Tesla can’t develop an all-new car in a year
Any new car Tesla brings into volume production in 2025 can't be all-new
Tesla excited Wall Street by promising a new car going into production by the second half of 2025 -- or even earlier.
I explain why it takes at least four years to develop an all-new car, counting from the very beginning to the moment it’s in the showroom.
Four or more years is what it took for the CyberTruck for its development lifecycle to be completed -- 2019 (or earlier) to the Fall of 2023.
Tesla clearly made this decision at some point after the last quarterly report, approximately 90 days ago, possibly even this week. The clock just started ticking.
It would be possible to deliver a less ambitious variant of an existing car (Model 3 or Y) in volume production by 2025, but not something all-new.
Tesla (TSLA) traded up a whopping 13% in immediate reaction to the quarterly report, not because of the miserable financial results (automotive revenue down 13%, EPS down 47% and $2.5 billion negative free cash flow) or the nonexistent to dreadful outlook for 2024 -- but rather in response to a new product promise: A new car delivered in the next 9-18 months!
Here is the text from the earnings release:
“We have updated our future vehicle line-up to accelerate the launch of new models ahead of our previously communicated start of production in the second half of 2025.”
Then, on the call itself, Musk suggested that this new vehicle could arrive even sooner than the second half of 2025. As a result, the indicated range is around 9-18 months for this new car to arrive -- but it doesn’t matter, because 18 months or even 24 months would be unrealistic as well, as I will explain below.
This is a brand new plan
The decision to launch a different kind of new car was taken at some point between the last earnings report in late January 2024, and yesterday. That’s a 90 day window. And yet, Tesla claims it will have this new car in production by the second half of 2025.
This promise clearly got Wall Street excited, at least for a few hours, but to anyone with some knowledge of how the automotive industry works it is an absurd claim. Here is why: Depending on the precise circumstances, it takes approximately four years to develop a new car. Any variance leans closer to five years than three.
If Tesla took the decision to make this new car, about which they offered essentially zero specifics except to say that it was going to be based on a novel blend of its existing platform and the “all-new” platform it previewed in 2023, at some point within the last 90 days, then in a best-case scenario nobody should expect volume production of such a vehicle until late 2027 in an extreme scenario, and that’s would be skimping on quality testing. So, not late 2025 but late 2027 -- with 2028 a more proper target. You simply cannot promise a quality product to be in volume production by the end of 2025 if you didn’t start development already two years ago at a minimum.
The CyberTruck took at least four years
We need not look at Ford, Honda or Renault to understand this basic four-year development cycle for an all-new vehicle. Tesla itself took at least four years for the CyberTruck. It showed the vehicle in 2019, which means that development likely started before 2019, and production started in the Fall of 2023. In other words, 4-5 years at a minimum, which is what every person working in the automotive industry would tell you is standard, perfectly normal.
Why does it take 4-5 years to develop a car?
Developing a car is a lot more complex than churning out a new smartphone or bookshelf. Development involves several lengthy phases, including:
Design, including optimizing for manufacturing - this takes 1-2 years
Validation, making sure theory works in practice - this takes one year
Ordering tooling - this takes a year to arrive
Testing the final vehicle in the real world - this takes 1-2 years
Perfecting the manufacturing process - this could take a year
All in all, this is why the so-called “sketch to showroom” is a 4-5 year process, and sometimes even more if there is a new platform involved. Tesla has not been immune to this law of nature in the past, including with its most recent model, the CyberTruck. Adding insult to injury, the Tesla Semitruck and Tesla Roadster 2.0 were both shown in 2017 and are still not in volume production (Semitruck still only in “pilot” production according to Tesla’s earnings release, and the Roadster remains in the development stage). Both of these are seven years in the making, and counting. Not even four or five years.
Here is the part of the car development process that takes 1,400 days just for this design — four years if you work nearly 365 days per year:
What could you develop in 18 months or less?
It is possible to create a new version of a car in 18 months or less, that would fall way short of “all new” and a new platform-combination. It would be something like getting rid of the sunroof, offering cheaper cloth on the seats, or getting rid of the floormats. It’s what we in the auto industry call offering a “stripper version.” That might save a few dollars, but it wouldn’t be a new car requiring new design, development, tooling and too much testing/validation. It would enable you to cut the price from $35,000 to $34,000, all other things equal.
There is nothing wrong with such a decontenting exercise. Automakers do it all the time. However, there is no Wall Street fanfare or additional market capitalization associated with it.
Hopium for Wall Street
It was clear that the Reuters report from a few weeks ago, about Tesla pulling back from ambitious development programs, hurt Tesla stock. Wall Street was not content with a robotaxi unveil promise for August 8 (presumably 2024), but it wants to see a traditional metric as well, which means a lower-cost vehicle, residing under the Model 3, closer to $25,000 than $35,000.
It really should not be difficult to figure out that Tesla decided it was best to concoct a story about coming out with something new, however vague and without pictures or any prototype to show, with a 2025 delivery date promised. They will deal with any disappointment later. This was about buying time and arrest the stock from falling.
The IR guy resigns
On the heels of the CFO leaving the company in August 2023 and the lead tech development guy (Drew Baglino) leaving the company just last week, the investor relations person also saw fit to resign in order to “spend more time with the family” (literally his statement this time). He would know whether Tesla’s claims about this new car were credible, and he clearly did not want to be around to answer to the disappointment that will likely surface in 2025 when people will surely be asking what happened to this new unspecified car announced on April 23, 2024.
Robotaxi: The eternal false hope
Ever since Tesla announced full self-driving (FSD, or robotaxi, aka Level 5 autonomy) on October 19, 2016, the whole point of it was that it was going to be a software update to every Tesla manufactured since October 2016. It was going to apply to all cars, via software -- not require a whole new car to be purchased.
One can therefore argue that Tesla tying “robotaxi” functionality to a whole new vehicle constitutes a breach of promise. It is not what people were sold. Go back and watch Tesla’s “Autonomy Day” presentation from five years ago, April 22, 2019: You were told that your existing Tesla would make money for you when you sleep, by driving around performing taxi functions autonomously. On yesterday’s conference call, Elon yet again described how this would work, essentially repeating what he said five years prior. So the promise remains clear.
Yet, Tesla has failed to deliver, many years after its first claim in October 2016 and five years after “one million robotaxis by next year” made in April 2019. Wall Street ought to have run out of patience with these promises several years ago already, but apparently some people continue to allow themselves to be fooled. It is always that “next” software release that will make it work. It never does.
A scorecard for this earnings call
There will be many articles published over the next 9-18 months pointing to the extent to which Tesla was or wasn’t able to meet the promises it made on this conference call. For the reasons I described above, I estimate that most of these promises will be wildly unfulfilled.
As a result, at some point -- and I don’t know when -- this stock therefore continues to face continued underperformance risk, just like it has done since its peak in November 2021. Revenue, earnings and cash are all falling, and at some point those will take on a more acute importance on Wall Street than they already have.