Sizing the worldwide battery market

[adiggs]

I've found myself thinking a lot lately, about how big the worldwide battery market can become. I've been particularly thinking about it in the context of the North American commercial truck industry and its demand for batteries as it electrifies - my initial napkin math tells me that the current worldwide battery supply is something like 1/3rd or 1/4th of what North America needs for a subset of commercial trucks, and that leads to the idea for this thread.

I bet we could, in effect, crowd source, a pretty good and evolving estimate of how big the worldwide annual consumption of batteries can / will grow to. Anybody and everybody is invited to join in and play along as much or as little as you'd like - I'm going to start by simply identifying the markets that I can think of that will need to electrify eventually. Some are more sci-fi than reality, some are more lab/prototype, and that might be part of what we evaluate (how good do the batteries need to get for electrification of the market).


Markets I can think of (no real order to them):
GROUND
- personal auto
- personal trucks (SUVs, pickups, non-commercial blending towards low-end commercial)
- motorcycles
- trikes (I'm especially thinking of the commercial uses common in Asia and Africa)
- commercial trucking (class 3-8 diesel today)
- long haul buses
- city buses
- RVs
- "heavy machinery". (i.e. front loaders, fork lifts, steamrollers, back hoes, road grinders, graders, etc... Here, I'm thinking generically of a heavy duty frame with a Diesel engine, some sort of attached tool(s), that does commercially valuable work, so there's a financial valuation to the machine and the work that it does, and thus a financial value to converting it to electric that can be made).
- agricultural tractors, combines, and equipment (the whole ag industry). big and small (somewhat analogous to the class 3-8 commercial trucking; there are big tractors and there are little tractors, and they're all used in commercially valuable ways)

WATER
- ferries
- tugs
- pleasure boats
- "medium" commercial. Thinking fishing vessels that are out and back same or next day sort of thing.
- trans oceanic cargo ships (I'm putting this in sci-fi for now, but putting it on the list to acknowledge thinking about it)
(Mostly I don't know this field at all, so I don't even know how the industry is commonly subdivided - I've read a bit about ferries on shorter routes electrifying effectively, so I know this field is starting to investigate the idea, and cheaper batteries at scale from other fields is going to help)

AIR
- planes (also putting in sci-fi; I know this is starting, but I think this isn't a source of GWh of battery pack demand in the next decade, when the rest of this is TWh scale demand)


Each of these has some steady state replacement rate in their developed markets where, if the value proposition is strong enough, I think 100% replacement of the replacement rate is a conservative estimate of what the electrified rate of production can be. I.e. - if commercial truck buyers really can save dimes/mile in fuel cost, there comes a point pretty early in the development of the electric commercial truck industry where the diesel truck industry Osbornes - the buyers only buy a diesel truck because they can't buy electric (no supply), and they can't find a used truck to tide them over until a new electric is available.


My goal with each market is to size up the current market, preferably worldwide, but significant subsets also work (i.e. - what got me started was North American commercial trucking, rather than worldwide commercial trucking). Learn a little bit about the value proposition for electrifying that market to size up how close electrification is, at least from a customer demand point of view, and how complete electrification might / will be. (If electric trucks are really dimes/mile cheaper to operate, the moment they are available at scale, anybody that isn't operating electric trucks is going out of business - nobody in the business of operating trucks commercially can give away dimes/mile to competition and last long, at least as I understand it).

As we size each market, I hope to arrive at an estimate of how much battery cell and pack demand / year each will have. Sum those up across all of the economically viable markets, and you'll have a low end estimate of the battery cell and pack markets.



My first question for folks to consider - what additional markets would you add to the list that you don't see above?

I'm going to start working up my napkin math for North American commercial trucking, mostly because I've worked it up before and because it provides some insight into why Tesla is pushing on the Semi the way they are. At least for me, it also points to what I see today as a chasm between likely demand and supply for that market, and just how much Tesla needs to grow. 50% growth sounds outrageously fast, until you see some of this math. Then you might conclude, as I'm starting to (at least in the context of battery cell and pack GWh / year output), that 100% annual growth is aiming too low.

In time, I hope that we'll get similar estimates established for all of the markets we identify.

 

[SDRick]

I think you forgot to mention a major component of the growing market that makes solar powered homes more robust? Hint, Tesla is already addressing this use.

 

[adiggs]

You're totally right - the characteristic of everything I put down is its mobile. Then there's all the STATIC applications. And it looks to me like many of these future mobile applications will bring with them static applications.

e.g. a commercial truck operator installing a static battery so it can trickle charge 24 hours a day, and then charge trucks at a much higher pace when they're ready to charge, so the truck can get rolling again as quickly as possible, while also avoiding as much of the demand charges as possible for the truck operator (all of this located on-site for the operator).

Of course, this also sounds like a nice backup for site power in case the grid goes down also, and a good excuse to add some solar panels to your big warehouse roof as well.

STATIC
- personal / home (I'd include duplexes / triplexes in this most likely, but no apartment buildings). power backup, power firming for solar install, grid defection
- medium scale (apartment buildings)
- commercial buildings. power backup, power firming for solar install. (in extremis - grid defection)
- commercial demand charge avoidance. thinking supercharger like usage, but in the commercial mobile category for charging vehicles. Big batteries will need big charging rates, and the big charging rates won't necessarily be available. So charge a static battery slowly all day, and then charge fast from the battery.
- utility scale power. renewable power firming (frequency regulation, consistent power moment to moment - variety of grid services). Shockingly small batteries enable surprisingly big grid scale renewable energy to bid firm energy into the grid instead of variable energy, improving their bid price. From what I read going on in Australia, it sounds like batteries make the fossil generators better too

I'll add the static category and see what subsets we can identify for that as well. Good catch

 

[adiggs]

Napkin math for sizing the North American commercial truck (Class 8) market.

First a note about truck classification - the systems for classifying trucks vary between Europe and North America (I'm shocked). Wikipedia is my go-to source to start learning about something new:
Truck classification - Wikipedia

Cool thing about the Wikipedia page is they have pictures of examples you'd find in each category (very handy).


Focusing on the North American Class 8 commercial truck market, my first question is the size of the current market. I've found two sources (there will be many more) to help us start sizing it:
2018 Class 8 Truck Update
(sizes it between 330k and 500k in 2018, with the most believable estimates IMHO closer to 330k)

Second source, that also includes Class 3-7 (which I'll ignore for now) that is US only:
Truck sales in the U.S. - Class 3-8 2017 | Statistic
(around 200k in 2017 for the US only)


Realizing that there's quite a bit of variation possible year to year, I'm going with 300k/year as my estimate of the North American Class 8 commercial truck market. That mostly means US + Canada in my mind, but I sort of also sweep Mexico into the bucket as well - I think the numbers don't move much whether it's included or excluded.

The biggest question that needs answering, in any of these commercial markets (at least in my mind), is the economic value proposition to the buyer of the vehicle. The easiest source of economic value I can identify is lower fuel cost per mile - there may be other sources of economic value and if anybody can put numbers on them, that'd be really helpful.

As my starting point, I'm using 2 kWh per mile (electric) and 6 mpg (diesel).

The electric efficiency is Tesla's claim (actually "less than 2 kWh"), and the for diesel, the first link I clicked on from a Google search got me this: How to Improve Fuel Efficiency on the Road (5.9mpg).

If electricity costs are $0.11/kWh (my personal retail rate in Oregon) and diesel is $3/gal, then the electric truck will cost 22 cents/mile and the diesel truck will cost 50 cents/mile for fuel.

So I'm seeing about a 30 cents/mile fuel cost advantage for the electric truck. As an outsider to the trucking industry, that sounds monumentally huge. My general impression is that the companies that operate trucks fight over pennies per mile - a technology change that brings a dimes/mile advantage is life threatening / disruptively big for the buyers and operators of trucks.


Of course there are many other questions around adoption - is the truck usable, can it actually get fueled in a fashion that can be fit into current truck routes and patterns, is it reliable. That's what the kick-the-tires phase over the next couple of years will be for. Given that all of these adoptions questions pan out, or there are routes where they pan out, and given that the electric truck purchase price is in the vicinity of the purchase price of a new diesel truck, the electric truck is going to be preferred for the fuel cost advantage all by itself.

(That's my read on things - I don't think is all that different of a competitive analysis from what has gone before).


So - this market looks ready for at least partial electrification, and if the charging infrastructure is buildable and usable and Tesla's Semi stats are real (I'm assuming yes).

So what will it take to satisfy this market? One easy assumption is that all of these trucks are long range trucks (they won't be) at 1 MWh battery pack / truck. Mostly this makes the math easy. For every 1k of these you want to ship per year, you need 1 GWh of battery cell and battery pack manufacturing capacity (1 MWh * 1000). I like easy math.

That means the North American Class 8 market will grow into around 200-300 GWh/year worth of battery pack capacity that it will consume.

 

[jhm]

Let's try a top down approach.

What we need do is replace gasoline, diesel, jet and fuel oil. 2017 demand for these were 25.6 mb/d, 27.8, 7.5, and 7.7 respectively.

Assuming that all of these use have 22% efficiency, a barrel of gasoline can be replace by 326 kWh of battery discharge, and the other fuels are replaced by 372, 347, and 405 respectively.

Lets assumed kWh of battery is replaced after 400 cycles when displacing gasoline (primarily private passenger cars) while it is replace every 2000 cycles for the other fuel (commercial use).

So in a given day to replace 25.6 mb of gasoline, we'll use about 8,352 GWh and need to replace about 21 GWh of batteries. Of the course, a year well need to produce 7,622 GWh of replacement batteries. This is just for gasoline.

For the other fuels, we'll need 1,886, 477 and 568 respectively.

Altogether, this is a need for 10.55 TWh for annual battery replacements, what's needed incrementally to replace 68.6 mb/d of fuel.

I believe I am using fairly conservative cycle numbers for replacement cycles. If you can squeeze 1000 out of an auto EV battery, that's super but it will take a long time. During the ramp up we want to replace ICE much faster than that. So we need to ramp up production capacity very fast, like 10TWh/y by 2030, but as the world has converted over the useful life will grow longer, mitigating the need to keep ramping up. Even if the demand for ICE fuel and battery substitutes grows 20%, the lengthening useful life of batteries can accommodate that. For example had I used 500 and 2500 cycles instead of 400 and 2000, the 10.55TWh level would displace 85.8 mb/d instead of 68.6 mb/d. The point I'm trying to make here is that my assumptions bake in an allowance for energy demand to grow.

 

[adiggs]

I see how you're getting there. I like this alternative approach to the napkin math - it seems like this provides us with an estimate of the upper end of the eventual size of the battery market in order to replace the mobile applications. The other approach I've been thinking in terms of will be better suited to identifying individual markets for companies deciding to invest in pieces of the transition.

Most importantly, this gives us an immediate estimate of the scale of what we have today. Here's an article I just found addressing this for all markets and uses of lithium-ion (so also including the consumer electronics folks - they count too):
Who is winning the global lithium ion battery arms race? | Benchmark Minerals

The key for this conversation:

production of 160GWh in 2019 from a capacity of 285GWh

I think that might be a misprint - I think they meant 2018 instead of 2019

As it's a little early in 2019 to say what 2019 production will be on what capacity.


Anyway, if we're thinking there's a 10 TWh / year business for mobile applications, in addition to whatever the consumer electronics business is, in addition to the stationary energy storage business - the real point of this thread that becomes immediately apparent in case anybody had any doubts - worldwide capacity isn't anywhere close to being adequate to the need / demand.

 

[mblakele]

Lets assumed kWh of battery is replaced after 400 cycles when displacing gasoline (primarily private passenger cars) while it is replace every 2000 cycles for the other fuel (commercial use).

I like this approach, but where does 400 cycles come from? I think it may be too conservative. I estimate that cycled my TMS 90D battery pack around 400 times, and it was still at 95% of original capacity — not yet ready for the scrap heap.

 

[Uncle Paul]

Not to throw cold water on this, but the original Tesla invented a way to send electrical current over the air.

Next breakthrough might be to commercialize this for the bigger cities. Every vehicle could draw power from the air to power their cars. No batteries needed.

 

[adiggs]

Not to throw cold water on this, but the original Tesla invented a way to send electrical current over the air.

Next breakthrough might be to commercialize this for the bigger cities. Every vehicle could draw power from the air to power their cars. No batteries needed.

I've seen a wireless charger in person back in 2013. At least at that time, the technology was a pad you'd put on your garage floor and then park on top of, and that would work for inducing a current in the car that could charge it wirelessly. The way I understood what I saw then, it was a demonstrator, not product I could buy in the marketplace.

I don't know of any wireless power transmission technology with a longer range or less elaborate infrastructure than that, even in a lab / prototype environment. If you know of something closer to production that we can be modeling around, please do point us at it.

 

[adiggs]

I like this approach, but where does 400 cycles come from? I think it may be too conservative. I estimate that cycled my TMS 90D battery pack around 400 times, and it was still at 95% of original capacity — not yet ready for the scrap heap.

The way I understand cycles of a Li-Ion battery, each cycle is considered to be a "full" cycle. So if you have a 100 kWh pack, 1 cycle is 100 kWh discharge and charge. Whether that happens as 10 discharge/charge cycles of 10 kWh each, or 1 discharge/charge cycle of 100 kWh. That's how I internalized / interpreted @jhm's larger comment.

So 400 cycles on a X100D would be 400 cycles of 100 kWh each, or 40,000 kWh. At 3 miles per kWh, that's 120,000 miles on the odometer. That's probably not enough - I know I plan to put a LOT more than that on my Model X, but that might be realistic once you consider improving battery technology providing a possible future opportunity to take out my 90 kWh pack, and replace it with a 200 kWh pack

2000 cycles on the commercial packs, and specifically 2000 cycles on a 1 MWh pack in the Semi, will be 2,000 MWh, will be 2,000 increments of 500 miles. Unless my math fails me, that's 1M miles on the odometer of the truck, and that's sort of what I've been using as a starting guess about vehicle life.



Tackling these assumptions, and coming up with alternative assumptions (and their reasons), is what I'm hoping this thread evolves into. I know I don't have the answers. And that my brain won't start whirling away trying to find some of them, while simultaneously being overwhelmed at all of the questions and what I don't know to think that I'll ever get all of the answers on my own.

But if we each take on bits of this that seem particularly interesting, and share our thinking, research, and conclusions, we can all get smarter without all of the effort that would be required to go it alone.

 

[RobStark]

 

[MC3OZ]

2000 cycles on the commercial packs, and specifically 2000 cycles on a 1 MWh pack in the Semi, will be 2,000 MWh, will be 2,000 increments of 500 miles. Unless my math fails me, that's 1M miles on the odometer of the truck, and that's sort of what I've been using as a starting guess about vehicle life.

Jeff Dahn recently mentioned 4 Million mile battery as an NMC variant.
We know the Tesla high Nickel battery doesn't have cobalt, so not NMC.
However, I would not be at totally surprised if a "Million Mile" high nickel battery was NMC, and Tesla used those cells for the Semi.
So what I am saying is for the Semi long life and high mileage s essential for that battery lowest cost is not 100% essential.

Perhaps unwisely, I speculated on various possibilities for Tesla battery day chemistries here:- Investor Engineering Discussions and here Investor Engineering Discussions

We know LFP will do lots of cycles, easily a million miles of 20 years of full daily cycling..

When it comes to pinning down the actual chemistries mentioned at battery day and the mileage of the cells, LFP is the only one we can be sure about.

Same deal with estimating future global demand for batteries, I tend to go with what Elon and Drew said on battery day.

 

[RobStark]

Global Passenger xEV Battery Market Doubled In 2021: 286 GWh

The global passenger xEV (BEVs, PHEVs, HEVs) battery market significantly expanded in 2021, reaching a new record level.

insideevs.com

 

[RobStark]

 

[RobStark]

 

[replicant]

Volta Foundation - Battery Report

As the most-read industry report, Volta’s Battery Report summarizes the most significant developments in the battery industry. Crowd-sourced from top industry and academia experts, this report seeks to provide a comprehensive and accessible overview of the latest battery research, policy, and...

www.volta.foundation