Tesla stock falls 3% after Nvidia unveils open-source AI for autonomous vehicles

Can't wait to drain the battery letting AI decide how to make a right turn

But ... Nvidia requires radar and lidar..

Oh look, there couldn't of been insiders for this bad news so the stonk actually fell. Neat.

Good catch! You're right, I should've cut across the entire road from the left turning lane instead of trying to make a left turn. I'll make sure I make better decisions in the future.

The stock fell 4% ! The scrambling efforts to downplay NVidia’s announcement by Elon fanbois is hilarious to watch!! 😂🤣🤷‍♂️

Odd, why would a company that doesn’t even have a self-driving car be impacted by a company releasing an actual functioning product? Is Nvidia even putting a man on Mars by 2025? /sssssss

Unreal this is only a 3% drop. Should be exposing this company for the smoke and mirrors it really is.

Cars = lost the lead now Taxi = not working yet and competitor has been doing it for years now. Now nvidia is coming for them Robots = so many others are already about to release finished products like Boston Dynamics The walls are closing in

“If you don’t believe Tesla will solve autonomy, you should sell your stock”

Call me when it's 50%

Nah, they are all buying the dip before it goes to 500+ …

It might be (also) because of their cars are’nt selling

I hope the beg of the end

What, you mean sales down 30% in Europe might be a bad thing?

Oof

4% is nothing compared to the usual volatility of Tesla.

NOTE - This has been on Nvidia's website for over a year........

Nope, he'll just "pivot" to the next shiny object.......

okay, welll uhhhh... ROBOTS GUYS!!! ROBOTS!

Well Elon will have the AI powered child-like sex doll market all to himself. Checkmate.

I expect talk of cybernetics with brain interfaces. Maybe cryogenic live forever type products also.

How many of you are actually shorting tsla?

They also lose market share in energy but that side of the business is still growing

Think bigger. Desktop nuclear fusion.

Apparently Jensen and Elon are not so good pals then

Mars, and... Andromeda!!

I used to think the same as you when I had a 2017 tesla. I thought it’ll never reach a good level of self driving. I recently rented a juniper model y with fsd and I’m genuinely impressed, fsd blew me away how well it worked with HW4. Have you tried one of the newer ones?

Last month in Australia the BYD Sealion outsold the Model 3 and Y combined. Tesla down year on year on year and lost their crown as the best selling EV brand to a manufacturer that only entered the Australian market in 2021. I've probably just made Tesla shares go up 30% off this post.

*The walls are closing in* Not until the AI bubble finally pops. It's not until then, when Musk's rebranding effort was for the toilet.

"FSD is the difference between Tesla being worth a lot... and zero" \- SnakeoilSalesman

Somebody didn't watch the Boston Dynamics Demo? :D

That's what I was getting at lmao

Dark matter harnessing quantum tokamaks, on Mars. With personal xAI cybernetic implants. #420

Yeah, it's got a beta of almost 2.

We can only hope this is where Elon’s shell game falls apart for real.

It's because NVIDIA makes things the general public actually want

I have a HW4 Model Y with the latest FSD beta (expiring Friday). It is indeed pretty good, I enjoy enabling it even on local trips. Last night it spotted some rando wearing all black wandering around in the road . . . I still have to take over here and there for lane changes but it's a solid product for what it is. When Elon puts his money where his mouth is, though, is when it's bundled free with Tesla insurance and Tesla Inc. takes financial responsibility for its operation. If it's so great Tesla cars should be much much cheaper to insure.

Fun fact: FSD will require radar and lidar too!

yeah, the actual announcement was something of a nothingburger since it's middleware for legacy makers who can't do FSD in-house like Tesla. but it is a long-term dent in the FSD story . . . "where's the moat??"

I have a few puts expiring March -> May that I picked up July -> Dec (bought with play money for the LULs mostly) After seeing the joke of the October 2024 preview, and the minimum-effort Austin deployment, I took the under on all of Elon's promises last year and was proved right, but had my head handed to me anyway. Still hopeful though that the May puts save the position, and even the March expiries are still active given how fast the market moves in & out of this stock.

The stock was already selling off, and would have dropped with or without this news. \_\_\_\_\_\_\_\_\_\_\_\_\_\_ I did find Musk's desperate need to respond to and downplay the Alpamayo announcement interesting through... suggesting that: **"...but the auto industry is doing little on their own"** ... because they were likely waiting for an open source solution that would enable all of the car companies to share/crowd source training data ... aka Nvidia's Alpamayo... \_\_ "**Tesla will have spent \~$10B cumulatively just on Nvidia hardware for training."** ... Now all OEMs can cumulatively spend money to use Nvidia's servers to do the training, and I imagine that Nvidia already has a robust core set of training data. \_\_ **"We combine that with our own AI4 chips for processing vast amounts of video or we would probably need to spend double that amount"** .... so no word on how much he spent on developing and buying those AI4 chips... Further, Nvidia's system has the mechanism to not only process video data, but the data from a suite of other sensors. Given that Tesla's valuation is obscenely overvalued based on their vaporware, other companies (including Nvidia) cumulatively spending $10+ billion shouldn't be an issue. Further, we're living in a cloud server environment. Tesla built their own servers... cool. Every other OEM can just lease servers to do the brunt of their training, and when the training's done, those servers aren't their problem anymore. \_\_ **"Then we are making \~2M cars/year and rising, all with our dual SoC Ai4, 8 cameras, redundancy in steering actuation and other systems, high bandwidth communication, etc."** ... first off, Tesla is only producing \~1.655 million vehicles per year. The rest of the industry is producing the other \~80 million or so per year. Sure, the OEMs don't yet have cars being built with full Alpamayo sensor suites, but that's something that can easily be solved. I'm more interested that Musk didn't mention the number of cars they've already produced with their tech, given that they haven't actually succeeded in achieving full autonomy and training the system is the hard part... not building new cars with the suite of sensors. ... second, IMO, autonomous driving's main value for personal vehicle owners is on highways, of which I think many OEMs already have solid highway ADAS systems and Alpamayo may only improve on that. When it comes to autonomous city driving, the lion's share of value is in the ability for the vehicle to act as a taxis. Which is ironic because Teslas give folks autopilot for free, then charges them $8k for FSD for the less valuable use case. I think it should be clear by now that Tesla/Musk have no intention of ever allowing their vehicle owners access to the robotaxi software. Why would they share the taxi profits with their customers when they are "making \~2M cars/year and rising" and can quickly push out their own fleet that'll generate 2x as much revenue and profits for the company versus using customer cars as taxis. Remember folks, Musk claimed (lied about) each of their cars would generate $600,000 in profit over a 10 year lifespan. Why would they share $300,000 of that profit with their customers when they don't have to? There's nothing in the FSD contract that suggests customers are guaranteed to get access to robotaxi software and capabilities. **\_\_\_** **"What they will find is that it's easy to get to 99% and then super hard to solve the long tail of the distribution"** ... Can Musk please stop suggesting they're at 99%. He first claimed they were at 99% in January 2019. Think about that... In 2019, they were still using radar and did for years before they decided to remove them from their cars on account of the radar parts shortage after the pandemic. If the last 1% takes as long to solve as the 99% they're claiming to have already solved, then they were never at 99%. They were at 50%.... This is the type of idiot catcher phrase that Musk is known for. His suggestion that they're at 99% never made any logical sense. It sounds cool... but it's actually wrong and dumb, and anyone who believed it was also wrong and dumb... and more than anything... gullible. I mean, people... the man's been claiming FSD would be working "in one year" every year for the past ten years, and that a million robotaxis would be on the road "in one year" every year for the past 7 years.

Not if agent K gets his way

if I were working on FSD the first thing I'd tackle, after stopping at stop signs, is not running over potholes and shit in the road. they saved that for last for some reason.

Yawn. You're still talking about cars? Tesla's an autobot company now! Few understand.

Not me. I just have a morbid fascination in extraordinary popular delusions and the madness of crowds.

Not sure why anyone would feel compelled to use nvidia's alpamayo over google's waymo, but I guess it will be interesting nonetheless to see how it plays out

MMs ain't gonna let retail traders make any money from puts, come on now. They're going to slow bleed this bitch dry and force it flat (relative to options IV anyway) for years to milk as much premium as they can before letting it finally die

he already did lately...data centers...in SPACE™!!!!!!

I have been reading almost identical comments since the Model 3 came out and I used to think you were just bots, but I had a conversation with a guy who swore it was really working 100% of the time and that just doesn't mesh with my experience, or the experience of anyone I've personally met who's tried FSD, pre or post "supervised". I even ask people if they live near one of the factories or one of the offices where Musk occasionally shows up when he's not playing video games (s/) because Tesla employees have leaked that they "Waymo'd" all the areas Musk is likely to drive himself so he'll leave them alone. The guy I talked to was nowhere near any Musk owned facility. I just don't get how peoples experiences can be this different.

Granted last time I was in a Tesla with FSD was 3 years ago, but it was still freaking out about fresh asphalt patches in pavement. Real potholes, not so much.

latest FSD actually dodged a small darker asphalt section of the lane this weekend when I was driving (I have a HW4 Model Y). It was fine to run over LOL. If cameras can't figure it out they need some sort of specialized scanner to read the upcoming road surface right; or just pay $2B to fix every pothole in the US, that would be cheaper actually.

Old fsd was one of Elon’s scams that kept saying “coming this year”. I thought he was a con and hated tesla. I’m genuinely considering getting one now. It’s that good. I don’t even pay attention to the gps, speed limits or anything. I just look forward and next thing I know, I’m there. I suggest you turo one for a day with fsd. That’s what I did. I paid about 250 for 4 days and drove to Vegas.

"... drove to Vegas." Assuming you were coming from CA, you were in a corridor that the Tesla programmers have "Waymo'd". Your 2017 Tesla would probably do just about as well as the new one you rented. You might consider taking a trip somewhere Musk is unlikely to go and see if it works as well on routes that haven't been pre-mapped before buying. I'm not saying I'm sure it isn't better, I haven't driven it, but I know to be suspicious about FSD performance anywhere Musk is likely to be.

>. If cameras can't figure it out they need some sort of specialized scanner to read the upcoming road surface right This has to be a joke.. right? Becuase that is the entire point of Agent K's (K stands for ketamine) whole "a camera only solution is far superior than a camera plus other stuff like LIDAR"

pretty sure that is the joke.

Nah, I'm not rich enough to bet against institutional investors.

How about just the fact that they make things that actually work

Really? Lolz

Embarrassing that you oversell your impact so much. 20% at best..

You shouldn't be downvoted and yes FSD is impressive. The question is whether it will ever be *unsupervised* level of impressive.

Came from Phoenix area. Maybe that’s true, I’m not sure. I’m still an Elon skeptic but I was definitely impressed to the point where it’s in my top 3 choices of next car purchase happening in 2027. Other choices are BMW with nueu klasse and Hyundai with pleos. I want to wait and see how those systems are. I prefer ICE over EV.

I'm surprised he hasn't tried to buy Boston Dynamics (or maybe he tried and failed)

Tesla 2025 sale dropped 9%.