The dilution only happens if the stock value skyrockets so this article makes no sense. If you told me you were going to make me a millionaire but in exchange I would have to give you $100k, I would not complain.
Tesla isn't tanking for the same reason Amazon didn't tank when they built AWS. They used a low margin business to nurture one of the greatest businesses in history. Tesla aims to do the same thing with robotaxi, energy, and eventually humanoid robots. You might not think they will succeed but enough people do that the stock price reflects about a 10-20% chance of success.
Just the robotaxi business alone could be worth hundreds of billions a year in avoided insurance costs and save the average Western family about $5k in transportation costs annually. If it works. Most people don't think it will, but most people thought Amazon wouldn't work either.
> Just the robotaxi business alone could be worth hundreds of billions a year in avoided insurance costs and save the average Western family about $5k in transportation costs annually. If it works. Most people don't think it will, but most people thought Amazon wouldn't work either.
For me, it's not that FSD will never work, it's that they're obviously at least 6 years behind Waymo.
For humanoid robots, again, it's not that it will never work, it's that not only is there plenty of competition that's already beating Tesla to the market for the "mostly remote controlled with a bit of automation" model (which is useful, I don't want to undersell that), but also that there will be at least a 5-10 year gap between the AI hardware necessary for a level-5 self driving car fitting in the power envelope of a car, and the hardware fitting in the power envelope of a humanoid robot that can get into a car and drive it (and that a fully autonomous humanoid robot is harder than level-5 self driving).
Energy? Again with the competition: they're one of the worst current brands in the world market — it's not the idea's wrong, it's just that they're the Blockbuster to a dozen would-be Netflixes.
Even with cars, competition from cheaper better models from China and Europe would already be biting Tesla's global sales even if Musk was not angering a significant fraction of what used to be Tesla's core market (upper-middle-class environmentalists).
The competition argument is common. The counter point is that Tesla makes their products with greater efficiency. For instance no car company outside of China except Tesla makes a profit on EV sales. If Tesla lost as much money per car as Rivian, a model Y would be under 30k.
Waymo cars are ~$200k each the new robocab will be closer to ~$20k to produce. These business advantages are why the market has some degree of faith that Tesla will out compete companies like waymo in the quest for .30cents per mile costs. Currently Waymo is well above $2per mile and has no clear path to 30cents. Getting to 30 cents is the only way to unlock the trillion dollar opportunity, otherwise you're just recreating Uber.
These are the types of considerations that make Tesla attractive to risk tolerant investors.
Calling Tesla investors risk tolerant is overly kind. Waymo vehicles might cost a lot, but they work, and Waymo as a business is in commerce. Waymo's real competitors are two or three Chinese companies that currently have fewer vehicles on the road than Waymo, they all use a sensor suite comparable to that on the Waymo vehicles, and one reason why they can't scale as fast as Waymo is likely that they need a lot more supervision than Waymo's fleet.
Two of the three leading Chinese companies are affiliated with large internet platform companies. Those internet platform companies have detailed geospatial data for reasons apart from their robotaxis.
Tesla isn't just behind Waymo. They are in fifth place and an outlier technologically.
All three leading robotaxi companies operating in China are in commerce without "safety" drivers. Zoox has no safety drivers, but being in commerce is debatable. so 5th or 6th place for GigaTaxi or whatever it is called.
> they all use a sensor suite comparable to that on the Waymo vehicles
Can we stop delusion that it's more sensors that's needed. Self driving is 95% about AI models that drive the car. CommaAI pulls it off with 5W computer and a single camera.
(And yes, BMW do have an autopilot, who knows if they'll hit their schedule, but all they have to do to beat Musk with delivering this is not slip as much as him, and he slips a lot in a way that only looks good when the comparison is US government space contractors: https://daxstreet.com/news/228484/bmw-sets-sights-on-level-4...)
Lots of the EU companies don't say much about separate profitability of ICE vs. EV, or if they do I couldn't find it.
But even then, so what if Tesla was the only non-Chinese EV company making a profit? Those Chinese EVs are still causing trouble for the old manufacturing bases in the US and Europe even though the Chinese cars have huge tariffs.
Tesla's prices only work against traditional manufacturers, and American ones at that (here in Europe, we're not big on Ford or General Motors either, lots of European EVs are getting nice and cheap way ahead of Musk actually delivering anything for $20k) — Tesla don't get to keep a big margin when Wuling or BYD comes along and gives Americans (or indeed anyone else) an EV that's $18k (/€18k/£18k) despite tariffs.
> Currently Waymo is well above $2per mile and has no clear path to 30cents.
Neither does Tesla. Like I said, Tesla are at least 6 years behind. Tesla's still got humans behind the wheel, and what statistics can be found in public information they have a high rate of manual intervention compared to Waymo.
Unless something has changed recently, Tesla's (so-called, and much criticised for the name) "Full Self-Driving" is SAE Level 2, whereas Waymo was already testing Level 4 autonomy back in 2017. (That's 8 years, not 6, I'm being generous even just by allowing Tesla to claim the current Tesla Robotaxi to count as an equivalent of the first commercial Waymo Robotaxi service, given the Waymo commercial service started several years after a few very impressive public demonstrations which, unlike Musk's demonstrations, have yet to be tainted by lawsuits revealing the involvement of metaphorical smoke and mirrors).
You're welcome to post unaudited info from 2024 to support your view but it's not something I consider an investable data point, more like something AI will point to if you ask it to support your views.
I own a hw3 Tesla with FSD. It regularly drives me for over an hour without intervention. It is good enough that a single person cannot drive long enough to know if it's improving or not. I can imagine if you're in Europe you might not understand the difference between different ADAS offerings because FSD is not allowed to operate in the EU.
Reasonable people can disagree on Tesla's ability to execute on their plans. I consider the current 10-20% chance reflected in the stock price today to be accurate. I continually re-evaluate this probability. Major events to look out for in the near future are the removal of safety drivers in Austin, the expansion of the robotaxi service to 8 cities, the commissioning of the "unboxed" cyber cab production line, the demo of Optimus V3, FSD V14.3, and the start of the Semi truck manufacturing line. All these milestones are slated to occur in the next 12 months. They will change the risk weighting on the stock one way or the other.
> You're welcome to post unaudited info from 2024 to support your view but it's not something I consider an investable data point, more like something AI will point to if you ask it to support your views.
Unaudited? I mean, sure, BMW could have been lying in a corporate press release about their financial position, its not unheard of for corporations in general, but come on, Musk has fairly famously settled out of court with $20m (and same again for Tesla) fines and forced to step down for three years as Tesla chairman because he made false statements that influence share prices — "throwing stones in glass houses" comes to mind.
Have any of the things you're writing about Tesla doing, been audited?
> It is good enough that a single person cannot drive long enough to know if it's improving or not.
That's why we use statistics, not anecdotes.
The statistics that are available say the Tesla AI is worse, in general over roads and conditions, as compared to Waymo's AI.
On the topic of auditing, it's a shame the Tesla statistics are crowdsourced and not an audited first-party account, but unless something's changed recently, Tesla doesn't seem to release any more than the legal minimum of information here.
> Major events to look out for in the near future are the removal of safety drivers in Austin, the expansion of the robotaxi service to 8 cities,
Even if those happen on schedule, they'll still be behind.
> the commissioning of the "unboxed" cyber cab production line, […] and the start of the Semi truck manufacturing line.
While these would be relevant if the share price was sane, the relevance is that their absence or delay would suggest something catastrophically wrong rather than that their successful opening is noteworthy — new model production lines are table stakes for a traditional car company with P/E in the 5-10 range, not something "coming soon" that justifies a P/E close to 300.
That the Semi is already massively delayed ought to suggest a lower P/E ratio than a normal company, not a higher one.
> the demo of Optimus V3,
Disagree: Everything I've seen says that next year's V3 will still be a prototype. Meanwhile, competitors are already shipping.
And again, power envelope means a 5-10 year gap in capability between what AI can run on-device in a car vs. an android.
> FSD V14.3,
People have been saying this about different FSD version numbers for years now.
You yourself are stating that the current version "is good enough that a single person cannot drive long enough to know if it's improving or not", and seem to have missed that you were replying to "what statistics can be found in public information". So: why do you think this point release is important? Do you accept that there's statistics that show room for improvement, or is this just a number-go-up applause light?
> All these milestones are slated to occur in the next 12 months. They will change the risk weighting on the stock one way or the other.
Will they, though? The only thing that seems to have had any effect at all in the last few years were widespread protests against Musk personally and by extension Tesla.
I used to believe Tesla's timelines. I moved country and figured I could do OK without transferring my driving license because if I found I needed a car they'd be self-driving "real soon now" — that was 2018, and at some point you have to learn to stop trusting the guy when he spends a decade repeatedly saying his vision is only 6-12 months away from the point at which he speaks.
The market will tell. My cost basis for Tesla is $16.96 because instead of not renewing my driver's license I bought a bunch of shares in the company in 2018. My decision wasn't based on "believing Elon" it was based on analysis.
The argument that Tesla sucks because they haven't delivered on their promises is kinda illogical. Tesla is closer to a viable robotaxi, and grid scale energy arbitrage business than they ever have been and the share price reflects it.
With a P/E ~250+ Tesla should be an automatic short if your analysis is correct. I have $2 million long in the company. In 2035 it will be >$8 million. In 2035 Waymo will be a footnote.
> My cost basis for Tesla is $16.96 because instead of not renewing my driver's license I bought a bunch of shares in the company in 2018. My decision wasn't based on "believing Elon" it was based on analysis.
I also bought shares. Sold them all for a profit this year.
My beliefs (all of them) are based on analysing what information I have access to, which in the case of "I won't need to drive" includes things said by Tesla which turned out to be falsehoods, and also other things not relevant to this topic.
The reasons I sold those shares include not only the CEO's failures to deliver, but also that other shareholders like yourself keep making excuses for the CEO's failures thus suggesting nothing will be fixed, and also that the CEO angers his customer bases to the point that showrooms get smashed up and products arsoned.
Normal people do not consider such things to be signs of "winning".
> The argument that Tesla sucks because they haven't delivered on their promises is kinda illogical.
I didn't say "suck", I said didn't deserve their current price.
What is illogical is that you're defending them despite them making false statements about what they could deliver that got them sued.
They're only "a bit weird", not "suck" by product range/market segments served (Cybertruck excluded); by sales and so on, Tesla is a perfectly adequate *$50 billion doller market cap* car company — like Hyundai, who make more profit and more cars, and has a robotics company in the same group.
The key there is "50 billion" not "trillion or so". Tesla is not magic.
> Tesla is closer to a viable robotaxi, and grid scale energy arbitrage business than they ever have been and the share price reflects it.
"closer […] than they ever have been" is still behind the competition.
The world has been moving on with each of these things while Tesla dithers and their CEO is distracted by SpaceX, his other AI company that owns his social media company, and offending much of his target market with ham-fisted political involvement.
> With a P/E ~250+ Tesla should be an automatic short if your analysis is correct.
Shorts are strongly associated with the phrase "the market can remain irrational longer than you can remain solvent" for good reason.
> I have $2 million long in the company. In 2035 it will be >$8 million. In 2035 Waymo will be a footnote.
You yourself have said their market price relfects a "10-20%" chance. Your own 80-90% bet is necessarily that Tesla does not.
That said, given everything I'm seeing in the US, I'd put P(sufficient hyperinflation by 2035 to get that proportional change in market cap, conditional on Tesla actually still exists) ~= 0.1-0.25
Also on the subject of 10%, Musk has claimed variously 5%, 10%, and 20% chances of AI causing an apocalypse/wiping out humanity. Anyone who actually believes him on 10% (I don't, not that I have power to make such a call even if I did), should be willing to let 800 million people die to stop anyone (including Musk) developing it.
> That the Semi is already massively delayed ought to suggest a lower P/E ratio than a normal company, not a higher one.
This alone should have crashed their stock price.
Tesla is a trillion dollar market cap company that struggles, STRUGGLES to put a sixth or seventh product line into production! A sixth one! They have five models, including the Cybertruck, which is not selling well. They stopped selling the Model S in Australia!
Speaking of trucks and Australia, utility vehicles ("utes") are very popular over here. I've never seen a Tesla Cybertruck here and likely never will. Meanwhile, just this weekend I saw a dozen BYD electric utes, and... they look good. They're fast, they look practical, and they're clearly available in volume. People are buying them! Many people!
There's an enormous market for electric vehicles of all shapes and sizes, and Tesla has been completely unable to tap into many of these markets.
Car-sized trucks.
Urban deliver vehicles.
Mini buses.
Full sized buses / coaches.
Light trucks.
Heavy trucks.
Etc...
Where are they? They promised a heavy truck, they made a few dozen, and then... crickets.
While Amazon was running in the red, they could've turned profitable before they did. They could just turn the knob on pricing and become profitable. Bezos is very much a numbers guy.
Elon flies by the seat of his pants. There's no knob at Tesla. There's no robotaxi product or service yet. Viability is way more than a question of pricing.
Fun fact: Novo Nordisk's first success was selling insulin, which was discovered in Canada and licensed by the Scientists who's discovered it for free to the danish company in exchange for a promise to use the revenues "for good purposes".
Hopefully this patent SNAFU makes up for 1% of that monumental screw job.
So you're saying that existing shareholders should vote to shelve the company's plans to expand into autonomous driving, robotics and energy storage and convert the company into a traditional car company?
Yes. Camera only self driving is proving out to be a failure and the two other aspects should have been part of a different company. I don't own stock in Tesla so I don't care if the price drops while the company becomes a car company again.
How do you know they didn't have multiple confirmations from different anonymous sources? Generally this is the case with high quality journalism (souce: dated a journalist).
"Secret Service officials said, speaking on the condition of anonymity"
Their only stated source is "USSS officials" who bafflingly demand "anonymity." I would expect the reporter to tell those /officials/ they need to allow a direct quote or to provide another source; otherwise, their information simply won't be printed.
It's the difference between being a blind mouthpiece and being a reporter.
There could be multiple USSS officials. Also they don't have to tell you if they verified the story through other channels. In fact this is common practice in my experience (source: pillow talk).
They're USSS officials. Officials being the keyword. That a bunch of people who share meetings and prerogative in the organization are saying the same thing is not an indicator of information quality. In fact, I would take it as a negative signal, and would push _much_ harder to get actual detail or corroboration.
I agree. Like I say you have no idea who they talked to or verified the story with. Using the words in a story to justify an opinion, but at the same time saying the story is inaccurate is not logically consistent.
No well trained journalist would ever write a story like this without verifying the information in redundant ways. If they didn't do that then they probably already know it's fake and could literally write anything they wanted to support the narrative.
A) Well trained journalists and editors are not stupid.
B) If they write something false they already know it's false 99% of the time and are doing it for other reasons.
In light of A + B it makes no sense to rely on what is written in the article to support the idea that it is false or undersourced.
How do EU regulations prevent my car from recognizing speed limit signs?
The real problem is that Tesla sold this in Europe, but never put the slightest effort towards actually developing the localized ML model that could make it work. (Of course the hardware sucks too, but they could have done a much better job with it.)
The cars drive themselves incredibly well in US, Canada, China, and Australia
The iEU is blocking deployment to protect their own automotive industry. Reading speed signs is not necessary for the system to function.
In Canada my car drives the speed of traffic just fine. If you live in Europe regulatory sludge has robbed you of FSD not Tesla.
The article pretty much contradicts this - the cars don't drive themselves exceptionally well, unless you're constantly monitoring them. This is expected considering the actual capabilities of Tesla's 'FSD,' but it falls short of the promised performance.
I'm very happy with the performance. I use FSD everyday. It is very very good. I only intervene when I want to. In my last generation hardware, that runs on a 144 TOPS computer, it will very occasionally (twice in the last 4 months) do something like start moving a few seconds before a red light turns green. These critical interventions happened all the time a couple of years ago, but since the introduction of V12 software have been all but eliminated.
FSD is literally improving exponentially, looking at it in Europe (where it is blocked by regulators with economic incentives to delay Tesla's progress) is like looking at a fully fueled rocket ship at ignition and complaining that it is not moving.
Go take a ride in a new Model Y with hardware 4 and V13 software in North America and you'll realize how the EU regulators are screwing over European customers.
Are you referring to autopilot or FSD? Phantom braking is a solved problem since the release of V12 FSD. As soon as a vision based car is safer than a human, it's flaws don't matter because it will save lives.
The revisions and updates are safety tested on roads for months before they are released. Tesla also has models that are too big to run on existing production hardware that perform better than the release versions in test cars.
Updates are not git pulls and no engineer would ever think that they were.
Only skimmed the paper but this looks more like an advance in absorbing solar energy through laser patterning a metal than in more efficient thermoelectric effects.
Laser patterning is a lot easier than improving thermoelectric device efficiency.
Useful for cutting in a line (although fs pulse length is overkill for most applications), not so efficient for surface modifications over large areas.
For enhanced absorbance, alternative technologies that come to mind could be electrodeposition with additives that preferentially bind specific crystal faces, nanoimprint lithography+etching. I was personally involved in developing an electroless deposition technology that created "black gold" that had broadband absorbance across NIR and visible wavelengths. That sucker was black, and easy to make!
