Game of EV Factories

by

The battle for global EV dominance is now a game of electric car-truck factories and also a game of self-driving. The starting point is having a desirable EV design and the next step is a profitable good range EV. The financial reality and logistics of scaling many profitable EV factories is why Tesla has a lead that is insurmountable.

The Tesla Roadster was desirable but was not profitable. The Tesla Model S and X are desirable and profitable on a single-car basis. Tesla Model 3 took the whole company to profitability at scale.

Making 100k-200k EVs/year and losing $1-4billion per year is one thing but what will it take to be a major-let alone dominant player in 2025. 2030?

The Volkswagen ID3 and I expect the ID4 will be desirable cars. There is more than one desirable combustion engine car in each category and there are multiple categories. Toyota can have its Camry and Honda can have its Accord. There is more than one good selling car.

But what about Norway? Tesla had 30% market share in 2019 and then other EVs took over Norway. Norway is a tiny market and Tesla only has two global factories. There are now 50k-60k Tesla’s in Norway and Norway sells about 100k-140k cars per year. The price of the used Model 3 is higher than a new Model 3. Tesla sales to many European markets is lumpy because they have to send batches of 10-15k cars on one big container ship from Fremont or Shanghai. Tesla is selling all of the EVs they can make and Tesla is doing that without any advertising. Other big carmakers are spending $5+B per year on ads. Tesla has maintained over 80% market share of the US EV sales market.

Volkswagen is losing $4500 for every ID3 they sell. The eGolf was a relative failure. Volkswagen not expecting to make money selling EVs until 2025. If they are losing $4500 per car then selling 1.1 million per year would cost them $5 billion. GM is selling its Bolts at 12k below MSRP and the Bolts had a starting price that was not profitable for GM. If Volkswagen is selling 1.1 million EVs per year in 2025 then Tesla would still be getting its market share out of the total of 15 million to 25 million EVs that would likely be sold in 2025. Ford sold 5.4 million cars and trucks in 2019 but had $1.7 billion GAAP net income losses in 2019. It is possible to move units unprofitably for a time and to make some back on financing games. No one says Ford selling vehicles as an unprofitable company is a threat to put Mercedes, Toyota or VW out of business.

EV are not simpler to make really good. Tesla leads on the efficiency rating of batteries and drivetrain. kwh/miles/weight.

It cost billions in R&D to make a competitive electric drivetrain and control systems. It takes R&D and software coders to enable over-the-air updates of EVs. There is a lot of spending needed to be competitive with self-driving. Companies also have to be desirable to attract the best engineers, factory managers and AI and software coders. It takes money and top technical talent. The top technical talent is all going to Tesla.

Does GM-Volkswagen have over-the-air update of the software in their cars?
Do they have their own AI chip. Tesla is heading to 2nd and 3rd gen on the custom AI chip?

Volkswagen cannot really scale (convert many gas car factories to EV factories) until they are making profitable EVs. They can endure pain to be in the game at 200k-500k/year but they dare not convert 10 factories to get 5 million cars per year in 2023-2025 while they are still losing a lot on each EV. If they cut losses to $3000 per EV and go to 3 million EV per year that is $9B in losses per year plus $6-10b for the cost of the factories.

The top-selling car model in 2019 was the Toyota Corolla at 1.2 million. 900,000 Ford F-150 were sold. Most EVs sell far less than 30,000 units/year. When other carmakers create a new truck do they create a Corolla-killer or a Ford F150 killer. Do they plot to put any other carmaker out of business? They just want moderate or large success for their new model.

Getting a Big Share of the Growing World EV Pie

World 3 million EV/year in 2020 (Growing 30% per year but signs of accelerating to 50-70% per year as EV get more desirable than combustion cars). Maybe 5 million in 2021. 8-9 million in 2022. 15-25 million in 2025.

There is room for 2 to 4 companies making 5 million cars per year in 2025 and then 2 to 4 companies making 10 million cars per year in 2030. It will cost $20-40 billion to build the factories by 2025. $10+ for the R&D. Does each also have to eat $10-20B in losses as they scale?

It will take another $20-40 billion to make/convert factories to get to 2030 scale.

If it is so simple to scale EV, then why does Volkswagen target 1.5 million cars per year in 2025 and GM 1 million? Tesla will have 1 million to 1.5 million cars per year in 2021. GM and Volkswagen have to spend $30-50B and 4 years to get to where Tesla is 2020-2021. Tesla continues to innovate. Some of the China EV makers will have the government backstopping them with billions for factories.

If Tesla maintains 20% global market share and has 90% of the profits from EV, then they are like Apple versus the Androids. Apple has almost all of the smartphone profits. Tesla is growing global market share.

Tesla makes 26+% auto margin and growing. So they make $13000 on every $50000 car they sell. After $1 billion in quarterly operating expenses that is mostly profit.

They are raising another $5 billion today for 0.8% dilution. That is enough for them to start two more Gigafactories for another 1 million cars per year in 2023. Tesla will have over $20+ billion in cash at the end of this year. They will be making over $4 billion in free cash flow every quarter. This is after spending about $8 billion every year on building more factories and research and development.

By 2023, Tesla will have rolled out most of the Battery day innovations. Halved the cost of the batteries. Batteries part of structure of cars and trucks. 70% lower cost to build out factories.

Tesla can race ahead on factories and new models and massive amounts of batteries because they have 500,000 EV/year profitably made and sold.

Tesla is past the tipping point for its self-sustaining avalanche of new EVs. Volkswagen and GM want to get to the point where their EVs are each profitable and then their division is profitable at 500k to 1.5M cars per/year. Then they can start massive factory conversion. Volkswagen and GM have to do this while Tesla cuts the price of EV’s by $1000-3000 each year. Tesla uses some of their growing profit margin to make it tougher for others to become profitable.

The business moat is getting to the scale and capability for 500k/year at good profitability with two gigafactories to get the right to try to spend $40-60 billion to scale 5 million cars per year in profitability with ten gigafactories and then another $40-60 billion to get to $10 million cars per year. The cost to do this drops 70% if Tesla has battery day innovations. This would mean $15 billion to scale to 5 million cars per year and another $10 billion to scale 10 million cars and another $15 billion to scale to 20 million cars per year.

The Game of Tanks in WW2 Was Actually a Game of Tank Factories

German Panzers were individually better than the US Sherman tank and the Russian T34. There were 12000 Panzers made in WW2 versus 50,000 Sherman tanks and 55,000 T-34s.

SOURCES -Tesla, CarAdvice
Written by Brian Wang, Nextbigfuture (Brian owns shares of Tesla)

27 thoughts on “Game of EV Factories”

  1. Yes and no though.

    Car electronics, much like the ICA/SCADA industry, had been living ina closed world for a long time, allowing a lot of industry rot to seep into their design methodology and ideology. The unauthenticated nature of the CAN bus for example, means it's much like the wild west of ethernet circa 90's.

    The real lost opportunity was NTSA not forcing a standardized onboard computer platform (similar to how we have BIOS/UEFI for the x86 PC world), rather than just OBD-2 diagnostic access. That could have led to something interesting.

    I also wouldn't have a problem with wireless OTA access from the manufacturer, provided there was a means to hard disable it without a loss of functionality, and some sort of 3 firmware image OS management (factory, last version, latest version)(like 3 SD cards) so OTA updates are less scary.

    But a lot of the platform stuff got mired in weird maker tendencies and various security lock pairing with music radio units (due to radio unit theft), so now it's a mess.

  3. That's an interesting read.
    Almost enough to make you want carburettors and points in your car.
    Certainly makes me want to shut down all electronic communication to anything more critical than the GPS or radio.

  6. I think Production is over 250k annualized on MIC model 3 based on October and November. The 550k 2021 model 3 and model y in China looks solid. Some bump in Fremont 500k model 3 and y. And 60k+ on the S and the X. Especially if S and X get the refresh. More margin on S X Y.

  7. Alabama is good for that. They often offer sweet deals to car companies to encourage them to build cars there. The easiest and cheapest right-now solution is massive tax cuts or deferments.

    The voters are happy because there are 2000 new high-paying jobs in the district. The politicians are happy because they can claim these jobs as progress when it comes time to get re-elected. And the car company is happy because the tax cuts change the profit ratio in their favor.

    In the longer run it's still good for the local economy because those 2000 people with high-paying jobs are spending money in the local area.

    I'm sure there's some space near the Hyundai plant for Elon Musk to build a new gigafactory. Hint hint.

  8. "Tesla maintains 20% global market share and has 90% of the profits"
    What this tells me is that even if some other company comes along with a better mousetrap, Tesla will have already raked in the massive profits for years.

    Suppose Sparky.com (new startup) makes a better electric vehicle than the Tesla with better fit and finish, the same range, and at half the price. Tesla could start a price war with Sparky and undercut all their profits, while massively rolling out its market share. Do this for a year or two and then buy out Sparky.com.

    More realistically though, Tesla won't have any real competitors in the EV market. This is because Tesla also makes the batteries. They're bleeding edge in EV battery tech, and it would make more sense for them to sell Tesla batteries to their competitors. If Ford wants to make a fleet of EV F-150's the logical place to obtain batteries is from Tesla. That way if the F-150 outsells the Cybertruck 20-1 Tesla is still making all the profits.

    It makes sense for Ford to buy Tesla batteries and it makes sense for Tesla to sell them to Ford. Elon Musk's real goal was to get rid of the ICE, not necessarily start up a car company. If every car company buys Tesla batteries and electrifies their fleets, Elon wins. He wins the moral victory as well as raking in the duckets.

    Sparky.com would need to build a better car, a better battery, better AI, and have the deep pockets to stand up against Tesla. That's not likely.

  9. Well, the Tesla November sales in China of model 3 was about 21 600. That is about 16% given a total of about 120 000 BEVs. When the model Y comes online, this will likely double in a couple of months. A market share of 15-30% in China is amazing, IMO.

    About Europe… We'll see.

  11. I'm not sure home repairs will be that hard.

    People said that home repairs were finished when cars went EFI. Within a few years the information and the tools were available to the public and it usually turns out easier than with points and carbys because it doesn't require skilled adjustment.
    Then they said home repairs were finished when it came to hybrids like the Prius… same thing. The procedures and tools made their way out of the authorised workshops and it turned out fairly straight forward. After a couple of electrical fires the safety procedures were promulgated too…

    Remember that the vehicles do have to be serviced by thousands of dealer workshops all over the world. They can't be too difficult or the junior apprentice in the Port Moresby dealership is going to not be able to do his job.

    Of course some car companies (eg. VAG) do choose to make vehicles that their own dealerships can't work on, but they don't need to be EVs to have this problem.

  12. Yet still selling every EV they make, with no advertising and able to increase sales and outselling Mercedes and BMWs, even the regular combustion mercedes and BMWs

  13. Difficult but not costly would be cases where there is a tremendous amount of encapsulation/integration but the problem part itself isn't that expensive. That might happen more with structurally integrated components, and some skateboard EV chassis also feature extreme integration.

    I suppose there's also the false romance of private owners being able to accomplish most repairs themselves, fueled by memories of relatively simple cars using analog components that are mostly safe to touch when stopped. With all the high voltage battery components and the restricted diagnostic interfaces, it may no longer be reasonable for private owners to do work themselves anymore.

    Which loops back to well equipped independent mechanics and dealerships only, which makes the repair opaque. Which likely ends up being expensive even if it's scenario one or two as you illustrated. Either way, the only escape is to avoid the repair via 3 year lease (hoping to avoid the early death bucket), or a full on robotaxi. Which implies owning an EV long term is a fools game regardless, or an extra premium (which means it isn't initially cheap)

  14. They are investing now, but they have in mind to make profit later. It doesn't make sense that they don't. Tesla has been profitable only for the last year or so.

  15. So you are saying Tesla will have serious problems if VW and China carmakers will limit Tesla to 100% year over year unit sales growth? VW doing this by losing money on every car they make and Xpeng, Nio and Li at near zero profit.
    I think Tesla will have 200k+ sales this quarter. MIC model Y will ramp in Q1 to offset the normal Q1 drop in sales. Berlin and Texas will have significant sales in the 2nd half of 2021. If Tesla 2021 car sales are "only" 1 million that will mean more batteries to start up the Semi production or for fixed storage. Tesla has projected out 55% year over year growth to get to 20 million in 2030.
    Tesla will have to get out more models (cybertruck, cybervan, maybe three model 2s ($25k designs for Europe and China).

  16. I don't think you can have difficult but not costly repairs. Not in a situation where 99% of the customers will be paying a professional "EV" technician to diagnose and do said repair.

    1) We swapped out a number 34 black box as directed by the onboard diagnostics and it's fixed now. $25 in labour, $2500 in parts.

    2) We spent 15 hours disassembling and reassembling your car. Had to take the engine out to get to the number 4 motherboard rack. Then we spent 9 hours circuit tracing to find the problem, and lastly 1 hour to source and solder in new capacitors. It's all good now. $25 in parts, $2500 in labour.

    No customer is going to care about the difference.

  18. Fit an finish , at least in the past, seemed to imply either a highly refined manufacturing system (favoring experienced legacy automakers with long term experienced line workers), or a brute force approach using a lot of people (this in the past favored makers who literally throw an army of people at the problem, enabled by low wages).

    Tesla neither has the long term institutional memory of an experienced line, and US wages prevent throwing people at the problem.

    Though how much those assumptions get thrown out the window when a highly automated and highly adaptable manufacturing line staffed with unfixed general purpose robotics, AKA Elon's alien juggernaut, remains to be seen, as nobody has demonstrated that kind of flexibility yet. Particularly as Telsa makes so many active changes on their lines. Human flexibility is still winning there.

    A CEO recently said that current automation means a highly automated line is most suited for products that are not expected to change, and can see the end of production on the horizon, thus the drive to squeeze out the last bit of profit before the end comes.

  19. You can fly as many numbers as you want. I have just bothered to look at one. Here is a honest forecast for Tesla projected car sales for 2021, Optimistically, 1M.

    https://oilprice.com/Energy/Energy-General/Tesla-Is-On-Track-To-Deliver-1-Million-EVs-In-2021.html#:~:text=Meanwhile%2C%20Tesla's%20new%20Shanghai%20factory,of%20840%2C000%20deliveries%20in%202021.

    With other electric car companies already dominating other major car markets, Tesla Sales outside the US will not grow so fast. It didn't happen in China, despite the Shanghai factory.

    https://www.forbes.com/sites/jimcollins/2020/11/10/teslas-market-share-is-sliding-in-china-and-has-cratered-in-europe/?sh=5f3b407536d7

    Nobody is a fool. Whoever wants a share in the global EV market, will invest its money wisely to play its strengths.

  20. I'd also argue that people wanting cheap reliable cars are already hobbled by a lack of money (inability to do lump sum up front and/or monthly expense limiters) are specifically searching for reliable transportation, and not cars per se, due to a car travel dependent environment as well as the economic mobility provided by having the ability to work far from home. If the reliability can be reasonably substituted through leasing, or substituted with a robotaxi service, then the actual costs for the car (including battery replacement, which is a big cost center) start to become less coupled.

    What are the functional differences between an easy to repair but costly to repair car (lots of swappable blackbox submodules), versus a difficult to repair but not that costly to repair car (cheap components but physical arrangements make repairs difficult)? That assumes a car can no longer be cheap and easy to repair, either due to regulatory issues or makers deliberately interfering with the ease of repair.

  21. Nothing wrong with consumer grade parts IF they are easily replaceable, or are used with consideration for longevity. For example the recent internal computer early death issues stem from 2 basic issues

    1. The OS design was stupid, not making efforts to reduce writes to poor write life media, and dumping huge debug logs. This is rookie grade embedded system OS engineering.

    2. Not making the log storage or other high write activity exclusively to an easily replaceable storage media. They were hammering their soldered on eMMC, when a removable SD card would have been a wiser choice. It's not like they are averse to this, as their sentry security system video logs can be dumped to a USB drive to improve storage.

    I think the bigger issue is that these cars, with the advent of regular OTA networked systems, means you need to design and manage things as an end user client fleet with an appropriate server backend, as services revenue comes to the forefront. Tesla has repeatedly shown they are not up to snuff regarding backend services, and if the allegedly stories from a backend IT guy on Reddit are to be believed, the backend is a shitshow waiting for an epic meltdown. In this respect, a major cloud company with experience in modern cloud service management is better equipped to handle the services side of the equation. Say Google, or Apple.

    Thus the looming threat of the iCar.

  22. They will go for loans from governments, they will leverage handouts from states, they can go to the banks, and they can dilute shares if necessary.

    Why get a loan if a state is willing to help to get the factory built in their state? "Free" money is the best kind of money.

  23. Tesla's quality (per Consumer Reports) is second to last in the luxury market.
    Model Y is new and looks nice but the quality is poor. VW might actually beat Tesla when it comes to EV quality which is mind boggling.

    Tesla needs quality cars that last. For example cars not built with consumer grade electronics. Once EVs become mainstream then Tesla will need to compete against cars that are better built and the lesson of American auto manufacturing from the past is that people want reliable cars, in particular people who are looking for cheaper cars who don't have money to spend on large repair bills.

  24. Two interesting things. Even with the 69% reduction in Capex for building battery factories, Tesla would still need about 90 billion dollars to build a 3 TWh per year capacity (=their stated goal). At this time, Tesla is investing about 10k USD per car-year-capacity, so getting to 20 million cars per year would require about 200 billion USD.

    Let's make it an even 300 billion USD of total needed investments.

    Is Tesla going to get loans for 300 billion USD or are they going to dilute their shares? My hopes is for cheap credit; I don't like my shares to be diluted to that extent. But will they get that kind of credit?

  25. Which car company will make these reliable EV cars at scale and when will it be? Tesla iterates and improves on quality mainly when they re-tool and make new factories and new lines. They had problems in the trunk of the Model 3. Bad fits and bad design with many parts. The new design is to make a megapress to form it all in one or two pieces. Paint had problems and the Berlin factory will have top of the line paint shop. The Cybertruck will have no paint. Tesla is not a stationary target. Tesla does not have to be perfect and win every sale. EV are 3% of all car sales and heading to 10% around 2024. Tesla has 20-25% of global EV market. Getting to 30% of global EV market in 2024-2025 would be huge. That means 90% still pick combustion engine car in 2024 and 70% still pick a non-Tesla. Would Tesla lose sales first or would it be a combo of some of the hundreds of other EV makers?

  26. I agree. Lots of things break down on modern cars, and it's by design so that car manufacturers can recoup profits on the spare parts. That's why there is no standardization for seat belts, seats, windows, etc. All is custom made for each model so that there will be no competition and lowering of prices.

    Things may change, though, when Tesla will be designing cars for their own robo Taxi fleets. At that time, it makes no sense to design the movable parts to be expensive and short lived.

  27. Wouldn’t step 3 be building reliable cars, or are we going to repeat the 70-80s all again? Where US auto manufacturers (Tesla) make pretty, sporty cars and get their lunch eaten when somebody else makes reliable cars?

    Who cares if your battery lasts for a million miles if the car falls apart at 100k? I sold my last car with the original engine and transmission because after 320k miles I didn’t want to replace all the other stuff that was wearing down.

Comments are closed.