% and when we were designing the Model 3, they were about another 30% better. That improvement just continues on every single year in the background.”

For those interested, you can check out J.B. Straubel’s entire speech:

Source: ElectrekTrendinTech

Categories: Battery Tech, Tesla

Tags: , , , , , ,

Leave a Reply

76 Comments on "Tesla CTO J.B. Straubel: 30% Increase In Battery Energy Density From Model S To Model 3 – Video"

newest oldest most voted

Naturally, Model 3 will have an aerodynamic advantage over GM’s Bolt EV, so it will be more efficient over 40mph. The newest batteries could help a bit too. This means Tesla will no doubt top the 238 miles of Bolt EV and we all win.

Thanks GM for sticking that carrot in Musk’s face because we’ll all benefit from it. After GM’s announcement, I KNEW Mr. Musk wouldn’t allow M3 to have less electric range.

I think this mention re: improved energy density is kind of a shot at GM – Just in case they think Model 3 will actually have a range of 215 miles.

To date GM has suck nothing in Musk’s face, but rather the other way around. How GM handles this very car will define their very future.

I think Mary Barra’s comments have been exactly “in your face” to Tesla over the years. Search for her name in the news feeds with Tesla, and you can see for yourself.
This one captures what I was reminded of:

This announcement is unfortunately as clear as mud. Is this a 30% pack density over today’s Model S or the 2012 Model S? Elon was quoted as saying 6 1/2% per year battery pack density increases. This is saying 5% or so, suggesting less, or more of an approximation. 30% over five years is 6% per year unless that’s a compound increase in battery density.

Any reference to X percent improvement per year from advancing tech, whether it’s Moore’s Law or the increase in li-ion energy density, is always a compound improvement, not a linear one.

If J.B. Straubel cited 5% improvement in energy density per year then I find that suprprising, because the figure usually cited is 7.5% or 8%.

Can’t fault GM for making a good EV. Now about this 30% increase in battery storage. Anyone have an idea on how that relates to increased range? I’m guessing that 30% increase in battery storage does not mean 30% increase in range cause that would put it up to around 280 miles of range. IF ONLY!!!

30% wouldn’t be over 2016 cars, if JB is saying “~5% per year”. To me, this implies Model 3 will have range dependent upon the proportion of battery space used vs. a Model S tray (X1.05). I’m not sure anyone knows the relative proportions of M3’s skateboard.

It doesn’t. It just means the same priced battery can be smaller, and possibly a little lighter as well. That alone tends to be kind of important to cars.

Even if the Model ≡ gets 215-220 range, I’d buy the Model ≡ with SC option over the Bolt.

With the CHAdeMO adapter and SC network one will have the pick of most DCFC till they come out with a CCS adapter. Then one can have all options.

“After GM’s announcement, I KNEW Mr. Musk wouldn’t allow M3 to have less electric range.”

You don’t really understand how long automotive design takes, do you?
Tesla reportedly did their “pencils down” of the design in July, a couple of months before GM announced the 238mi range in mid-Sept.
Changing anything about the drivetrain engineering that affects range at this stage (not very many options: Larger battery, significantly changing the aerodynamics or limiting motor performance) would delay the car for many months, and cost a lot of money to retest everything, redo supplier contracts etc.
At most, Tesla can decide to change the pricing equation: Maybe reduce the (presumed) larger-battery option price.

And the mention of increased energy density has nothing to do with GM. Straubel has been mentioning the slow-but-steady improvement of battery density in public talks for years, way before GM even started the Bolt project — most likely because he wants people to be realistic about this, and not expect geometric improvements like in computer capability.

I am sure there will be a version that goes MUCH further than 238. But I think it would be nice if they do come out with a 215 mile range base version that can allows them to hit the $35K price target.

And what happened to the the next Gen Leaf? When is it being revealed?

Model 3 will make BOLT seem elementary by comparison.

it will make it seem like a 2 year old car, which it will be by the time model3 delivers

I will see the Model 3 where I live before I ever see a Bolt

That is what is the reallity.

Chevy just played a marking game.

Where’s that? Australia?

You must live in a Compliance State

All of US will see the Bolt before the Model 3, Canada too. Europe will see the Ampera-e (Bolt) probably also before the Model 3 with the exception of England.

Australia will not see the Bolt, but the Model 3. The Model 3 will be faster in Japan, China, India. So either he is asian, australian or english.

Roadster used the 2,000 mah 18650 batteries. When Model S came out, it had the 2,800 mah batteries. Currently LG has the highest energy density 18650 battery out there, it is 3,500 mah (somebody correct me if I’m wrong). If Panasonic is to beat this number, they will have something around 3,600 mah for Model 3 (in a bigger format obviously). The sad part is that Panasonic launched 3,400 mah 18650 battery in 2012. Next year, it will be 5 years and energy density will have improved only by around 1% a year. I would love to see that 30% jump to 4,700 mah very soon, but it doesn’t look like it’s on the horizon…

You appear to be using the mistaken premise that Tesla will be using the highest energy density batteries available in quantity. On the contrary, price is much more important than energy density, when it comes to building mass produced PEVs (Plug-in EVs).

Tesla has always used the cheapest available batteries which have at least the minimum energy density and power density the car needs. We can be confident they will continue to do so.

I thought the chemistry in the cells for the 90 kwh pack was the best on kwh/kg and that was due to the small increase in silicone in the anode. I also thought that the chemistry in the model 3 would be the same as in the 90 kwh S.

Is that the way you see it?

I’m certainly no expert, but I would assume that current Tesla cars are using the 3,400 mah batteries that Panasonic launched in 2012. It is likely due to the added silicone. The reason I mention 3,600 mah battery is that Elon mentioned that they would have highest energy density cells out there, that is why it’s only my guess that they will beat LG by a tiny bit. And who knows, maybe Tesla is using them already, but Panasonic certainly didn’t officially launch a bigger than 3,400 mah 18650 cell…

I’ve always said that the 300 mile range is the benchmark for range in an electric vehicle. When automakers pass that benchmark, then they can focus on efficiency and price.

This is strictly my opinion of course.

Outside of the enormous open spaces of the US 300 miles is probably 100 miles more than most people need or would want to pay for.
I probably drive more than 200 miles in one day maybe 1 or 2 times a year and as long as charging is available every 150 miles or so I simply can’t see the point of having a bigger battery the 363 to 364 days a year.

That sounds like the typical Suburban or Urban dwelling response. Where I live I have to drive at least a hundred and fifty miles in any direction just to reach a city with a population of more than 7,000 people which is roughly the population of the town I live in. so yes 300 miles would be optimal could I get by with a 200-mile car yes but when I got anywhere I would arrive at my destination city sorely in need of a charge

I wonder if this change is essentially Tesla saying

“we’re going to move to bigger individual cells, like everyone else, for the model 3 because it is cheaper and lighter”.

If the model 3 is a lower performance model of the S and X it might make a lot of sense to move to bigger cells. This would mean less surface area for cooling so slower fast charging, less effective regen and less acceleration but could come with significant weight and cost savings.

I am waiting to find out what will differentiate the S from the 3.

They are moving to 2170 because they have found that is a better size for making high energy density cells. They have said that 2170 cells will come to Model S and X next year so it’s not poorer battery.

My understanding of what Tesla has claimed, and I see no reason to doubt the claims, is that the slightly larger size of the 21-70 cells will hit the “sweet spot”; the optimized size to maximize energy density, minimize cost and ease of manufacture per kWh, and still use cells small enough to have good passive heat dissipation (that is, they’re too small to have much waste heat build up inside them.)

As I understand it, other auto makers are using significantly larger cells, about the shape and size of a thin paperback book. While the 21-70 cells are (if my memory serves) roughly 10% larger in all dimensions than the older Panasonic 18650 cells, and about 30% larger in volume, they’re still rather smaller in volume than cells used by most or nearly all other PEV makers.

The surface area of a cylinder goes down as you increase the diameter. Double the diameter, and you have half the heat dissipation per volume. This is NOT the case for a flat pouch cell. You can make the surface area to volume just about whatever you like. Imagine a brick made up of 1000 thin slices, with a molecule thick coolant passage between each. Advantage to the pouch cell.

I should have said the surface area goes down RELATIVE to the area.

Volume…not area. Mixing circle terminology, and cylinder terminology. Time for dinner.

Even if you didn’t increase the thickness when you increase the size of a pouch cell, it still increases the temperature difference between the center and the edge, because a larger cell has a longer distance between the center and the edge.

If I understood correctly what I read in a technical paper on the Volt battery pack cooling system, the engineers’ goal was to maintain a temperature differential across the cell of no more than (if memory serves; don’t quote me on this) 2° Centigrade. Make the cell bigger and that will become more difficult, no matter what the geometry is.

Furthermore, it doesn’t matter much how thick or thin the pouch cells are, if you stack them tightly together with no space in between, so there’s no place for the heat to escape except on the edges of the cells.

That is why the Volt’s battery pack sandwiches a cooling plate in between each pair of cells. Or at least the Volt 1.0 did; I’m not sure that the the Volt 2.0 handles temperature stabilization of the pack in the same way.

Sorry. If you are trying to convince anybody that it is easier to cool the center of a cylinder than a flat plate, you can forget it.

Why do you think Apple was looking at hollow cylindrical cells?

Warren said:

“If you are trying to convince anybody that it is easier to cool the center of a cylinder than a flat plate…”

I wasn’t; I was merely pointing out logical flaws in your argument. Looks like you just want to score points, so please find someone else with whom to make your pointless arguments.

I submit that how the cells are installed in the pack has as much or more to do with how well or how poorly they can radiate away waste heat than the actual geometry of the cells.

With the Model S debuting in 2012, and the Model ≡ planned to debut in 2017, that’s five years of advancement in battery tech.

If li-ion batteries, all other things being equal, increase in energy density at 7.5% per year, that should result in a 32.28% improvement. The 30% improvement claimed here suggests a slightly slower improvement rate of ~7% per year.

Of course, “all else being equal” is almost never true when it comes to engineering something as complex as a highway-capable, street-legal automobile. And anyway, 30% appears to be a round number, so perhaps the actual improvement in energy density is closer to 32%.

But the battery price is more important to Tesla’s development of the Model ≡. It’s great that the batteries have gotten somewhat smaller and lighter, but that hasn’t really been the limitation of PEVs (Plug-in EVs) since the Tesla Roadster. They can always make the car bigger (or the luggage/passenger space smaller) to accommodate a bigger battery pack, if they have to.

You’re not very good at math.

J.B. Staubel said battery density is improving around 5% per year. Today, in another thread Alaa said: “The energy density is going up by 5% a year roughly.” You scolded Alaa and claimed battery density is “going up by about 7.5%-8% per year.” You owe Alaa an apology. It’s about 5% a year.

Basic math requires you to compound the increase every year, and not just apply the increase to the base figure 5 times.

100 x 105% = 105
105 x 105% = 110.25
110.25 x 105% = 115.7625
115.7625 x 105% = 121.550625
121.550625 x 105% = 127.62615625
5% improvement per year results in a 27.63% improvement in 5 years.

100 x 1.075% = 107.5
107.5 x 1.075% = 115.5625
115.5625 x 1.075% = 124.2296875
124.2296875 x 1.075% = 133.546914062
133.546914062 x 1.075% = 143.562932616
7.5% improvement per year results in a 43.56% improvement in 5 years.

How sad for you. You can’t even do the math for a simple geometric decrease.

A 7.5% improvement each year for 5 years is n x 0.925 x 0.925 x 0.925 x 0.925 x 0.925, which comes to 67.72% of the original figure.

Not whatever the heck you claimed it is.

You certainly understand finances better than I do, sven, but WOW do you need a remedial course in elementary math!

Are you certain that the energy density keeps decreasing year after year?

LOL! What I’m certain of is that the improvement here is, like Moore’s law, a situation where things get incrementally smaller as the years pass, so the law of diminishing returns applies. Sven’s math would suggest the amount of improvement increases every year… which is backwards. The actual, measurable improvement will incrementally decrease every year, all else being equal. The rule of thumb is this: The smaller/lighter batteries get, the harder it is to improve them by the same amount which they improved in the previous year. * * * * * However, that said, it may be that sven and I are talking past each other, if you can get past his pejorative language to the actual math. Perhaps the difference in sven’s math and mine is a matter of viewpoint, not who’s right or wrong. I can see that if you were to do an analysis on the basis of measured Wh/l or Wh/kg per year, that it would increase every year, rather than decrease it be would if you measure the volume or mass of the batteries needed to yield X number of kWh. Given that ratios and percentages don’t match if you consider a change as… Read more »

And 7.5-8% improvement per year is the figure most commonly cited, despite your assertion to the contrary. Not opinion, but fact.

For example:

“Tesla co-founder Marc Tarpenning gave a presentation about Tesla. He talks about battery improvements. If you switch to HD you can see a chart that shows 7.5% improvement per year.”



Energy per unit weight (Wh/kg) is specific energy. Energy density (Wh/l) is energy per volume.

Both the Powerwall 1.0 and 2.0 provide the same power. It’s the energy that has doubled.

yeh I think Steven may have used the word power when he meant energy density.

It looks like he changed it to capacity, which is better than power.

Ambulator said: “Energy per unit weight (Wh/kg) is specific energy. Energy density (Wh/l) is energy per volume.” Not always. Energy density can be either volumetric energy density (volume/size) or gravimetric energy density (mass/weight). So energy density can be correctly expressed in terms of either Wh/kg or Wh/liter. Are the terms “specific energy” and “energy density” interchangeable? Wikipedia’s “Energy density” article says this: “Energy density is the amount of energy stored in a given system or region of space per unit volume or mass, though the latter is more accurately termed specific energy.” https://en.wikipedia.org/wiki/Energy_density Okay, that’s news to me. It seemed to me as if on the (now defunct) TheEEstory forum, at least some people who appeared to be literate in electrical engineering used the terms interchangeably. Clearly from the above citation from Wikipedia, the term “specific energy” is preferred when referring to gravimetric energy density. But that doesn’t mean it’s actually incorrect to use the term “energy density” for both characteristics. At any rate, I don’t expect to be able to remember that difference, as I’ve been using the term “energy density” for both characteristics for years. Here’s hoping that doesn’t confuse anyone. * * * * * Has anyone… Read more »

You hit the nail on the head!
Often the terminology is misleading in articles written by non-experts. Gravimetric vs volumetric should be specified whenever discussing capacity. Also missing is a specific reference to cell or pack. Usually the term “battery” is used for both, which ends up confusing many, and at times makes the info useless.

Volumetric energy density would be redundant.

“Okay, that’s news to me. It seemed to me as if on the (now defunct) TheEEstory forum, at least some people who appeared to be literate in electrical engineering used the terms interchangeably.”

My memory of that time is getting rather faded. I would be surprised if, say, EETom did that.

“At any rate, I don’t expect to be able to remember that difference, as I’ve been using the term “energy density” for both characteristics for years. Here’s hoping that doesn’t confuse anyone.”

I feel that articles need to set higher standards than mere commentators do.

“Has anyone noticed any difference in the rate of improvement, over the years, in volumetric energy density vs. gravimetric energy density? So far as I can tell, the two seem to be improving equally fast. At the least, I’ve never seen any article talk about different rates of improvement for the two metrics.”

I recall that the 18650 cells using silicon that Panasonic promised but never delivered (as far as I know) were 54 g compared to standard 45 g cells. It’s a bigger difference when you are comparing lithium ion with lead acid or lithium sulfur cells, though.

“I would be surprised if, say, EETom did that.”

Probably so, yes. I could have, or should have, added “…while others appeared to be more careful in always using precise terminology.”

GO TESLA GO…all the naysayers and haters can kiss your lily white arse. I want my model 3 with 90KWH, 0-60 in 3.2 seconds, for $50k.

Good luck with your Model ≡ order!

But I very seriously doubt the M≡ will have the option of a 90+ kWh battery pack in the first couple of years of production. Someday, yes; but not soon.

Yeah, initially I’d expect maybe 55 kwh and 70 kwh versions.

My guess is 50 and 70, if Hyundai can get 4.5 miles per 1kwh, so will Model 3. That means base model will have 225 miles range and the 70kwh version will be at 300+. Just my guess obviously, but we shall see:)

I don’t think there’s an icecube’s chance in hell that Model 3 will appear, even in base form, with a range less than Bolt EV. I think the Bolt EV is a great statement and GM’s engineers should be applauded for it. There are many reasons to buy one and it’s certainly notable that it makes the BMW i3 look pretty silly for all it’s expense and plastic-carbon sandwich – That the best BMW can do is 120 miles of range or so. Be it so – Bolt EV is a front wheel drive mini crossover with two rows of seats. There is no will at GM to mass produce the car and it falls inside someone’s definition of a compliance car as it’s definitely a limited-production machine not intended to be scaled up to 100,000s of units sold per year. In my mind, that is a compliance machine, sculpted and produced with mandates and requirements ( soon to be challenged by Congress ) in mind. There can be no argument that Model 3 was imagined as a mass-produced machine to be sold everywhere possible, not just C.A.R.B. states or where ZEV credits benefit the most. Once somebody ( Tesla… Read more »

Well said as usual. Indubitably:

ooh James. Range of a 100D at a significantly lower cost. Sounds like a winning recipe!!

“Once somebody ( Tesla ) begins selling in the 100,000s – the entire landscape for EVs changes. Suddenly we’ll see more serious and proprietary vehicles emerging to challenge it’s success.”
You do realize that GM has sold more Volts and Nissan more Leafs in the US than Tesla has sold Model S.

And the manufacturing capacity of GM/Nissan vs Tesla is?_______.

You seem to miss the point. James is making it sound like Tesla alone is doing all the heavy-lifting, when in reality there are several players, two of which have sold more EVs in the US than Tesla.

Nissan is producing the Leaf’s at combined production lines that also produces other vehicles.

They also have factories in Japan/EU/US, and definitely would be able to scale production to the levels of several hundre thousand units per year.

And they went on sale earlier, are backed by nationwide dealer networks much earlier, and cost 1/3 the price.

And what month in early 2017 will this not be true anymore?

And which of them has an internal competitor in the form of a similar vehicle from a drive train standpoint with a different form factor?

I can’t seem to find it but it would seem that S+X>leaf. If not now then by next week

I’d say a nationwide dealer/service network is a positive, as well as offering lower priced vehicles. This is what is required to reach the “mainstream”.

Making vehicles with different types of powertrains is irrelevant. Are you saying for the last 10 years that Toyota has sold the Prius (a hybrid drive train) that selling all of their regular cars was a hindrance to them? Profits = money for R&D to work on new/advanced powertrains.

Kdawg said:

“You do realize that GM has sold more Volts and Nissan more Leafs in the US than Tesla has sold Model S.”

You do realize that the Model S has outsold both the Volt and the Leaf for the past three years in the U.S./Canadian market, despite the fact that it costs about 2-1/2 times as much as either of the others?

But it’s incorrect to assume (or push the narrative) that Tesla has been the only one selling EVs over the last 5 years when there’s other companies that have sold more.

“But it’s incorrect to assume (or push the narrative) that Tesla has been the only one selling EVs over the last 5 years when there’s other companies that have sold more.”

No, it isn’t. GM has been making a big show of its Volt and now Bolt but they are most definitely trying to restrain demand for their products. Nissan has been no better. Tesla has been doing essentially all of the heavy lifting.

GM delivered the first affordable PHEV with enough EV range to cover daily mileage. Now they’ve delivered the first affordable 200+ mile range BEV. Yet you continue to say they aren’t doing any heavy lifting.

You probably didn’t follow Tesla’s news in the past. Tesla always advertised their cheapest price, but they always deliver their most expensive cars (battery + options)to the customer first.

Back in 2012, Tesla promote Model S 40kWh for $50k after tax incentives, but started delivery P85/S85 first, S60 started delivery after 6-8 months after S85, S40 was scheduled more than a year after S60, then they discontinue shortly after.

I’ll take 70kwh for $40k.

It is so nice to have a company run by dedicated engineers with vision. Yeah, sometimes they bite more off than they can chew. But Tesla really hits on the important areas . . . aerodynamics, battery price, elegant design, whiz-bang extras for profitable high-end models (P versions, Ludicrous mode, high-end styling, etc.), the Supercharger network, etc.

They’ve made a few mistakes (we all do) but they have FAR MORE hits than misses.

Will GM kill the Bolt if the Lobbying group convinces the government not to “force” GM to make EV cars?

Is that not why they are already slowing the roll out of the Bolt EV?

China’s BYD will force GM to make electric cars in most markets of the US…unless GM intends to leave the market for other profitable interests. BYD must be a regular nightmare for them.

Model s cells around around 245 wh/kg, so that means model 3 cells are around 320 wh/kg! Impressive.

Currenty I don’t have the numbers but don’t have the 22kWh and 41kWh batteries of the Zoe roughly the same size?!
-> Other manufacturers obviously already did that step to that new cell generation…


Perhaps battery capacity will advance as computer processor speed did according to Moore’s Law. Moore’s Law is a computing term which originated around 1970; the simplified version of this law states that processor speed, or overall processing power for computers will double every two year