What Goes Into A Tesla Model S Battery--And What It May Cost

Tesla Motors - Model S lithium-ion battery pack

Tesla Motors - Model S lithium-ion battery pack

Enlarge Photo

In his recent interview with Barron's, Tesla Motors CEO Elon Musk hung up on the reporter who was interviewing him.

He said he had "no interest in an article that debates what we consider to be an obvious point -- which is that there is a dramatic reduction in battery costs," then went on to tell the reporter, "You clearly do not understand the business," before apologizing and ending the interview.

The resulting Barron's story argued that Tesla's stock price was overvalued, because the lithium-ion cells used to power Tesla's vehicles cost a great deal.  

Reading the story, it becomes clear that the author believes Tesla has to spend $400 per kilowatt-hour to build the battery pack for its Model S electric luxury sport sedan.

The figure is widely cited by journalists who write about Tesla--The New York Times last year, for instance.

This "price" may come from the $10,000 price difference between the 60-kWh and 85-kWh versions of the Model S, since $10,000 divided by 25 kWh produces a per-kWh figure of $400.

Quick market test

To test that notion, I contacted one wholesaler and offered to purchase a small number of the '18650' lithium-ion cells Tesla uses in its packs.

That company offered to sell them to me at a price of roughly $350 per kWh, including the attached circuit boards used to group battery cells into larger assemblies, which can retail for almost $4 each

Considering the small number of cells, and the free offer of attached (or unattached, my choice) circuit boards, it seems clear Tesla's price in great volume could be much lower.

Nevertheless, the $400-per-kWh "price" seems to have been widely accepted without further inquiry--perhaps because it is still much less than what competing automakers appear to be paying.

2013 Tesla Model S

2013 Tesla Model S

Enlarge Photo

Tesla: thousands of cells

In contrast to every other automaker, which use specialized large format Li-Ion cells, Tesla's battery pack is made up of thousands of inexpensive commodity cells similar to those found in laptops.

Unlike automotive cells, these cells are produced in the billions, subject to the fierce competitive pressures that are a signature characteristic of the computer and consumer electronics industries.

Even including the overhead of the pack enclosure, connections between cells in modules (and modules in the pack), sensors, and circuitry, Tesla likely has lower pack costs than any other maker of plug-in electric cars.

Simplifying a cheap cell

But for the Model S, Tesla  redesigned what was already a relatively simple cell to be much less complex, and to have a much lower manufacturing cost--largely by removing expensive safety systems built into each individual cell. 

When used as a laptop battery, each cells requires a safety mechanisms to prevent fires. But in a large, electronically-controlled, liquid-cooled battery pack like the one used in the Tesla Model S, having certain safety features on each cell would be redundant.

2012 Tesla Model S body-in-white

2012 Tesla Model S body-in-white

Enlarge Photo

In this case, the company's cell design eliminates the relatively complicated battery cap of the commercial cell, and replaces it with a simple aluminum disk.

Intumescent goo

Having radically simplified the cells, Tesla then designed simple and inexpensive fireproofing systems into its battery pack.  Among many innovations, Tesla appears to have incorporated a form of intumescent goo that it sprays onto the interior of the pack to aid in fireproofing. 

When exposed to heat, a chemical reaction occurs in the goo that helps cool the heat source, while simultaneously forming a fireproof barrier to protect the rest of the pack.

In testing by Tesla, this material often cooled cells experiencing a runaway reaction--to the point that many failed to ignite at all--and provided a fireproof barrier surrounding those that ignited.

The potential safety advantages of Tesla's small-cell approach were highlighted during the Boeing Dreamliner battery-fire fracas.  

'Revenge of the Electric Car' premiere: Tesla Motors CEO Elon Musk on red carpet

'Revenge of the Electric Car' premiere: Tesla Motors CEO Elon Musk on red carpet

Enlarge Photo

As Elon Musk pointed out, it can be quite difficult to cool large-format cells efficiently, and even harder to contain them once they do ignite.

Thus far, Tesla has never experienced a battery fire in a production pack.

Tesla's price advantage

But even without the simplified design Tesla created, the standard Panasonic NCR18650A 3100mAh cells that Tesla uses probably don't come close to costing it $400 per kWh. 

Panasonic is an investor in Tesla Motors, so there could be an incentive to work hard on lowering the price for specially-developed cells to a company it partially owns.

And for years now, people associated with Tesla have said its battery packs would cost under $200 per kWh--it's a figure that's hardly news.

Yet that's the price that prestigious consulting firm McKinsey suggests will be reached in 2020 by the industry at large.

Follow Us

Comments (62)
  1. The $400 is justified even if the price of a "solo" Kw is below 350 or lower. You need to include the cost of the infrastructure to install the batteries (in the car)... the cost of testing (crash testing etc.) the batteries prior to full production of the Model S. And the price of the "case" (manufacturing, installation, fireproofing etc.) the batteries come in. Once you take all these factors into play Yes the price will come down as volume increases (of cars sold) Each car chips away at that initial start up cost. Eventually coming down to what it cost them (plus 10%... they are running a business) and thus that savings trickles down to the consumer...

  2. You didn't read the article. He didn't say the price per cell was $350, that was what it would have cost him to buy a handful of standard cells. He makes the claim that what Tesla is paying is probably closer to $170 and maybe even lower. You can't get from $170 to $400 with just the pack costs.

  3. Well, it was more than a few cells. I could have made a handful of cars with my proposed purchase. But it was still a TINY purchase on the scale these guys operate at, and they were throwing in free equipment (PCB's) and labor to integrate the batteries with the PCB's. And it was still only ~$350/kWh.

    The academic and industry sources indicate the cost is probably under $200/kWh. And the ads (including for competing cells) show it might be much less.

  4. Great article Thomas...I have no doubt about the long term profitibilty of Tesla. I wish I saw this early and invested in the company from the start....hope u did that...

  5. My working estimate for the cost of the pack enclosure and cell integration is ~$2,000. The pack is impressively simple and manufacturing is probably almost entirely automated.

  6. Excellent analysis. For years on GCR and elsewhere we have seen discussions of $1000/KWH decreasing at 8% per year. I think the numbers are way too high for two reasons stated in the article.

    First, low volume prismatic cells (modules) have been available for $400/KWH to the consumer for years now and 18650 cells for much less. And, as the author wisely states, undoubtedly manufacturers get much greater discounts.

    Yes, there is a build up in cost as we go from cell to module to pack, due to supporting equipment that must be considered.

    However, testing and tooling costs are likely to be a small fraction of the price at these volume levels.

  7. Agreed. The numbers in this article are radically different from the numbers the "exerts" are floating. Mind you, some of the confusion might be the result of Tesla being the only one using high energy density cells that will only work in big packs because they tend to have low power density.

    The rest of the industry seems to stick to some sort of covenant that EVs are for city use only and therefore they use the high power density cells that are suitable for small packs. The economics for these high power density cells may be completely different from what Tesla is using.

  8. Yes. Even $400/kWh is so low by industry standards that nobody questions it. Even $350/kWh is a newsworthy number, and I got that plus extra goodies as a walk up customer.

  9. Very interesting article. It would appear then if the generic brands are any indication that Tesla's cell cost could sink to as low as $75/KWh in the not too distant future. Much lower than I would have thought possible and the sort of numbers that make the affordable car Tesla has announced not only entirely feasible but extremely profitable.

  10. Over at TMC we are trying very hard to ignore the large number of $1 ads for these cells. But the Generic prices are unambiguous, and extremely cheap.

  11. Now we see why Elon Musk hung up on Barron's: he does not suffer fools lightly.

    Great analysis of the state of battery costs. Much more factual than interviewing a GM executive ;-)

  12. Yes. He can't just come out and say what his cost is, but he has been dropping hints for years, and probably tried to do so again in the interview. But Barron's seemed wedded to the $400 prices, and even tried to argue that Tesla wouldn't be able to deliver a 200+ mile Gen III. That's probably what set Elon off.

  13. If you want a 200+ mile Gen III for under $40,000 just tell Elon that it can't be done and it will, because Elon love nothing more than proving the naysayers wrong.

  14. The quotes from GM are about how quickly, and at what rate prices would fall. Barron's seems to have gotten the $400 number from a Morgan Stanley analyst, who probably got his number from the NY Times. The NY Times got its number based on the price of the Model S as I showed.

    They would have done better by calling Dell or some other laptop maker.

  15. Good article Thomas, but although I've seen the Panasonic NCR18650A 3.1Ah cell mentioned as the cell in the Model S several times, has this been verified? The cell is not well suited for 4-5C discharge rates that the drive system can demand. I know the capacity makes sense, but I'd like to know if Tesla has ever confirmed it uses that specific cell and not another from the Panasonic cell family.

  16. Nobody has cut open a Model S pack yet. A RAV4 pack has been cut open, and i used the photos to verify the use of intumescent goo, and Teslas simplified cell design, but you cant confirm chemistry or capacity.

    However, there is a tremendous amount of evidence supporting the idea. And this press release - http://www.teslamotors.com/fr_CH/about/press/releases/panasonic-presents-first-electric-vehicle-battery-tesla

  17. Trust me the GMs and Ford's have already cut open the Models S pack already!! Of course they won't say so openly......

  18. Can you post the pictures of the RAV4EV pack or tell me where I can find them? Can you tell if the cells actually say NCR18650A or perhaps something slightly different, like NCR1860PD (slightly less capacity but lower impedance and better power handling capability)?

  19. I am not sure you will gett 4-5C tho. The bigger the battery the lower the amp/cell will be. And teslas 85kwh is a big battery
    If we know the max voltage from the battery we can calculate the amp/cell tho. Any one know the voltage?

  20. I'm quite sure that many drivers have tested this empirically by stomping on the accelerator while watching their power consumption graph. But it's obviously something that they would want to test a very large number of times, so that you could do a proper statistical analysis...

  21. At about $170/kwh, the 85Kwh pack costs around $17,000. This translates to less than 25% of the S $80,000 retail.
    I am not a business man nor an accountant; but with these numbers I am not sure I understand how Tesla makes money ONLY as a result of carbon credits and NOT in the actual sale of the car.

  22. Just for kicks, let me add that $17,000 equals to the gas consumption of a 20 mpg car for 90,000 miles (with gas at $3.80). Add about $2K for electricity.

  23. To be fair you also have to account the extra maintenance cost of an ICE vs. an EV over 90,000 miles (oil change and filters, fuel filters, spark plugs, belts, etc.) Also aside from the battery cost there may be a net gain for an EV in the cost of the rest of the drive train vs. an ICE. Electric motors vs. engine block, pistons, camshaft, coolant system, exhaust system, etc.

  24. I wonder how the cost of the rest of the drive train compares. The gas car needs a lot of things that the electric doesn't: a tank, fuel pump & filter, IC engine, oil system, transmission, and exhaust system. There's a lot of maintenance on all that over 90K miles as well. The total drive train could be a lot cheaper on a Tesla or Leaf than a gas car.

  25. The motor is the size of a water melon, and the rest of the drive train is just a differential. Three moving parts total in the whole system.

    It must be just a tiny fraction of the cost of the battery. I wouldn't be shocked if someone told me it cost less than $2,000 to build those components, with most of that being the motor/inverter.

    If Tesla is building a complete pack for under $170/kWh, their whole drivetrain might be less expensive to manufacture than the drivetrain from a competing car, like a BMW 7 Series. Just the engine from that car can cost $30,000 new, and the transmission is maybe another $5,000. That leaves out the drive shaft and differential.

    There is a reason Tesla expects 25% margins even without credits.

  26. "Just the engine from that car can cost $30,000 new, and the transmission is maybe another $5,000."

    You keep getting confused between the cost of producing those engines and transmissions vs. the selling price of those parts to you.

    Most automakers can make engine at a cost ~ $2-$4 per hp. A 300 HP enginee from the 7-series will cost no more than $3k. That $30k price is there to rip off the rich fools....

  27. @Juan Pelotas,

    Tesla doesn't "only make money as a result of carbon credits". During the Q1 2013, Tesla improved it's margins on the manufacturing of the car by dropping labor hours/car by 40%. Many of their suppliers provide cost breaks when the # of parts ordered reaches certain thresholds.

    Just car margins will be over 25% by the end of this year. When you add in carbon credits and ZEV credits, their margins will likely exceed 40%.

  28. Yes, Tesla's cost problems are mainly tied to bad supply networks and other teething pains. They should be able to make money once they clear all of that up, irregardless of regulatory credits.

  29. There was an article (I believe on this site), that stated that Tesla was profitable last quarter only thanks to the credits. I am a fan, and very glad to be wrong.

  30. @Juan: Yep. This one:


  31. Tablets (that are lighter and more economical in energy use) are narrowing the laptop market, so may have less demand for that kind of cell for computing purposes, that can be redirected for Model S assembly.

    It may be a cost saving factor, too.

  32. Good point. The IEK report from Q3 of 2012 lists 18650 prices at $120-$200/kWh, while the soft cells used in tablets were around $500/kWh. The IEK cited decreased demand from laptop makers for the fall in 18650 prices.

  33. So.. Good thing that EV demand will keep manufacturing up, so that pre-existing plants can have their capex amortized a bit longer :)

    Hopefully, Tesla's keeping tabs on these (and other) developments, and has the wherewithal to reengineer the battery while keeping it within the same overall formfactor, so that future improved batteries can be mounted onto older cars when their batteries finish their usable lives. So that in, say, 2025, you can have your old and busted battery swapped out for some new 200kWh hotness in the same size and weight balance (though maybe with some suspension tweaking for overall mass reduction).

  34. Maybe in 2025 make a hot rod will be easier. :)

  35. I had always thought larger format cells would be the way forward for all EVs. I didn't realize thermal runaway could be so easily controlled on the smaller cells.

  36. If you think about it, it just makes sense. Its like the relative difference between a thimble full of gasoline and a bucket.

    It's a lot easier to keep a bunch of thimbles from having a cascade reaction than it is to protect a few buckets. If you have a small problem with a small cell, it will tend to want to stay a small problem.

    If you have a small problem with a large cell, it almost immediately is a danger to be a large problem.

  37. It's difficult to know where to start with a comment.

    Tesla doesn't use a laptop market 18650 in the model S. They use a NCA chemistry EV designed cell that happens to share the 18650 form factor. It has radically improved cycle life and lower internal resistance than the older cobalt oxide based 18650's used in laptops.

    You mention Ultrafire 4000mAh 18650 cells... Ultrafire isn't even a mfg, they are a company that re-shrink-wraps whatever cells it rips out of defective laptop packs and shrinks them with whatever numbers they like printed on the side. Search for discharge data on 4000mAh 18650's and you will find the real capacity ranges from

  38. I clearly said "similar to those used in laptops". As to your comments about the chemistry, we actively debate those topics at TMC, and I stayed away from that debate in the article, specifically because it is not a resolved issue. If you have evidence, you need to present it, and I'd suggest TMC as the forum to debate that.

    As to Ultrafire, that was provided as an example. Again, I am aware that the discharge rates don't match the specs, and I am also aware that someone tried to build an EV battery out of them and it caught fire, lol.

    But there are dozens of brands advertising 18650 cells at extremely low per/kWh prices. And there are many ads for both the NCR18650A and NCR18650B with prices as low as $1 per cell.

  39. Just to be clear, those of us who discuss this at TMC think it likely that the market price is closer to $2/cell. But there are many generic brands going for much less than that, and the point of the article was to document the ongoing price reductions.

    The IEK data is from an industry source, putting an upper bound of $200/kWh. The 2009 study showing costs at the time to be $200-$250/kWh for 18650 cells was commissioned by the Department of Energy. Throwing up your hands and claiming its all nonsense is not evidence.

  40. Tom,
    Great article. Do you know anything more about the Panasonic 3.4 amp and 4.0 amp Lithium cell? I guess that they would be about a 10% and 20% improvement compared to the present 3.1 cell. Do you think that the 3.4 amp cell is planned for use in the Model X and the 4.0 amp in the Gen III?

  41. So wait. You know that the Ultrafire cells have completely fictionalized capacity numbers and you posted it anyway as if it was a fact? Why should anyone believe anything you say if you would knowingly post something as fact that you knew to be a lie. Please correct this mistake in the article as it makes you look terribly misinformed. It was an otherwise good article.

  42. The NCR18650A batteries discussed in the article are already NCA chemistry, not the older cobalt oxide based 18650s.

  43. Yes they do. I missed his reference to NCA chemistry. There is some debate as to whether Tesla has a tweaked chemistry or the stock one. But the stock chemistry is NCA.

  44. I'm sure Tesla doesn't make much money from selling its cars. The profits come from elsewhere. You can figure that one out easily. Another bet I'm willing to make is that Tesla is going to move towards energy management, where the real money is. It's dumbfounding Detroit and utilities don't see what they are doing. Tesla is moving forward, once chess piece at a time and does a brilliant job at becoming tomorrow's leader. Give it another 5 years and the auto industry won't understand what happened to their traditional and outdated business model.

  45. There are non-automotive moves that Tesla is making that I've been taking a hard look at and which I find to be quite interesting.

    That said, I expect Tesla to make a lot of money on the Model S, and even more money from regulatory credits. The real risk is a sudden decline in demand from the levels it's at right now. Whether because of a bad recall or whatever. Even so, they are in a much better position to ride something like that out than they were before their recent money infusion.

  46. Great Article Thomas... Thank you for actually doing some research and checking sources before posting your article. I am sure your skills are being taken advantage of as a consultant but part of me wishes you were a full time writer so there were more factual articles out there.

  47. Thank you Blake!

  48. I just want to echo the kudos Thomas. I've written about a dozen pieces for Forbes (generally for a broader audience than Green Car Reports enthusiasts, but still usually interesting I think for them), and have considered a piece like yours.

    You confirmed my fears it would be far too dense for the wider audience but did a flat-out amazing job telling the story in compelling detail. I've long believed Telsa cross the $200 kw/hour threshold some time ago and I'm basically certain of at this point.

    I think this really point out the gross margin target the company has set it not only achievable but possibly something Tesla can exceed by Q4. In fact, I'm sure they have a stretch goal to do that.

  49. There is no reason to believe that Panasonic make a different design technology for the Tesla battery pack cells which could not have been bought by an outsider. Surely the battery pack offers superior reliability and strength which give an edge over other EV packs.

  50. You mean besides the links to the Tesla patents for a simpler battery?

    I've wondered about that myself. I think it's actually possible that Tesla's patents could be enforced, mostly because their "advances" involve breaking the battery, and making it non-functional (in many ways) outside of a strong BMW'd pack.

    It's not clear to me that anyone would go out of their way to patent a broken battery, unless they had a specific need like Tesla did.

    If the patent is unique, and an improvement (which the Tesla battery indisputably is in the context of their pack), it might be enforceable.

  51. Not sure how BMS got switched to BMW. Darn autocorrect, lol.

  52. Wow! This article is the best thing I have read on EVs in 5 years. Along with the comments it is hinting that Tesla are close to being able to produce a drive train that is more economical, over its life cycle (making it and recharging it) than an ICE (making it and filling the tank). If that is true in the US it is even more so in parts of the world with sky high fuel taxes. Well done Mr Fisher - I particularly like how you did a bit of internet shopping to cast doubt on the authenticity of the Barron's article.

  53. Do we know the weight and volume of the 60KwH battery in Tesla S

  54. The elephant in the room is that the extra 25 kWh also gets you faster acceleration (in addition to range) and access to the Supercharger network for life is included. I would hope that the $10k bump in price is around a $5-6k bump in cost for Tesla (including superchargers)...Tesla isn't a non-profit, you know.

  55. One issue missing is the discussion of what is included in the price of the "pack." There is a lot more to a pack than just cells. There are the connections between cells, holding fixtures to support cells, sensing wires, cooling lines, safety circuits/fuses on each cell, then the pack structure itself, battery monitoring system, fire proofing goo(?), charger with all of its complexities to balance all the parallel and series cells.
    Cylindrical cells by their very geometry lose about 30% in terms of Wh/L. Liquid cooling helps but the liquid cools only remotely, because you can't cool each cell by flooding the whole pack.
    While 18650 cells maybe dropping in price, Panasonic is not building billions of these cells.

  56. Very good article. Well researched, in-depth. What a shame Ultrafire batteries are used as some sort of 'evidence'. It is well known their capacity claims are just a beautiful print on the wrap. It weakens the article.

  57. What makes you say this?

  58. In my career I have done a lot of cost accounting. Thoughtful engineering and an understanding of mass production can make an amazing difference to the cost of an item. I get the impression that Elon Musk understands these things way better than even the best reporters!

    Peter Burgess - TrueValueMetrics
    Multi Dimension Impact Accounting

  59. This is far and away the best article I've seen yet on the Tesla battery. Thank you for writing a well researched article.

  60. Further evidence of a widespread collapse of prices, well beyond what the I.E.K. reported in 2012, can be seen in prices for generic Chinese Ultrafire 4000mAh cells.
    ..................-and could theoretically represent next-generation technology.

    Yes yes..it's an old article. But even back when this was written it was beyond disingenuous to consider UF 4000mah cells as even remotely decent batteries....
    The last part makes me laugh out loud- UF and next gen technology was a joke then and still is now.
    Go ahead, google ultrafire 18650 4000mah review

  61. Any half-attentive student of Tesla Motors knows the 18650 cell used is unlike any other available on the market. Laptop battery failure rates are not acceptable in an automobile, or when 7100 cells are in play.

    It is a reasonable guess that Tesla's incremental cost per cell is less than the casual internet price of a generic 18650 cell but to think the incremental cost is anything close to the actual cost is fantasy of those who have never had to make a business work. A good rule of thumb for electronics is that one must sell to the next tier of the distribution network for at least 3 times the "cost." Even at that many go bankrupt for not charging enough no matter sufficient demand.

  62. Unfortunately you can not purchase a Panasonic NCR18650A cell at retail for $4. The link cited in the article is for a 2-cell BMS circuit, not an actual battery. Those cells retail for $6-$10 (or more,) which equates to $560-$800/kWh, or roughly $50-$70k for the cells to build a 85 kWh pack. There is great difference between "academic" li-ion battery prices and retail prices.

    See current retail prices at: http://tiny.cc/i51gux

Commenting is closed for old articles.

Get FREE Dealer Quotes

From dealers near you

Find Green Cars


© 2015 Green Car Reports. All Rights Reserved. Green Car Reports is published by High Gear Media. Send us feedback. Stock photography by izmo, Inc.