Model S Battery Pack - Cost Per kWh Estimate

[Discoducky]

Ooo.. recording is back up -

http://www.media-server.com/m/p/786ftzc9

Thanks CapOp! Just listened to the whole thing and here are my notes (I may not have gotten everything)

Q3 earnings call

1. Update for Q4
   1. 25% excluding ZEV credits
      1. Pretty secure as TM is nearly ½ way through
   2. Production constrained and NOT demand constrained
      1. Alleviate these next year
2. Current assembly line capacity
   1. Main constraint is the cells
      1. Addressing this constraint as it is critical
      2. Announced a deal with Panasonic that alleviates most of the contention next year
3. Pricing of the GenIII vehicle at parity or discount versus ICE
   1. Price around 35K and consider the savings versus ICE car priced at $28 or in the EU priced at $22
   2. See’s leasing as the way to go for EV’s since the savings are essentially viewed as immediate (month to month)
4. Priorities and breadth of investments in 2014
   1. 2014 expanding Model S into Asia and more broadly into Europe and also other parts of the world
   2. X development and fine brush strokes
   3. Styling of the “generation” vehicle
      1. Will take a while to build out capacity for this vehicle
5. Panasonic agreement – Shape the agreement
   1. We’ll make a lot more cars than the current agreement and we’ll need more agreements with possibly other companies
   2. Clear we’ll need more production capacity and we are in the process of figuring that out
      1. Gigafactory!!!
6. Current level of demand – Quantify please
   1. US/NA demand has continued to increase and TM has starved NA to feed Europe
   2. Could sustain 20K/year in NA and maybe more than that
   3. Currently Europe is like NA back in January and it’s like 10K/year demand
      1. But TM is not trying to drive demand as they can’t fulfill it as they are production constrained
7. Production/week #’s – Comment on produced vs sold
   1. Lots of cars in transit
   2. Not a capital issue since the cars are already directed at paid customers
8. China – Allowing for importing percentages
   1. TM would qualify for the % of EV’s that China allows to be imported
   2. February launch
   3. Homologation is finished and have all approvals necessary to ship the car
   4. Put first cars on a boat in January
   5. Seeing good initial demand and TM isn’t doing any marketing
9. Cell supply issue – Building you own battery cell plant
   1. We are not quite ready to make a big announcement on the Gigafactory, but are exploring options
      1. But if Elon were to guess he would guess that a Gigafactory would be the best option and it would be in NA
10. How does the cell issue rank in terms of challenges
    1. Biggest single constraint and that is the critical path item
    2. TM can ensure that everything else is less important
11. Are raw materials an issue?
    1. No, main constituents other than lithium are not an issue either
12. Panasonic supply agreement – What needs to happen on their side?
    1. They are covered for 2014
    2. TM has high confidence they can deliver
13. Model X
    1. Few units in 2014 and high volume production is Q2 2015 (normal ramp)
14. Daimler
    1. B-class is going to be a great car and most compelling electric car on the market
15. Shipping logistics – Risks?
    1. We’ve got it ironed out and shipped partially built cars to Netherlands and ramping up investment there to add more capabilities.
    2. Every countries electrical grid has its own challenges and nuances
    3. 0 cars in Q2 and >1000 cars starting in August
16. Balancing shipments to different countries
    1. TM is aware of the tax credits and sympathetic to customers to achieve satisfaction their
    2. Currently in a normal delivery mode
17. Battery capacity for GenIII
    1. Not the right time to talk about the Gigafactory
    2. Green factory, with a lot of solar panels and recycling with no harmful emissions or toxins
    3. TM is trying to figure out what way to do “Version 1” of the Gigafactory which is comparable to all Li-Ion production in the world
18. Q3 or outlook benefits from tru-ups
    1. No such luck and no one time benefits as it was a clean quarter
19. Comment on wait time in Europe and China
    1. 6 to 7 weeks for Europe is best
       1. Subassembly ships to Netherlands
       2. And there is a backlog so about 3 to 4 months currently
    2. Ordering in China now would receive in mid to late Q2
20. Update on X-country trip
    1. Might do that trip during spring break due to schedules and weather
21. Mexico crash
    1. Car actually went through several structures
22. Is demand leveling?
    1. Huge amount of untapped demand in NA
    2. Saw a huge amount of reservation increase
    3. Less delivered cars was due to starting deliveries in other countries
23. Free cash flow – Update high level guidance
    1. Prefer to give 2014 guidance in the next call
    2. Depends on lots of capex details to be discussed in next call
    3. Balancing growth with cash flow
       1. Infrastructure investment is high
24. GHG and capex
    1. Effects continue to decline in every quarter and slightly over 1%
25. Improvement in tac-time and overtime
    1. Production efficiency is continuing to improve
       1. Room to improve in labor costs
       2. Head count per car should get better
26. Relations with battery suppliers and how they are contributing to IP of the cell
    1. Internals of the cells are evolving quite a lot
    2. TM is being helpful to Panasonic and has a lot of exciting things in the pipeline
27. Cycle time for improvements
    1. General goal is make a material improvement every 4 years and we started this cell step 1.5 years ago, so you’d assume to see another step change in 2 years
28. For GenIII – Is 18650 still the right form factor
    1. No reason to go away from this form factor, but it would be odd that it was perfect for long term
    2. Most likely if change, it would be bigger, but not hugely bigger
    3. Large format cells are not cheaper
29. Configuration of production – S’s and X’s
    1. TM has a game plan and if demand eclipses 100K for both, then they would reconfigure production lines
       1. Not a limiting factor and TM has a handle on how to get there

 

[ItsNotAboutTheMoney]

Boom. Tesla Giga-factory seems to be the leading option. This thread was all over that issue.

- - - Updated - - -

We've done the math on this. The capital requirements are massive.

Of course, a big question is how much do they have to spend on the Gigafactory* before it starts earning money. If they're planning full cycle, I figure they'd be spending a lot of money. Maybe start with the recycling facility and giant solar farm? Recycling absolutely makes sense because it allows them to hedge against commodity price increases.

* Interesting pattern in ueber bull Andrea James' (Dougherty) questions in the call:
(1) Are you going to build your own cell (Giga)factory for Gen 3?
(2) How does cell production rank in Gen 3 challenges?
(3) Are raw material supplies an issue [for the Gen 3]?

 

[brianman]

I didn't pull out the muffle but here's my quick capture with only a couple rewinds:

we're not quite ready to make a big announcement on the kind of cell factory / gigafactory

exploring a lot of different options right now

if i were to guess we would do that kind of gigafactory

raw cells coming in to finish packs

with partners

my best guess

that factory would most likely be in north america

but investigating other options as well

-----

raw materials are not an issue

correct

wouldn't worry about lithium supply

lot of lithium out there

main constituents by weight are nickel, then cobalt, aluminum, then lithium

lithium 1% of the pack mass
2% of the cell mass

-----

as far as next year, they got it covered

been working closely with them

spending a fair amount of time in panasonic factories
(24:51 muffle)

we have high confidence Panasonic can deliver on their commitment

 

[CapitalistOppressor]

I didn't pull out the muffle but here's my quick capture with only a couple rewinds:

Thanks, I can't understand the interruption either (sound engineer anyone?) but the rest appears to be spot on.

 

[JRP3]

Was it a squink, or galonk?

 

[Porter]

The most recent Panasonic deal outlined supply agreement for 2 billion cells. Now the reports are that it represents $7 billion in revenue for Panasonic. In rough terms that would be $3.50 a cell average ($24,684 cell price for the MS pack). While not completely unreasonable given some of the engineering cost that Panasonic may have done to get the Tesla cell down to an optimized / lower cost cell, it is higher than most of the analysis that has been done in this thread. I realize there is still some amount of disagreement on this issue, but all my information points to Tesla using 7,104 of the 3,100mAh cell (NCR18650A). The 3,400mAh cell yields are still not great and the cost/price is too high. That being said, that average of $3.50 could represent the next generation cell price averaged with the current 3,100 cell. JB's comments about the pack costing 1/2 or even 1/4 of the vehicle are still not too far off when you consider they are charging different prices in different countries.

 

[trils0n]

The most recent Panasonic deal outlined supply agreement for 2 billion cells. Now the reports are that it represents $7 billion in revenue for Panasonic.

I thought the 7 billion number was just kind of a random guess by a journalist. No official number correct?

 

[JRP3]

Correct, the 7 billion number is completely made up.

 

[tftf]

I don't think $2 to $3/cell (including a mark-up for Panasonic's margins) is that far off once the "supply" glut of re-started 18650 factories is over and higher Ah (4+) density is reached, maybe $3.5 is too high.

There was also high volatility in USDJPY in recent months with the Yen falling, maybe the Japanese press quoted in JPY terms.

Then there's the cost of integrating these "naked" cells into battery packs and adding a BMS, maybe these numbers were also mixed up.

I guess $/KWh is an easier, universal estimate so it can be compared to EV batteries using larger cells.

In my view, the $150 area is the future "tipping point" for EV adoption in the mass-market.*

Recent studies such as the one from consultancy Mc Kinsey (2012) probably include far too high numbers:

Our analysis indicates that the price of a complete automotive lithium-ion battery pack could fall from $500 to $600 per kilowatt hour (kWh) today to about $200 per kWh by 2020 and to about $160 per kWh by 2025.[SUP]1[/SUP] In the United States, with gasoline prices at or above $3.50 a gallon, automakers that acquire batteries at prices below $250 per kWh could offer electrified vehicles competitively, on a total-cost-of-ownership basis, with vehicles powered by advanced internal-combustion engines (exhibit)

Battery technology charges ahead | McKinsey Company

- Is assuming $150/kWh by 2020 a reasonable estimate? That's the rough (optimistic) estimate I came up with.

- Do you think TSLA can already achieve similar targets ($150-200) with the Gen III car and in-house cell manufacturing by late 2016?

- Maybe some of you think $150 is not needed and the tipping point for mass-market adoption of EVs can come above that prive level?*

I would love to see studies or price estimates how and when $150-200/kWh can be achieved, thank you.


PS: For the record, this is a Musk quote on price drops most will be familiar with already:

A participant from JP Morgan asked (in torturous analyst-speak, using "dimension" as a verb) something to the effect of, "Could you speak about how certain you are of a $200-per-kWh battery price and where you see that number going?" Musk began his response by seeming to deny that figure represented the company's current cost but went on to say, "I do think that cost per kilowatt hour (kWh) at the cell level will decline below that, below $200, in the not-too-distant future."

Battery cost dropping below $200 per kWh soon, says Teslas Elon Musk

_______________
* That of course also depends on future gas prices. I would have agreed to higher numbers before the U.S. and other countries started getting oil from non-conventional reservoirs a few years ago. Nowadays, I think $150 is a good estimate, also because some car buyers don't factor in total costs of ownerships correctly (future savings on re-charging vs fuel costs at the pump over xyz years), which is a dis-advantage for EVs at the time of purchase.

 

[tftf]

I was completely amazed at how the analysts did not even flinch when Elon said that G3 will consume more batteries than all of the laptop market and by a good bit. He went on to say factories will need to be built. Not a single person followed up on what I would think is THE key element of Tesla's future. It is a shame that the financial side of things has receded into this myopic Q by Q lack of vision.

Allow me to highlight this excellent old comment from the thread because the same happened in the following conference again on Nov 5, 2013. I hope they wake up on the next conference call or an analyst conference in 2014.

Analysts model in Gen III sales but (apparently) most of them don't care about battery supply, pricing and the related logistics. Same for high-paid consultancies (Roland Berger, Mc Kinsey...) whose current $/kWh are not accurate, not just for TSLA but for large-cell EV manufacturers as well. (Otherwise, Nissan couldn't have lowered the LEAF price recently...)

 

[JRP3]

I would love to see studies or price estimates how and when $150-200/kWh can be achieved, thank you.

You should reference the first post in this thread which has a number of different links and analysis showing how close to that we already are. For an anecdotal reference back in 2009 some of us got together for a group buy of a few hundred large format LiFePO4 cells from China. With shipping to my door my 36 100ah nominal cells, 110+ah actual, worked out to around $370/kWh. If a handful of hobby EV builders could get that pricing in a low volume order for a non standard cell format 4 years ago it's not that hard to imagine what Tesla can get for a higher energy density chemistry in a standard 18650 format when ordering millions of cells.

 

[Porter]

I do the calculation with 3.6V because that is what Panasonic has in their tech sheet.

If I was going to simulate the pack using the 3400mAh cells at 3.6V, it would be 7104 or 7008 cells. I like 7104 but either would work.

So I've been diving down the "pack" spec rabbit hole and trying to reconcile numbers. I apologize for the cross post, but it seemed relevant to this thread as well.

Here are my assumptions:
- 74 (cells in parallel) * 3.1Ah (cell) * pack voltage (cell V * cells in series) = kWh.
- Cell nominal voltage = 3.6V and there are 96 cells in series.
- This would equal a nominal pack voltage of 345.6V.
- There is a label on the pack that says 400V. We also know that this type of cell can have a peak charge / discharge voltage up to around 4.2V (or less) before it starts to hurt cycle life. This would be about 402V - 403V.

Again remember we are talking about the nominal "pack" voltage of a very large pack and not the individual cell voltage. Nominal is not actual. There are lots of variables that go into what "actual" could be. Here is how one battery supplier describes it:

"The lithium battery output voltage varies during discharging. The nominal voltage is measured at the mid point between full charged and fully discharged based on a 0.2C discharge (where C is the rated capacity of the cell in mAh). A single lithium battery cell nominal voltage is usually shown as either 3.6V, 3.7V or 3.8V. The actual voltage profile during discharge depends on the cell chemistry being used, the discharge rate, the temperature and the age of the cells or battery being discharged. A single cell lithium battery cell maximum voltage is 4.2V, cut off voltage will be 2.5V ~ 3V."

At "nominal" voltage (3.6V) the pack would only be 79kWh. But at peak charging voltage (4.19V), this would be 92kWh. Its likely the real number is somewhere between the two. The other possibility is that they used peak voltage and calculated the 85kWh based usable vs actual. If you take the 6-7% off (safety margin) of the 92,273 number, you get pretty close to the 85kWh.

A large pack that has a moderately low discharge rate and a robust thermal management system could have a higher overall nominal voltage than the individual cell (I believe). Still, I doubt its the 4.17V that would be calculated from using 400 V - that sounds more like the max voltage.

I have very high confidence in the 7,104 cells in 74P 96S configuration and the 3.1Ah battery. By the fact that there is a label on the pack that says 400V and 85kWh, I attribute the rest to specmanship, marketing and rounding. Only Tesla knows the real world performance data for the actual numbers.

 

[bluetinc]

Unfortunately your 4.2V*3.1*7104= 92kWh is not how batteries work. The nominal voltage is stated because that is the average voltage, where it will be both above and below that. Since the energy out of the battery is V*A*t you must take the nominal voltage times the current out times the time you can discharge for. This means that the only answer to how much energy the battery could hole would be 79kWh, and not anything above that (3.1Ah, 3.6V cells, 7104 cells). If you pull some of the datasheets for the 18650 cells, and look through the graphs you will be a better sense of this.

Peter

PS. Your voltages look correct, Tesla uses 3V-4.15 for their cells, so the pack voltages would be 288V-398.4V with a nominal voltage of 345.6. Just your 3.1Ah rating needs to change.

So I've been diving down the "pack" spec rabbit hole and trying to reconcile numbers. I apologize for the cross post, but it seemed relevant to this thread as well.

Here are my assumptions:
- 74 (cells in parallel) * 3.1Ah (cell) * pack voltage (cell V * cells in series) = kWh.
- Cell nominal voltage = 3.6V and there are 96 cells in series.
- This would equal a nominal pack voltage of 345.6V.
- There is a label on the pack that says 400V. We also know that this type of cell can have a peak charge / discharge voltage up to around 4.2V (or less) before it starts to hurt cycle life. This would be about 402V - 403V.

Again remember we are talking about the nominal "pack" voltage of a very large pack and not the individual cell voltage. Nominal is not actual. There are lots of variables that go into what "actual" could be. Here is how one battery supplier describes it:

"The lithium battery output voltage varies during discharging. The nominal voltage is measured at the mid point between full charged and fully discharged based on a 0.2C discharge (where C is the rated capacity of the cell in mAh). A single lithium battery cell nominal voltage is usually shown as either 3.6V, 3.7V or 3.8V. The actual voltage profile during discharge depends on the cell chemistry being used, the discharge rate, the temperature and the age of the cells or battery being discharged. A single cell lithium battery cell maximum voltage is 4.2V, cut off voltage will be 2.5V ~ 3V."

At "nominal" voltage (3.6V) the pack would only be 79kWh. But at peak charging voltage (4.19V), this would be 92kWh. Its likely the real number is somewhere between the two. The other possibility is that they used peak voltage and calculated the 85kWh based usable vs actual. If you take the 6-7% off (safety margin) of the 92,273 number, you get pretty close to the 85kWh.

A large pack that has a moderately low discharge rate and a robust thermal management system could have a higher overall nominal voltage than the individual cell (I believe). Still, I doubt its the 4.17V that would be calculated from using 400 V - that sounds more like the max voltage.

I have very high confidence in the 7,104 cells in 74P 96S configuration and the 3.1Ah battery. By the fact that there is a label on the pack that says 400V and 85kWh, I attribute the rest to specmanship, marketing and rounding. Only Tesla knows the real world performance data for the actual numbers.

 

[JRP3]

So I've been diving down the "pack" spec rabbit hole and trying to reconcile numbers. I apologize for the cross post, but it seemed relevant to this thread as well.

I'm afraid you're just as wrong here as in the original post for the same reasons.

 

[Dan5]

Petersen is at it again:
citing $4 per cell and saying that Panasonic is taking a loss on the cells.. yada yada, yada.

BTW- great job capitalist oppressor!

 

[CapitalistOppressor]

Petersen is at it again:
citing $4 per cell and saying that Panasonic is taking a loss on the cells.. yada yada, yada.

BTW- great job capitalist oppressor!

Actually, if I am right about which article you are referencing then he was citing $4/cell as the cost to build the factory (ie if you want a factory that produces 100,000,000 cells/year that factory will cost ~$4(100,000,000) or ~$400,000,000 to build). His per cell cost to build the actual batteries was ~$1.30 which is also consistent with a number of sources we have, though Tesla is likely paying Panasonic over $2/cell and the most reasonable estimate of known constraints puts it closer to $2.50/cell for the 3.4aH cells they are using. Still, if Tesla had their own factory it seems certain that the per cell cost to Tesla would be well below $2/cell.

That is in line with the information we have collected in this thread, and honestly the last couple of articles he has posted have contained some facts consistent with our own information.

The fact that he started writing about the need for a factory (and using relatively accurate information about the issue) a couple of weeks after we started analyzing it in this thread has made me suspicious that he actually might read TMC now and again. Notice that the citation he provides for his article (Journal of Power Sources 231 (2013) (pp. 293-300)) is actually the same paper I linked up-thread which we've used as the basis for our own calculations. He didn't provide a link for his readers, but here it is -

Cost comparison of producing high-performance Li-ion batteries in the U.S. and in China

I've also used that citation previously in some articles outside of TMC that I'm also fairly sure he has read, so maybe he got it from there. But the timing of his switch to this issue, and the information he has been using still makes me think he is lurking in this thread somewhere.

If that is the case, then I'm happy that we've been able to educate him and help him clarify his thinking by focusing him (somewhat) more on the facts underlying Tesla's operations as opposed to the ignorant fantasy that he has been peddling to his readers for so long. And if he is incapable of applying logic to address the implications of his own facts, (ie while it seems obvious that Tesla will have to invest heavily in building battery factories, once they do so they can expect their per cell cost to drop from ~$2.50/cell to something closer to the ~$1.30/cell that he has cited, and that will result in a massive increase on gross margins for Tesla) then I guess we shouldn't be too harsh on him, because in the end he is still John Petersen. (Edited so as not to offend the Peterson's out there who don't like the association)

 

[JRP3]

Actually he's John Petersen :wink: I think "tftf", who has been pushing the Giga factory angle like mad on SA as well as reading and posting on these forums, may have fed Petersen the data from here, or one of Petersen's other minions. After all, if he read these forums himself he'd be much better informed than he is.

 

[TSLAopt]

Actually he's John Petersen :wink: I think "tftf", who has been pushing the Giga factory angle like mad on SA as well as reading and posting on these forums, may have fed Petersen the data from here, or one of Petersen's other minions. After all, if he read these forums himself he'd be much better informed than he is.

Yes I agree with this...tftf and JP I think are connected somehow

 

[CapitalistOppressor]

Actually he's John Petersen :wink: I think "tftf", who has been pushing the Giga factory angle like mad on SA as well as reading and posting on these forums, may have fed Petersen the data from here, or one of Petersen's other minions. After all, if he read these forums himself he'd be much better informed than he is.

Fixed the misspelling on the name..

Also, since I'm here and have been reminded about tftf, I wanted to post this video of Straubel that found its way into my bookmarks after reading one of his posts on SA. It's a nice video, but skip ahead to ~20 minute mark to see the discussion on battery production.

As to tftf's articles, outside of the minor problem that he appears to be a typical gold buggerer, his articles on Tesla's battery problem have been relatively on point fact wise, though he uses absurd straw man arguments to make his point. I mean really, why do you need to build a scenario that requires Tesla to produce batteries at a rate an order of magnitude higher than they actually need to accomplish and then use that idiotic premise to argue that it will be impossible to achieve?

With that type of ethics (or just general foolishness?) I could totally see him as acting as a minion for JP and stooging around TMC trolling for useful information to attack Tesla.

 

[surfside]

elon just tweeted that the batteries used in the new solar city installations are 200 Wh/kg vs. 250 for the model s, so no short term supply constraint on these. i'm assuming this essentially means that the batteries used are less complex than those used in tesla's vehicles (and therefore there is ample supply of these types of cells).

can you battery guys weigh in on this? this is very interesting/good news from my perspective -- one concern i had about tesla's battery storage business is that it could compete with tesla's auto business for cells (which as we know and have been discussing are going to be seriously constrained as tesla ramps up vehicle production).

surfside