Tesla Battery Cells Are Best Value In Class: 20% Cheaper Than LG Chem

NOV 21 2018 BY MARK KANE 36

Those Tesla/Panasonic cells are simply more affordable.

According to UBS analyst Colin Langan, the teardown of lithium-ion battery cells produced by Panasonic/Tesla, LG Chem, Samsung SDI and Contemporary Amperex Technology (CATL) revealed that Panasonic/Tesla cells are some 20% more cost efficient than the next best offer (LG Chem).

It’s estimated that the Panasonic 21700 (2170) type cylindrical lithium-ion cells cost $111/kWh, which is $37 less than LG Chem ($148/kWh). If the other manufacturers are at even higher prices, the advantage is tremendous, although we are not sure how much cheaper is to make modules/packs with other types of cells (the overall difference on a pack level could be smaller).

Anyways, 75 kWh of batteries should cost Tesla $8,325 or $11,100 in the case of 100 kWh – before they start building the pack, as we understand. In the case of LG Chem, it would be $2,800 or $3,700 more respectively.

UBS expects that the four mentioned manufacturers will control 70% of the EV battery market by 2025 and that costs will fall by 10% within 2-3 years.

Source: thefly.com, ft.com

Categories: Battery Tech, Tesla

Tags: , , , ,

Leave a Reply

36 Comments on "Tesla Battery Cells Are Best Value In Class: 20% Cheaper Than LG Chem"

newest oldest most voted

“and that [battery] costs will fall by 10% within 2-3 years.”
Does batteries not fall in price by 10-20% each year anymore?

It does, but UBS has l-o-t-s of oil interests to satisfy 😉

In short, these numbers are not too far removed from fiction.

The rapid drops in cell prices over the past 8 years may be coming to an end sooner than we like. The price for Panasonic and LG will probably drop more like 15% over the next 3 years, but I doubt they will drop more than 20%.
And these estimates aren’t crazy talk. $111 for Panasonic and $147 for LG Chem cells aren’t bad estimates. They sound a bit high but not out of the ball park. Panasonic at $98 and LG at $125 maybe?
3% to 4%, (maybe 5% this year) improvements per year may be all we get for Li Ion packs going forward. And realistically, that will be good enough.

2-7% annually has been the history.

Actually, the recent history says it drops >= 10% per year.

This doesn’t mean it will keep dropping like that forever, but it provides a good estimate of the price reduction rate.

Actually, this depends a lot on further technology breakthroughs opening new potential for further price reductions… Without that, the price drops should slow down considerably as they asymptotically approach the limit of existing technology.

I think you are confusing density improvements with price drops… 5% – 7% has been thrown around for density; for prices drops, it has been much more than that over the past years. (Estimates vary, but most recent ones seem to be in the 15% – 20% range…)

Null — There are two factors in battery pricing.
1) Decrease in per-unit cell pricing for each battery cell. So if you buy 5000 battery cells, the price for those cells has dropped.

2) Increase in storage capacity for each battery cell. So if you buy 5000 battery cells, and each cell stores more energy, the total amount of energy stored is higher for the same 5000 battery cells.

In order to calculate cost per kWh, you have to account for BOTH factors.

Historically this has produced about a 14% improvement in cost per kWh on a CAGR basis when both factors (per unit price drops and per unit increase in kWh storage capacity) are accounted for. My source is calculations based on Panasonic’s publicly available documentation.

With that said, Tesla/Panasonic are beating the rest of the market.

Another Euro point of view

And it is enough to tear down a battery to come up with a fabrication cost ? Most probably not, I take it they have other information like about raw materials supply contracts, volumes produced, amortizing cost of machines etc. otherwise it would be a meaningless exercise. I wonder how they do collect these kind of information which battery makers could be reluctant to give away.

And since Tesla owns the Gigafactory where Model 3 battery cells are manufactured, the overhead costs of running the factory (allocated to portion of the factory dedicated to battery manufacturing) should be added to the cost of batteries. These costs include depreciation of the building and manufacturing equipment, property and other taxes, maintenance, etc.

Another Euro point of view

Indeed, and I take it there are rather complex agreements in between Panasonic and Tesla as regards who takes which cost in their P&L. I understood Panasonic owns the machines and Tesla indeed owns the building. Also, part of the cost includes Panasonic’s profit margin. How does UBS knows how that accounts into the total cost per cell, did they just call and ask ?

They are estimating costs, not prices.

Audi for example claims they are buying cells for the e-tron Quattro at $114 per kWh. According to the UBS analysis, it costs LG $148 per kWh to make them — meaning they would be selling them at a significant loss…

Not sure I would trust that. The estimates are probably more reliable in terms of *relative* cost differences than absolute costs.

Obviously UBS is taking these costs into account. Their information is likely to be incomplete — but they aren’t dumb…

Okay, they aren’t dumb; but neither are they omniscient. The point is that the best they can do is estimate costs. Without access to Tesla’s internal reports on their actual costs for various things, how accurate can any of these estimates be? Not very, I’d say.

I remember what a surprise it was when it was inadvertently revealed that the price (not cost) LG Chem gave to GM for its EV batteries was only $145/kWh. At that time, it was a price far lower than what was being cited as the average industry cost. I think that shows just how vague these guesses are; the guesses that outsiders are making about battery costs.

“And since Tesla owns the Gigafactory where Model 3 battery cells are manufactured, the overhead costs of running the factory … blah, blah, blah, etc, and more blah” Short answer is: No. Long answer: It’s implicated due to transfer pricing, cost accounting and related economics, etc. Short answer: There are main 3 ways (assuming mutually exclusive comprehensively exhaustive argument) a company controls an asset: 1) it owns the assets and depreciates it if it’s a long-life asset like a plant 2) it can lease (rent) the plant, e.g. the company is the lessee or renter 3) it can use sub-contracting, here get access to the assets by using subcontracting, e.g. get someone else to produce the batteries. Under financial theory (built on economic theory) the three costs should be the same (in the long run) because the company wants to minimize its cost. In reality, taxes, financial engineering, etc creates distortions. So basically it’s six of one and half dozen of the other — in theory, it should not matter if the company owns the plant, leases, or sub-contracts out the production. This is what throws most people off but is something that is stressed in basic financial statement analysis… Read more »

“… blah, blah, blah, etc, and more blah”

Nothing to say, but saying it at great length? Apparently “Jason” is blind to irony.

Many of these things can be estimated pretty well from publicly available information. Raw material prices on the spot market are well-known, and can be assumed to be roughly similar in long-term contracts. (Or they can be derived from the financial statements of the producers…) Production volumes are well-known from Tesla’s own statements. Price quotations for the machines in question can surely be obtained easily from the vendors.

Of course there is still some margin of error. Last time I heard for example there was no reliable public data on the energy consumption of battery production, with estimates varying by an order of magnitude…

It’s interesting that their LG Chem number contradicts GM’s own data for cells purchased from LG Chem. This study claims that LG Chem cells are $148/KWh yet GM in 2016 were saying that they were getting LG Chem cells at $125/KWh with the price expected to fall to below $100/KWh by 2021.

As an additional note this study appears to focus on just the cells and does not address the packaging and assembly differences using different battery cell formats.

Yup, if I recall correctly $148/kWh was the number leaked by GM back in 2015/6 when the Bolt was only an announcement. It is very very likely, quite a bit lower now.

The number leaked was $125/KWh

I think Assaf is right, he usually is, as I recall, But then one could just look it up.
But I suppose it’s easier to just type in “You’re Wrong.”
So it was 145, not 148, but Assaf is very close.

ffbj – Chill mode now (and yes, Assaf is pretty close most of the time) 🙂

The price they pay should actually not be lower for several years, by GM’s own presentation. They apparently entered a fixed-price contract…

Note that the price they pay is not the same as production cost.

Except that, according to Jay Cole some months ago, LG Chem and Samsung got into a price war. So LG Chem may have had to lower its price to GM even though they apparently had a 3-year contract with GM for a price fixed at $145/kWh for the three years.

But that’s just speculation based on the fact that the revealed slide, which showed the $145 price, also showed the price holding steady for 3 years. Perhaps there was no fixed price in the delivery contract.

At any rate, various outside estimates are that GM is indeed paying LG Chem less than $145/kWh now, so I guess that is true. But since it’s only a guess, maybe it’s not.

This report or at least these excerpts are completely meaningless. LG and Samsung both sell all sorts of different battery cells for automotive applications with regard to chemistry and form factors. The batteries used for hybrids and EVs with small batteries are different from the ones used for EVs with large batteries.

The UBS studies I’m aware of have all been about long-range BEVs (Bolt, Model 3 etc.), not any PHEVs or short-range BEVs.

Actually the price doesn’t tell us which cells are better value.

True it’s a value judgment, and using judgement they are a better value. Here Come the Judgement.

The “better value” equation on the consumer end of the automotive EV battery pack, is dependent on a number of external factors, which go beyond the type of cell format, battery type(chemistry), cost, lifecycle expectancy/replacement(size/kWh), from the four largest (Panasonic, LG Chem, Samsung SDI, and CATL) EV battery OEMs.

These variables will be interesting to observe, as the major ICE players start to promote their potentially different degradation warranties (8 yr./100k mi.) for their large volume competitive mainstream EV offerings in 2020-21.

You also have to account for pack costs. Having cheaper cells doesn’t mean much if the packs are more expensive.

UBS has analysed the pack costs before; the cell cost analysis just fills in a large unknown, whereas previously they went by rough guesses based on isolated and outdated bits of data…

Banks don’t understand much about batteries … Just throwing it out here. Not what I would consider a trusted source

Their “evidence lab” has been doing this sort of thing for years. For their Bolt analysis, they worked with Munro for example. (And presumably also on their Model 3 analysis…)

I wouldn’t trust their conclusions about absolute costs too much (there is simply to little data to verify assumptions); but I suspect the relative costs of different cell types can be estimated reasonably well…

I seriously doubt they can estimate with any accuracy just how much the Gigafactory 1’s vertical integration reduces costs. I also doubt they have accurate figures on how much Tesla is paying for the materials used to make the cells. As has been said, many or most of the figures they’re using are going to be mere estimates. No doubt many estimates are highly accurate… but some others won’t be.

What I’d love to see is an analysis of just what the margin of error (or standard deviation) is in these sorts of estimates. My guess is that it would be high enough to make the exact figure cited in any estimate fairly meaningless.

Elon Musk said in a recent Q&A that Tesla hoped to get battery cell prices down to under $100/kWh by the end of this year. Presuming that is true, then I rather doubt Tesla’s current prices are as high as $111/kWh at present. A more-than-10% drop in a very few months seems unlikely. (Certainly not impossible, but unlikely.) I’d guess Tesla’s costs are already closer to $100/kWh than the estimate here.

I really don’t know how they conclude about that big difference. It seems very unlikely – Tesla is not selling cars for cheaper, doesn’t have big profits, and said price for LG batteries is very close from what I can get for 18650 cells from Panasonic, LG, Sanyo or Samsung.
Also it’s very hard to know the real price as probably many are losing money selling them.