Model S Battery Pack - Cost Per kWh Estimate

[DonPedro]

Years. The phase III production capacity of the massive Panasonic fab developed to produce the NCR18650 line of cells is designed to produce ~600m cells per year, and I believe it is by far the largest such factory in the world.

500,000 GenIII cars might require 2 billion plus cells per year. So we are talking multiple massive state of the art factories, and if memory serves, the Panasonic fab cost $300m.

Meanwhile, the S platform will be soaking up the rest of global production.

This is a huge source of concern. There is need for a fairly liquid market, not just one customer. Maybe this is one reason why EM would like competition to get their act together.

 

[Buran]

Years. The phase III production capacity of the massive Panasonic fab developed to produce the NCR18650 line of cells is designed to produce ~600m cells per year, and I believe it is by far the largest such factory in the world.

500,000 GenIII cars might require 2 billion plus cells per year. So we are talking multiple massive state of the art factories, and if memory serves, the Panasonic fab cost $300m.

Meanwhile, the S platform will be soaking up the rest of global production.

Yes, it is years, or has been historically at least. I have started making an analysis of the Li-ion battery production capacity of the world (as many others here I noted the growing concern in Elons TSLIVE talk), but the data is quite hard to compile. I'll might post something when I feel more confident. However, in the mean time:

This is the earliest refernce to the Suminoe plant that I find, mid 2008: http://green.autoblog.com/2008/07/2...-capacity-with-new-lithium-ion-battery-plant/

I found it in this old presentation by JB Straubel, a very nice find if I my say so: https://www.google.com/url?sa=t&rct...=sVU0VyQ6jj0H7bz54xtNpQ&bvm=bv.50500085,d.bGE (sorry for the long url, hope it works, I can't get the right one, it is the top pdf hit if you advance google "panasonic battery factory Suminoe" limited to searches up to 1/1/2009)

The Suminoe factory was up and running, producing electrodes in October 2009, and first batteries in early 2010 http://panasonic.co.jp/corp/news/official.data/data.dir/en100325-3/en100325-3.html. Around the same time Tesla and Panasonic apparently "finished"/whent further public with the deals of them delivering batteries for Model S. http://www.engadget.com/2010/04/23/panasonics-3-1ah-batteries-to-be-used-in-the-tesla-model-s-hav/

Panasonic have only realized the first half of the 600m cell project, they postponed the second phase in 2011, but as I understand it is is now on track to be realize this again.

So, from initial public mentioning of investment (discussions must precede this with quite some internal discussions of course) mid 2008 to first batteries delivered was about 18M, and >24M to any volume, and until initial deliveries for model S in mid 2012 it was 4 years. I read somewhere that one of Sanyos factories had a 12M build time and 18M to production.

I would say that it is 18M minimum to build a large factory, probably more like 24M to get a large scale factory up and running at say half capacity. I imagine that each factory must be designed to operate at only a proportion of the max capacity the first years, to be able to scale production according to demand, as per the Suminoe exampe above.

And this is for the current dynamics where the Model S eats "only" more batteries than all laptops in the world, obviously it will be a whole different logic to get an on the order of 5 to 10 Suminoe factories online in the comming decade! Half the Suminoe plant (one of the 300m phases) can supply only about 40000 Model S/X. Since they do other batteries as well, to me, this is also why Tesla are battery constrained next year, that is, until Panasonic can get going on their second 300m Suminoe phase again.

 

[kevin99]

Here is an example of a Panasonic Li-Ion battery plant: Panasonic Inaugurates New Lithium-ion Battery Plant in China to Respond to Global Demand | Panasonic News Portal

The name on the celebration banner is Sanyo in Chinese. Was it bought by Panasonic ?

 

[pGo]

I can foresee Tesla investing a few 100M bucks in Panasonic to expedite cell production. At the same time, Tesla will also knock a few doors @Samsung, @LG, @Sony to get dedicated lines for Gen 3 battery production.

The issue with electric revolution is that if 500K cars can fill current supply, what are we looking at when you have 5M electric cars. I think Nissan got that right by building its own factory (not talking about the product) to support capacity of 300K cars.

The art is to keep 18650 prices down while constantly being supply constraint with a massive ask for years to come. If the prices rise, the revolution may get delayed. I trust Elon, but this is a tougher task than what I originally thought.

 

[maekuz]

Seems to be the case as it is part of the Panansonic family per their website.

 

[pGo]

Do you guys think that the 18650 issue can be attributed to the DRAM market? For years, production increased while prices fell. But recently, DRAM prices are rebounding. Don't really know much what triggered it, but the scale of the production is similar to 18650.

 

[brianstorms]

Me thinks Envia is all hot air, since nothing - no press release, no interviews or any news - has come out in the last 18 months after all the massive too-good-to-true claims of 400Wh/Kg. What make this even more odd is that the CEO Mr. Kapadia is generally too eager to answer questions to remove FUD on his company and products on internet forums and comment sections, and has since gone under complete radio silence.

It is odd that Envia has gone radio silent. There was this San Jose Mercury News interview from 8 months ago, in which the CEO of Envia had this to say about Tesla:

Q: What do you think of Tesla Motors (TSLA) and their battery?

A: Tesla has done a brilliant job of managing the manufacturing problem. But fundamentally, they are still using the wrong chemistry. They cannot bring the battery cost down with the chemistry they are using.

So, how brilliantly is Envia executing these days, one wonders?

 

[CapitalistOppressor]

Don't have time yet to respond to the great post by Buran. But I wanted to link this story here because there is discussion of battery costs. The story itself probably deserves its own thread as well if someone has time to post it.

Tesla’s Novel Battery and Charging Technology | MIT Technology Review

The day before my drive, I toured Tesla’s R&D lab in the hills behind Stanford University. The company’s chief technology officer, JB Straubel, showed me versions of Tesla’s Roadster, its first car, and a Model S with everything removed but the frames, wheels, and the electrical propulsion system (which includes the battery, the motor, and the electronics that control them). It was a stark look at how far the company’s come. In the Roadster, the bulky battery takes up the back third of the car. The Model S’s battery and motor seem to have disappeared. Even though the battery stores far more energy, it’s more compact: it’s now a flat slab that sits inconspicuously between the wheels and serves as part of the vehicle’s frame. What’s not obvious is that the cost of the battery, per kilowatt-hour, has also been cut in half.

Straubel pointed to the wide variety of lithium-ion battery cells—the parts of a battery pack that actually store energy—that the company is testing. This included a row of small cylindrical cells about the size of AA batteries—the kind Tesla uses in the Model S.

Tesla’s choice of these small lithium-ion batteries is, arguably, one of its most important strategic gambles. Established automakers have chosen larger battery cells—they make engineering a battery pack simpler, since you need fewer of them. But the larger cells, because they contain more energy, are also more dangerous. So automakers use less energy-dense battery materials that are more resistant to catching fire. Trying to offset the lower energy density, automakers chose flat cells because they pack together more densely, but such cells cost more to manufacture.

By choosing smaller, cylindrical cells, Tesla saved on manufacturing costs—their costs have been driven down by economies of scale for the laptop industry, for which the cells were developed. Tesla could also use the most energy-dense battery materials available, in part because smaller cells are inherently less dangerous. And better energy density reduces materials costs. This approach meant Tesla had to develop a way to wire together many thousands of separate cells, compared to several hundred of the larger cells. Straubel also invented a liquid cooling system that snakes between the cells and can remove heat so quickly that a problem with one cell doesn’t spread to the others.

Choosing the smaller, cylindrical cells also gave Tesla more flexibility in packaging the cells. Large, flat cells will deform in a collision and possibly catch fire, so other automakers have had to find places within the car where the battery would be out of the way in a crash. That meant using up some passenger or cargo space. Tesla says it has passed its crash tests without its cells deforming or coolant leaking.

By most estimates, the battery for the Model S that I drove should cost between $42,500 and $55,250, or half the cost of the car. But Straubel indicated that it is already much lower. “They’re way less than half, actually,” he says. “Less than a quarter in most cases.” Straubel says more can be done to lower batter costs. He’s working with cell and materials suppliers to increase energy density more, and he’s changing the shape of the cells in ways that make manufacturing them easier.

- - - Updated - - -

The name on the celebration banner is Sanyo in Chinese. Was it bought by Panasonic ?

Yes.

- - - Updated - - -

Yes, it is years, or has been historically at least. I have started making an analysis of the Li-ion battery production capacity of the world (as many others here I noted the growing concern in Elons TSLIVE talk), but the data is quite hard to compile. I'll might post something when I feel more confident. However, in the mean time:

This is the earliest refernce to the Suminoe plant that I find, mid 2008: http://green.autoblog.com/2008/07/2...-capacity-with-new-lithium-ion-battery-plant/

I found it in this old presentation by JB Straubel, a very nice find if I my say so: https://www.google.com/url?sa=t&rct...=sVU0VyQ6jj0H7bz54xtNpQ&bvm=bv.50500085,d.bGE (sorry for the long url, hope it works, I can't get the right one, it is the top pdf hit if you advance google "panasonic battery factory Suminoe" limited to searches up to 1/1/2009)

The Suminoe factory was up and running, producing electrodes in October 2009, and first batteries in early 2010 http://panasonic.co.jp/corp/news/official.data/data.dir/en100325-3/en100325-3.html. Around the same time Tesla and Panasonic apparently "finished"/whent further public with the deals of them delivering batteries for Model S. http://www.engadget.com/2010/04/23/panasonics-3-1ah-batteries-to-be-used-in-the-tesla-model-s-hav/

Panasonic have only realized the first half of the 600m cell project, they postponed the second phase in 2011, but as I understand it is is now on track to be realize this again.

So, from initial public mentioning of investment (discussions must precede this with quite some internal discussions of course) mid 2008 to first batteries delivered was about 18M, and >24M to any volume, and until initial deliveries for model S in mid 2012 it was 4 years. I read somewhere that one of Sanyos factories had a 12M build time and 18M to production.

I would say that it is 18M minimum to build a large factory, probably more like 24M to get a large scale factory up and running at say half capacity. I imagine that each factory must be designed to operate at only a proportion of the max capacity the first years, to be able to scale production according to demand, as per the Suminoe exampe above.

And this is for the current dynamics where the Model S eats "only" more batteries than all laptops in the world, obviously it will be a whole different logic to get an on the order of 5 to 10 Suminoe factories online in the comming decade! Half the Suminoe plant (one of the 300m phases) can supply only about 40000 Model S/X. Since they do other batteries as well, to me, this is also why Tesla are battery constrained next year, that is, until Panasonic can get going on their second 300m Suminoe phase again.

Maybe I have time to respond in parts.

I don't have information on total capacity, but current production of 18650 cells is ~55m cells/month according to the energytrend links I provided last week which kicked off the current discussion. And my recollection of the total consumer LiION production from the IEK report linked in the OP was ~1.6 billion/year (to give a sense of scale of 18650 vs entire market).

 

[DonPedro]

When TM supply drive trains to third parties (like MB and Toyota), do they also supply batteries? If yes, then they would be cannibalizing their own stocks of 18650 cells, which we are discovering is a big bottleneck.

 

[maekuz]

Don't have time yet to respond to the great post by Buran. But I wanted to link this story here because there is discussion of battery costs. The story itself probably deserves its own thread as well if someone has time to post it.

Tesla’s Novel Battery and Charging Technology | MIT Technology

JB Straubel is being quoted that battery cost is a quarter of the cost of the car in most cases.

 

[CapitalistOppressor]

JB Straubel is being quoted that battery cost is a quarter of the cost of the car in most cases.

Yes, I included that link above, and you'll note that he said "less than a quarter".

We have a ton of evidence pointing to battery costs that can't be a much higher than $20,000 and might actually be less. This comment seems to place an additional bound at the $22-$23k mark. Prior to the changes with the options the product mix of cars under ~$90k probably represented ~60-70% of all sales.

- - - Updated - - -

I found it in this old presentation by JB Straubel, a very nice find if I my say so: https://www.google.com/url?sa=t&rct...=sVU0VyQ6jj0H7bz54xtNpQ&bvm=bv.50500085,d.bGE (sorry for the long url, hope it works, I can't get the right one, it is the top pdf hit if you advance google "panasonic battery factory Suminoe" limited to searches up to 1/1/2009)

Excellent find. Re-affirms a lot of other data we've collected in this thread.

 

[lolachampcar]

Buran,
"10X increase in cell mfg capacity requires $30B capital investment" for a business that returns 5%.
OUCH
So Tesla will have to either take a huge hit in margins when they do their own cells or joint venture with Panasonic (or other) to keep the low margin stuff off Tesla's balance sheet. Either way, they are going to need more cash without Tesla's expected ROI!

 

[CapitalistOppressor]

Buran,
"10X increase in cell mfg capacity requires $30B capital investment" for a business that returns 5%.
OUCH
So Tesla will have to either take a huge hit in margins when they do their own cells or joint venture with Panasonic (or other) to keep the low margin stuff off Tesla's balance sheet. Either way, they are going to need more cash without Tesla's expected ROI!

This whole issue is a one reason I've thought about doing some real research on a potential Tesla aquisition of Panasonic. As much as I dislike the combination, it's a simple fact that if not for Tesla there would be no need for Panasonic to make this kind of investment.

How can Panasonic justify the risk of a giant increase in capacity with all of the up front costs that requires? And if they are going to get the kind of guarantees from Tesla that they should get, then why not just cut out the middleman and have Tesla just make the investments themselves? And building that kind of capability from scratch is an unattractive option, which means Tesla should look to M&A's and Panasonic is target #1.

I can see the creation of a joint venture to accomplish all of this, but you are still talking about Tesla shelling out a level of resources that they haven't appeared to be contemplating thus far. And if I am going to bind myself that tightly to Panasonic anyways, I'd at least want to research the alternate strategy of just buying them out so that I have something to compare to.

This decision is existential for Tesla, while Panasonic is just out to make a buck. That's a bargaining position that I don't much like, while an aquisition at least more or less guarantees that Tesla will have a guaranteed source of batteries going forward.

 

[Buran]

Excellent find. Re-affirms a lot of other data we've collected in this thread.

Yes, I think the graph at p 3, which puts 2007 year cell cost at about 2.5USD/cell makes your estimates of below 2USD/cell for Model S quite plausible.

 

[austinEV]

I don't really see the advantage in acquiring Panasonic. By creating the demand for batteries, Panasonic and others will have a simple business case for expanding capacity. This implies that Tesla cannot in the short term depend on the lowest prices they have been getting by sweeping up latent capacity like they have been doing. Now, partners will have long term purchase contracts and some price increase will be built in over a few years to spread the cost of building the new capacity. This has the positive that they keep a low-margin business off their books, as other posters have said.

The disadvantages of going into the battery business are too long to list at 2am: They might break the golden goose. Acquisitions have a nasty way of breaking the company as key people leave and management doesn't understand the business or even the culture. Then Tesla is left trying to figure out how anyone was making 3.6Ah batteries at any decent yield. Futhermore, at SOME point some other battery tech will come along and make the current chemistry obsolete. Maybe Tesla's new battery division could adapt, but maybe they are just tooled up for the wrong thing. This becomes a version of why the big ICE automakers aren't successfully making EV's: They have invested billions in factories that make the competing technology. It would be the height of irony if Tesla was in bed with Li-Ion 18650 manufacturing and GM's supplier figured out cheap, energy dense, safe large-format batteries that allowed any fool to make 200kWh auto batteries.

 

[Buran]

Buran,
"10X increase in cell mfg capacity requires $30B capital investment" for a business that returns 5%.
OUCH
So Tesla will have to either take a huge hit in margins when they do their own cells or joint venture with Panasonic (or other) to keep the low margin stuff off Tesla's balance sheet. Either way, they are going to need more cash without Tesla's expected ROI!

Yes, tough nut to crack in terms of the scale of investment. However, in term of margins, I think it is doable. For GEN III, lets say they can make the car sans battery for 25000, including 15% margin. A base Gen III model 50kWh battery for 9000 USD. Add 15% margin for Tesla AND 10% ADDITIONAL margin for battery maker on the cells, and you still have a cost of the battery of only ~11000 USD and a total price of 36000 USD, with 15% margin for the cell maker. Remember, this is without any major new battery technology breakthrough and only the steady 8% y-y improvement, so the per kWh should be even lower with the next step change improvement. Elon has stated several times that they will not do Gen III until such a step change, e.g., to a >4000mAh cell.

I think the GEN III equation is very doable, with still some decent margins for both the battery maker and TESLA. It will not be 25% total margin, or 25% on the battery for that matter, as with Model S, unless they move quite far beyond "being stupid at making cars" that Elon jokingly phrased it in the earnings call. But if they reach 25% margin on Model S already in the first year of mass production, and learn the trade better and better in the coming years, they might be able to do 30% on Model S/X and half that on Gen III. And that actually give some room for margins also for the battery maker!

Another way to view it is that I cant possibly think that Panasonic invested in production capacity for a very uncertain new production category using their batteries, without having some guarantees that they would at least earn some decent money on this. I.e., Panasonic should already today have a better than industry average of 5% margin on their Tesla cells, perhaps 10-15%? The 5% margins of the battery industry is a consequence of the razor thin margins in the mature consumer electronics industry. EVs is a new game with a new product type.

This whole issue is a one reason I've thought about doing some real research on a potential Tesla acquisition of Panasonic. As much as I dislike the combination, it's a simple fact that if not for Tesla there would be no need for Panasonic to make this kind of investment.

How can Panasonic justify the risk of a giant increase in capacity with all of the up front costs that requires? And if they are going to get the kind of guarantees from Tesla that they should get, then why not just cut out the middleman and have Tesla just make the investments themselves? And building that kind of capability from scratch is an unattractive option, which means Tesla should look to M&A's and Panasonic is target #1.

I can see the creation of a joint venture to accomplish all of this, but you are still talking about Tesla shelling out a level of resources that they haven't appeared to be contemplating thus far. And if I am going to bind myself that tightly to Panasonic anyways, I'd at least want to research the alternate strategy of just buying them out so that I have something to compare to.

This decision is existential for Tesla, while Panasonic is just out to make a buck. That's a bargaining position that I don't much like, while an aquisition at least more or less guarantees that Tesla will have a guaranteed source of batteries going forward.

Yes, I view a joint venture as the probable solution. However, I think that they do have contemplated on this. In fact, I think that given Elons foresight and very long term perspective in his plans and investments, and given that JB Straubel already in 2008 pointed out the need for a 10x ramp up, there should already be draft plants for how to solve this. I think that the strategic partnership with Panasonic was the first step in this. Next step? Perhaps round up 1 billion for a new "Suminoe plant", that is 500m each, in a joint venture to supply the growing demand for Model S and X and start inception phase of future GEN III collaboration. Similarly to how they declared there intentions with Model S a couple of years ago, but one step further formalizing it as a joint venture. Should not be impossible to get in place that kind of funding with Tesla's current valuation, say early next next spring, almost a year after the offering this year?

- - - Updated - - -

I don't really see the advantage in acquiring Panasonic. By creating the demand for batteries, Panasonic and others will have a simple business case for expanding capacity. This implies that Tesla cannot in the short term depend on the lowest prices they have been getting by sweeping up latent capacity like they have been doing. Now, partners will have long term purchase contracts and some price increase will be built in over a few years to spread the cost of building the new capacity. This has the positive that they keep a low-margin business off their books, as other posters have said.

The disadvantages of going into the battery business are too long to list at 2am: They might break the golden goose. Acquisitions have a nasty way of breaking the company as key people leave and management doesn't understand the business or even the culture. Then Tesla is left trying to figure out how anyone was making 3.6Ah batteries at any decent yield. Futhermore, at SOME point some other battery tech will come along and make the current chemistry obsolete. Maybe Tesla's new battery division could adapt, but maybe they are just tooled up for the wrong thing. This becomes a version of why the big ICE automakers aren't successfully making EV's: They have invested billions in factories that make the competing technology. It would be the height of irony if Tesla was in bed with Li-Ion 18650 manufacturing and GM's supplier figured out cheap, energy dense, safe large-format batteries that allowed any fool to make 200kWh auto batteries.

Yes, acquiring Panasonic is not optimal. Joint venture, where Tesla invest in Panasonic, they way Panasonic invested in Tesla is better I think. They collaborate, but not a merger of any kind. Let's say Panasonic own the majority stake, but Tesla helps finance it? Someone highlighted before that a good comparison should be how car makers do build their own ICE engines and do research and collaborate on this. One can actually make the argument that EV manufactures have to be on top of the cell making game. Since most li-Ion cells will be for EVs, it will be one of their core businesses in the future, and hence, logically, they have to drive R&D, innovation and investments. But i think it is wise that this is not the case right now, for the reasons that you argue austinEV.

By the way, love the TMC community. A 24-7 discussion on one of my passions and it is 10.30 am here responding to 2:20 am posts. While you go to sleep, I should probably work instead of analyzing Tesla... lol. Was lurking around here for a year before membership and first posts :redface:, and it helped my decision of going long in Tesla after past years Q3 call :biggrin:. I have made 600% on my TSLA now and wanted to start giving back.

 

[DonPedro]

The arms-length way of securing the ramp-up would be for Tesla to guarantee volume and pay a penalty if the target volume were not met. At the same time, the supplier would guarantee supply, also subject to hefty penalties. In that way, they de-risk each other's operations. Such deals could also be spread around, to ensure a healthy supply side (Panasonic, Samsung, etc.).

Still, at the end of the day, you have a too large concentration on the buyer side (Tesla). The real question is: How long before competitors understand the genius of using the 18650 cells? I think that the events of the past year should have given good visibility on this. If they started this year, maybe they will have something out in 2015 that uses those cells? That would be good news for Tesla indeed, because seeing additional customers for the cells would probably de-risk investments into capacity by the battery manufacturers.

In any case, it seems that in order to secure supply of 18650 cells, Tesla might need to wed the format long term and forfeit some flexibility in that arena.

 

[lolachampcar]

Nothing is for free. Someone needs to invest in the capacity.

If Tesla gets someone else to do it, they are tied to that horse(s) and that technology. The vendors must have the confidence that Tesla can pay if they do not meet consumption goals. Tesla needs money.

If Tesla does it themselves (acquisition or ground up), they are tied to the technology they pick (locked into choice while GM gets lucky with another technology) and Tesla needs money.

No matter how you cut it, you have to chose a technology to bet the farm on and you have to fund building the farm. To fund the farm, you have to sell a story and Tesla to the market is the best sell. When does Tesla get enough credibility to sufficiently sell the big picture and thus fund it?




Post Run thoughts
On the other hand....

Someone with a proven track record of making large bets, a procurement team with ultra high $ experience and the market cap to pull it off could solve this problem.

If one of the big guys buy off on the G3 concept, Tesla's market cap is a very small price to pay for the opportunity to fulfill the G3 promise. I do not know the CEOs well enough to know if one or more has the wisdom, vision and stones to solve the battery problem while leaving the Tesla team/concept in place to execute. Can you imagine what Tesla could do if the ultra high dollar G3 issues were addressed for them?

Musk would have indeed accelerated EV adoption. However, it would be a shame to have the environment that attracts a Musk and his team only to loose the opportunity to another country that has the capability and wisdom to seize the opportunity.

 

[JRP3]

One point Elon made about the 18650 format is that it allows better yield. If there is a flaw in manufacturing and you get a bad cell it's only 3.1, or 3.4, ah's worth of material that gets tossed, where if it's a 20 ah cell you lose that much more material. On the other hand the smaller format and greater number of cells per ah of capacity means that many more chances for production errors, so I'm not completely convinced by his reasoning. Every time you have to case and cap a cell those are potential failure points.
I'm not concerned about other companies adopting the 18650 format, outside of contracting Tesla to build a pack for them, because they are all too far down the road with the large format cells, and I'm fairly certain there will be enough advances in cell chemistry to make them practical and get finished pack density and cost closer to Tesla's. I think at this point with all the LEAF's, Volts, etc. on the road the "safety" advantage for the 18650 format is a moot point, since none of those vehicles are having spontaneous or crash related fires, and Tesla gives up a large amount of energy density during pack construction, about 100wh/kg worth. That's a crap load of pack infrastructure.

 

[lolachampcar]

I wonder if Sony or Panasonic read these forums.