Thread by @KawasakiKR11

1/
Last week, @subach found a video in which an analysis estimated $TSLA's battery cost to be ~$240/kWh, citing METI (Ministry of Economy, Trade & Industry) data. So, I looked into it.

TL;DR: He's probably right.

$TSLAQ

https://twitter.com/subach/status/1091150643635634176

2/
METI is the Japanese equivalent of U.S. Commerce Dept., & collects tons of data, including automotive Li-ion battery production/sales. I'll link to specific files at the end.

(Note that 2018 data are for Jan-Nov, no Dec. data yet.)
meti.go.jp/statistics/tyo…

3/
Here are the data, plotted yearly. The raw data are in Ah & JPY, so I converted them to kWh & USD. There are a few take-aways.

#1 is the sales value (cost) never goes below $200/kWh. That's true even if you look at the monthly data.

4/
#2: If you followed the changes in raw material cost & exchange rate during this time period, it's not too hard to understand why the price decline has stopped.

Here are the numbers for the plot above.

5/
Here are the data from USGS on lithium carbonate & cobalt. You can see the prices spike in 2017. As you see next, the peak continues to early 2018.

6/
Lithium price remains rather high, while cobalt price has come down, as you can see in the data from @benchmarkmin & InfoMine.com. These are not exactly the price Pana pays, but should be reasonably close. I'd love @benchmarkmin & others to comment on this.

7/
Simplified story is like this:

2012-2015: the price declined as production volume increased, helped by weakening JPY.

2016: strengthening JPY attenuated the price decline.

2017-2018: the spike in raw material costs prevented battery cost from declining further.

8/
Admittedly, the data are for the entire automotive Li-ion battery production/sales in Japan. However, Pana's share should be over 70%, as the analyst mentioned in the video. Here is my estimate:

9/
How much do batteries cost to Tesla & others?

Here is the simple math based on 72% share. I find ~$220/kWh for Tesla to be reasonable.

10/
Nearly 10% are for Leafs, & their cell cost was reported to be under $300/kWh in 2013. The figure is from a paper in Nature CC in 2015. I noted the number that really can't be true.

11/
Looking at the figure above, you'll note that no estimate below $150/kWh in the next 10+ yrs. Here's the estimate from UCS. Except for 2030, similar story.

12/
If you look around, one of the main sources of low battery cost estimate is BNEF, which, as a part of @business, seems to like to shill for Tesla.

13/
You can find similar graphs floating around, many made by known shills @InsideEVs, @cleantechnica, etc. I don't think they have a clue. I'm not linking to them, because they are all unicorn farts & rainbows.

14/
While this applies only to 18650 cells for S/X, I doubt 21700 cells for 3 are all that different. After all, Pana/Tesla has a long way to go to pay for GF1, unlike the Pana factory in Japan that's been making cells for >7 yrs. As of now, those 2 factory are similar in size.

15/
In sum, these data suggest battery cost <$150/kWh to be way off the mark.

Good news for Tesla? At >$200/kWh, Tesla won't have much trouble meeting the purchase obligations to Pana.

fin/

p.s. source data on battery production/sales:

Monthly data (last 5 yrs):
meti.go.jp/statistics/tyo…

Yearly & monthly data for each year:
meti.go.jp/statistics/tyo…

exchange rate:
investing.com/currencies/usd…

Lithium price:
minerals.usgs.gov/minerals/pubs/…

Cobalt price:
minerals.usgs.gov/minerals/pubs/…
infomine.com/investment/met…