Elon Musk shut down internal Tesla analysis showing Robotaxi would lose money | Hacker News Reader

Elon Musk shut down internal Tesla analysis showing Robotaxi would lose money

electrek.co

115 pointsdanso55 comments

Such a shame that Tesla will need to report the Robotaxi rollout due to tariffs /s. I kid you not - this is a great excuse and they will 100% use it in some way.

It's already clear for a lot of people that the numbers at Tesla don't add up. They were able to cut a lot of (legal and accounting...) corners when they were growing quickly, but it will be much more complicated now when they are contracting.

Sure they can kick the can down the road a bit, let's say for 4 years during which US seems to be akin to a Banana Republic, but what's the end game ?

Is there anyone serious who still believes that this is a 1 Trillion $ company ? I would love to hear some serious arguments (other than projecting 100s of Billions $ from Robotaxis next year...)

Well lots of institutions hold hundreds of billions worth of the stock. It's easy to forget that the easiest person to fool is yourself.

There were equally compelling reasons for me to think that Google, Amazon, and Apple were wildly over valued in the 2000s. I was dead wrong.

When evaluating a stock's value it's always important to look for reasons you're wrong instead of minimizing contradictory data.

One data point of interest is that for the past few weeks Tesla cars built in Texas have been driving unsupervised from the factory across the highway to the end of line facility on uncontrolled roads. So far 50k miles without incident. They are also driving unsupervised at the Fremont factory.

I think it would have been very apparent at the start of 2024 that Tesla cannot compete with the Chinese manufacturers on cost. And Robotaxi is an attempt to expand beyond the luxury car bracket before someone eats that lunch. Without the article spin, the decisions could be quite sensible for the market at that time. It also helps explain the lengths Musk has since gone to to ensure Chinese manufacturers don't get access to the US market.

> I think it would have been very apparent at the start of 2024 that Tesla cannot compete with the Chinese manufacturers on cost.

But they weren’t going to have to anytime soon. The US would have had very little issue convincing their allies and trade partners to block or tariff Chinese EVs to the point of preventing them from undercutting the market.

Agreed. Ford, GM, etc. would have ensured it without Musks help. And, is that a bad thing? China et all is not exactly our friend and we should stop sending them all of our money.

Before I get brigaded, I am firmly opposed to the Trump/Musk planned theological oligarchy.

Americans, however, are our friends (as is China, our biggest and most important trading partner) and everyone benefits from trade.

Americans having access to more affordable high quality goods is a win for Americans.

I’m not sure why people don’t realize this.

The objection is that China is not a genuine trading partner that sees open and efficient markets as a good thing.

They are instead the sort of trading partner that reminds people of the British Empire: they will produce all manufactured goods and you will handle all the more easily replaced aspects of the business from raw materials to soft services.

Are Americans buying "high quality goods" from China all day?

Isn’t nearly every tech worker on this forum using an apple product ‘Designed in California’ but actually made in china

China earns relatively little per iPhone/Mac. The complex components come from other countries like Taiwan, Korea and Japan. Apple earns a hefty profit for marketing savvy, iCloud services, and perhaps financing. China makes a pittance for screwing all the parts together. Where would you prefer to be in this product pipeline?

Yep, I buy an iPhone "all day" from China.

That's kind of a straw man argument. Go pick up every single cord, speaker, lamp, etc. in your house and look at the label.

That low quality shit that lasts a year is what I am talking about. Furniture used to be handed down for generations. Now, if you move apartments twice it falls apart.

The furniture used to stay in place and net be moved around distances all that much. Due to being heavy and difficult to transport. And impossible to disassemble and reassemble.

The cost of all the tech bought in a year is not insignificant.

Back when I was in full time education, my summer holiday job was making HVAC on a production line. The transformers I screwed into backing plates came from China.

The furniature you complain of? The furniature I have comes mostly from Ikea, which is famously not Chinese.

Ikea may be famously Swedish, but are they making my ROCKSJÖN and my EKENÄSET in Stockholm?

Yes. Companies like Apple, Anker, Amazon, and DJI all manufacture their products in China.

The highest quality products in the world come from China.

Very hard to convince people to remove Chinese EVs from the market once they are there. Making EVs less affordable is not a green policy, and doesn't get votes from the other end of politics either unless you actually have a local car manufacturing industry to protect. The market is already undercut in places like Australia, and other manufacturers like Nissan and Volkswagen have already been dropping prices to stay relevant. You can now buy EVs in Australia that are not in the luxury car bracket.

Meanwhile on wall street, one sane analyst were pegging a valuation of $424 billion for robotaxis and ARK invest projected $240 billion in ebitda for robotaxis.

Creative writing > bean counting https://www.barrons.com/articles/tesla-valuation-ai-stock-pr...

Smoking crack > bean counting

  We had lots of modeling that showed the payback around FSD [Full Self Driving] and Robotaxi was going to be slow. It was going to be choppy. It was going to be very, very hard outside of the U.S., given the regulatory environment or lack of regulatory environment.

I wonder how they originally thought it would roll out?

Somewhat related, but did anyone see the EV event in China a few days ago? Have any opinions?

> or lack of regulatory environment.

I’m really curious what they meant by this exactly.

Presumably places not having any legal concept of autonomous driving?

Thanks, I don't know why I didn't read it like that. This makes sense.

1. I don't wonder, we've seen it before...

2. They originally didn't think, It was one guy's diktat...

3. It will roll out as gracefully as Cybertruck.

Elon is that idiosyncratic founder we've all met who careens between extravagance and cheapness because he can't stand looking at financial numbers and project plans.

Making money by selling products is a lot harder than by selling hype. Throw this Friday in jail.

If BYD and China are dominant in the EV (hardware) war, then the long term strategic play is to dominate the FSD (software) war. Whether or not this is a bold strategy or massive mistake is yet to be determined.

they already hobbled themselves by limiting their fsd to computer vision, no chance of dominating the fsd market when the competitors are already level with or superior.

I agree that a purely visual approach is weaker, however the new vehicles are expected to have many more cameras and using structured lighting to identify objects in near realtime. I won't count it out as that is an order of magnitude cheaper to scale if it becomes approved.

The only deployed fully autonomous passenger vehicles are robotaxis with lots of sensors operating is carefully managed service areas. There is no "FSD war."

This is actually reinforcement to my point. Getting real wold momentum on FSD software is exactly what Elon wants. He wants to be ahead in this technology curve before anyone else attempts to scale it outside of taxis.

What he "wants" is the hype to continue to prop up a pumped share price that looks increasingly vulnerable, nevermind his own experts saying it is financially incoherent.

BYD already have a FSD competitor. It's not that clear that Tesla will win that one.

Here's some funky footage of a driverless thing mucking around on a race track https://youtu.be/J_c2gsxImjA?t=53 . Not sure how genuine it is but they are definitely developing stuff. Also re their B system:

>It has been tested on real roads for more than a year and has driven across China with no human intervention in testing using a BYD Denza sedan.

Compare to Musk:

>One of the more wild claims Elon Musk has made — regarding Tesla, that is — was back in 2016 when he said a self-driving Tesla would be able to go from Los Angeles to New York City “in ~2 years” (in early 2018)

which I think still hasn't happened?

"One of the people familiar with the analysis said: There is ultimately a saturation of people who want to be ferried around in somebody else’s car."

I think this is wrong. i think Musk is right about this.

People don't want to or cannot drive, people don't want to find public transport, people just want to get where they are going as cheaply as possible.

Robotaxis have to be cheaper than normal taxis with a person driving and charging wages.

I don't know who is this people that you are talking about, but if it was the case, everyone would only use taxis already. Even if robotaxis would make the ride "half price", that would still be costly for a lot of persons, especially the one using their cars a lot.

Also you might like to own your car, to fit it to your need and style, you might need to use it to store items for period of times, you might want to be able to take your time, load and unload in private spaces, you might want to be 100% sure to have your car ready at the place that you want at the exact time and schedule that you need it, you might need your car for lot of small trips in a row without a clearly defined schedule, ...

Hi! one of those people! Live in a city with adequate public transportation that I never learn to drive and would like not to for as long as possible. To get to places I pretty much use uber for places that would take a while to reach with public transportation.

When public transit is every 5 minutes or less it's basically not a strict schedule anymore. Just leave whenever you want.

Vast majority of people just want to get where they're going on time.

There's no such thing as a robotaxi, it's just called a taxi. People who want to take taxis already take taxis.

People enjoy driving their own cars, is what I am seeing.

People who live in cities not dominated by cars are not fooled into wasting time driving their own cars. Is what I am living.

This article is more Musk bashing, but I’d like to add that this is very common in companies private or public, large or small.

CEOs and/or senior leaders get ideas stuck in their head and it becomes apparent that no amount of analysis is going to change it. Then the product/expense is more of a “we’re doing this, how do I do it the least risky/least expensive way.”

This isn’t just products. I’ve seen examples at prior companies I worked for. It’s for getting that additional jet (we got a 3rd jet, all calcs and opinions was definitely unnecessary and barely used) or putting in this additional massive plant in a non-union state to brag about how state of the art/no risk from unions we were (plant never hit anywhere near capacity just like prior analysis showed, waste of money and eventually sold at a loss.) That said, there are lots of examples where projects and products did work where analysis failed them.

This is where CEOs can steamroll their boards, especially when they are also the Chairman.

If I were in his place, I would make the small car, make a decent less edgy pickup/suv and add dashboard/stalks as an option on all cars.

(and remove the silly wide front light on the y)

I know it sounds boring, but I think more "normal" people would buy.

I mean, think about yourself - would you drive a cybertruck around town? Would you drive a conventional pickup around town? People I've talked to feel silly about driving a cybertruck.

Um, yeah, rolling out something that hasn't been done before (large scale robotaxi service) is going to be slow, choppy, and hard. What did they think the analysis would show? Ask Google about how hard this is even with a handful of taxis, let alone a million a year.

Lucky for musk he's made it even harder for himself by refusing to use LIDAR

Lucky for Musk, losing money is the real business he's in.

I'm not sure that's actually a joke. How much of the man still remains under all the hair plugs and buccal fat stripping and finasteride? How much is only a husk, the puppet of the hangers-on who are the real beneficiaries of Elon Musk (tm) LLC?

Oh, and here's a real question. Separated from those sycophants and temporarily deprived of all access to face and funds to recruit temporary new ones, how many days do we think he could survive unaided? For bonus points, identify what would be the presenting complaint if he eventually found the sense to try going to an ER.

smart move, to shut down internal dissent, on.something the rest of the industry is moving towards as fast as possible, up.to including flying robo taxis bieng given the green light in China. Given that tesla, and the varios x group have very deap resouces pertaining to navigation and automation, an internal report of iminent failure does seem very strange indeed

Not that I particularly support Musk in any way or the things they stand for, it's a great goddamn shame that they dropped the project. Even if the company in itself would not be monetarily as good as they would've liked, there's still basically an infinite amount of _value_ that they can still create.

But alas, that is not what the for profit public companies have a mandate to do.

If the maths come to about ±0 then there would be tons of goodwill to be gained by spinning up the robo-taxi project as a non-profit org.

Given the way AI is advancing it would be really dumb to not go hard on Robotaxi. That's it. That's the analysis.

If you think AI progress is going to hit a wall or Tesla's AI in particular is going to hit a wall then sure spin up a Model 2 and prepare for a long grueling endless war with BYD.

> Given the way AI is advancing it would be really dumb to not go hard on Robotaxi. That's it. That's the analysis.

Let’s say we’re headed for fully-autonomous vehicles in whatever optimistic medium term you want. I’m kinda missing the winning strategy here. Unless there’s some kind of moat to building/selling robotaxis, the sale price equalizes to marginally above the production cost, except with far less volume than what the market is selling today.

> Tesla's AI in particular is going to hit a wall

Actually at this point Tesla's AI seems to have hit lots of things: trucks, barriers, emergency vehicles, pedestrians, cyclists...

Choosing a two seater is very odd for this. Then again I’m in my 40s so maybe I’m too old

I realise that it is most improper to read the article on this website, but the analysis in question was _very optimistic_ (IMO naively so) on the robotaxi aspect, in that it just assumed that it would happen and most people would want to use it; the analysis was negative on the _financials_.

I used to be so upset at this king of bullshitters. Had paid for FSD I think a decade back and it never came. But now I look around and see he is a mere product of his environment. Nothing to get angry about.

That long ago yeah it was pretty bad and a mistake to pay for. I also bought it 6 years ago and regretted it. These days it’s actually something that people pay for. My dad tried it for a month in his new car and decided it’s worth it to him to buy. He makes the same 2 hour drive every 2 weeks to visit his grandchildren and it makes the drive tolerable. It has come a long way.

Well, 6 years ago it was also a thing people actually pay for, as you proved. Then as now, it's not what it's claimed to be

Reply on news.ycombinator.com