Nice, I read that article yesterday and thought it was very good. Didn't know you had written it. Congrats!
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Model S Battery Pack - Cost Per kWh Estimate
- Thread starter CapitalistOppressor
- Start date
-
- Tags
- TSLA
Nice, I read that article yesterday and thought it was very good. Didn't know you had written it. Congrats!
Do you have a source for this? This might have been true 5 years ago, but today, it would seem that it has shifted toward thin batteries vs cylindrical 18650. In fact, it would seem that TM might actually be soaking up nearly *all* of the non-retail market for 18650.
I suppose it matters if you count cells or kWhr capacity. Certainly flat batteries of many formats are in phones, tablets and thin laptops and larger formats like prismatics. Cylinder batteries have a lot of formats, but I would guess 18650 is now the largest selling cylinder in both cell count and kWhrs.
Last edited:
stopcrazypp
Well-Known Member
I'm talking about specific battery formats. Thin batteries (like those in cell phones) tend to be custom sized and not built out of standard size cells.
Don't have a direct link to the slides, but from a graph in a presentation on the rechargeable battery market by avicenne at the 29th International Battery Seminar the market break down for 2011 for lithium batteries is as follows:
cylindrical - 1.72 billion cells per year
prismatic - 1.55 billion cells per year
laminate/polymer - 0.83 billion cells per year
In the same slides, another graph points at 1.72 billion 18650 cell demand in 2011, but that may be incorrect labeling (otherwise 18650 would take up 100% of the market).
CapitalistOppressor's source estimates by 2014 the polymer cells would overtake cylindrical:
cylindrical - 55m/month (660 million per year)
polymer - 59m/month (708 million per year)
http://www.energytrend.com/price/20130506-5180.html
Not sure why their numbers are so much lower (lower demand in laptop market, or different market?).
But anyways, given cylindrical is still king, and 18650 is the most popular cylindrical (although I don't have hard numbers for this), it's likely that 18650 is still the top format.
Side point is that the average capacity of a 18650 cell is about 2x the cell phone prismatics (2+Ah vs ~1Ah), so it'll probably win out in kWh count too.
Last edited:
JRP3
Hyperactive Member
As I noted in the comment section I don't think it's helpful or accurate to label the chemistry that Tesla is using as "dangerous". Considering how many cells are actually in use in the world in various products I'd say it is in fact a rather safe chemistry when used properly, and should in fact be even safer than previous versions of Li Co chemistry.
Posted this in reply at GCR -
I'd just add that competence and I are only loosely related when it comes to discussing batteries at all. Still though, I should be able to recognize when my hyperbole is getting out of hand as it was in this case..
Mario Kadastik
Active Member
Question. If we take that there are in fact 7968 cells in the pack and that it takes 2000-5000 to manufacture (let's be pessimistic, let's take 5000). Then according to the spare part price (just got quoted from Tesla that the 85kWh battery is 44k eur + taxes + shipping) the cells cost 40k eur. Averaging to 8000 cells that's 5 eur / cell or about 8 usd / cell. I'm guessing that there's some margin involved though Elon's promised that his plan is not to make a lot of money on repairs, but that would mean that we're talking about the high end of the $2-$10 spectrum...
Or Tesla is charging a HUGE premium for replacement batteries. One of the two must be correct.
Or Tesla is charging a HUGE premium for replacement batteries. One of the two must be correct.
It is very likely that they are charging a huge premium for replacement batteries. With today's constraints, each replacement battery is one less new car sold. So they have to make up the margin for the lost car sale on the battery price.
Mario Kadastik
Active Member
Interesting observation. If they are indeed battery limited to an extent, then this is a valid statement and should mean that once battery limitations due to extended supply or leveling off production kick in the pack should drop in price. However that's close to impossible to model when that'd happen.
There's about 7100 cells in the 85kWh battery (which only makes your point more forcefully).
I'll need to update the OP when I get time.
Our information is much more refined now and we've put good constraints on the total pack cost even if there is some ambiguity about how that is split between integration costs and cell costs. While it's hardly definitive, there's no reason for folks to be discussing $8 cells, $40,000 packs, or the 7968 cells that were based on the 3.1aH architecture.
The numbers I posted in my last post on the subject must be at least close to reality or we would be violating important constraints that are based on public information released by Tesla -
Zzzz...
Member
What chemistry is cheaper per kWh, NiMh or Li-ion? There was an interesting presentation I've seen a while ago. It contained graph with analysis of the price depending on volume. The graph claimed that as volume grow really big, price of Li-ion drops below of NiMh.
I think that similar dependency could be very true for 2.8 Ah li-ion cells vs 3.4 Ah cells. So if volume is big enouph, 3.4 Ah cells would be cheaper then 2.8 Ah ones(per Ah or kWh). And Tesla's ordering volume is huge.
Notice that EnergyTrend data already point out to ~$180 average price per kWh for cells that are at or below of 2.8 Ah capacity. Sure this doesn't automatically mean that Tesla is already getting cells at or below of $180 per kWh, there are other factors that affect the price (limited number of vendor that could satisfy Tesla's volume requirements, strict specs and quality control, liabilities and warranty clauses adding to price etc). But nevertheless mine guess would be that at the volume Tesla ordering cells it is actually cheaper for them to order 3.1 or 3.4 Ah cells rather then 2.8Ah cells.
jeffhre
Member
I agree that its quite likely that Tesla is getting a better deal than what is displayed by Energytrend, which is why I had the "better" model in my most recent estimate, with "better" meaning "more likely".
Sorry for the late response, I have been busy with work again. I also wondered why Energytrend's numbers are lower.
I know I've seen other estimates pointing to cylindrical cell production being in the 1.6b cell range. My strong impression is that Energytrend is only looking at mainline suppliers of modern cells.
If you look at their chart, the low end is <2.8aH. While that potentially includes all batteries smaller than 2.8aH, I don't think they really mean to be all inclusive. Here is what they cite as their methodology -
"The price information provided by EnergyTrend is primarily a result of periodical survey of a pool of major manufacturers via telephone, questionnaires, and site visits. EnergyTrend cross-surveys major buyers and suppliers throughout the supply chain and strives to ensure all enclosed price information reflects actuality."
Basically, I think they are looking at the Panasonic's and Samsung's of the world (ie major manufacturers), as opposed to every generic supplier who offers 2aH cells at $1 a pop.
If my supposition is right, I don't care about those generic suppliers either, because they aren't capable of supplying Tesla's needs.
- - - Updated - - -
Also, today someone asked me how much it costs to build a new plant. Fortunately, I'd posted a recent study on that subject upthread to support our battery cost information. I'll repost now because it is somewhat critical to the current discussion -
Cost comparison of producing high-performance Li-ion batteries in the U.S. and in China
Key quote -
The upfront cost doesn't seem to be affected much by the size of the plant (on a per cell basis), but the economies of scale with a big plant (10 lines) lower the per cell cost by quite a bit.
- - - Updated - - -
Note: The study points out that the per cell cost between a U.S. plant and a Chinese plant is only ~$0.07/cell in favor of China, but that advantage is eroding as Chinese costs go up. That works out to ~$450/car on average right now, so it's not insignificant from Tesla's point of view. But there might be additional advantages that make it worthwhile, and even more so when you factor in potential costs 5 years from now.
What a tremendously important piece of info, and a bit scary if correct. Suppose for instance an optimistic 5 year growth scenario +50k TMS/TMX and +400k GenIII. That means a demand for around 50,000x7,000 + 400,000x5,000 = 2,350,000,000 cells. That translates into factory investments of close to $10bn. I think that is way higher than the investments required at the plant to reach that volume.
Another indication that cell production capacity is a core strategic issue for Tesla, not just something that will work itself out.
I agree. After posting that, I did some quick math, and a quick internet search. If you need 4,000 cells/car for GenIII, and build 500,000/year, that works out to ~$6.4 billion in capital costs using Chinese numbers.
A quick comparison of Nissan's Smyrna plant, shows that to reach it's full capacity of 200,000 packs/year will cost ~$1.7b -
Nissans New US Battery Plant Shows Major Dedication To EVs - HybridCars.com
Multiply that by two in order to simulate the requirements for a 50kWh pack. That gives you $3.4b of capital costs to support 200,000 GenIII, which is not far off of my $6.4b for 500,000 using Chinese numbers (its $8b using U.S. numbers, which balances almost perfectly).
JRP3
Hyperactive Member
Do we know the production volume of the Panasonic plant in China that they just built in 2012 for around $100 million? Just wondering if it fits with the pricing model.
Similar threads
