Tesla Disbands Dojo Supercomputer Team | Hacker News Reader

Tesla Disbands Dojo Supercomputer Team

bloomberg.com

51 pointsmfiguiere47 comments

> Tesla has faced an exodus of key talent this year as it has grappled with rising competition, falling sales and a consumer backlash to Musk’s political activity. Milan Kovac, the head of engineering for Optimus, and David Lau, vice president of software engineering, departed earlier this year, while Bloomberg reported in June that longtime Musk confidant Omead Afshar had abruptly left.

Wonder how much is belt-tightening from poor financials and how much is talent that was tempted/driven away.

Elon is very much like that boss who starts too many projects, under-funds all of them, and then gets pissy about under-resourced projects failing to produce exceptional results. Just three or four orders of magnitude scaled up from the typical tech industry leader wanna be who does this.

Let's see here. Tesla was the first profitable electric car company. It is still profitable.

Space X is also profitable and growing quickly. Boring company is doing surprisingly well having done the loop in Las Vegas and now doing a big one in Tenn.

Neuralink has 9 working patients at this point who have been helped tremendously.

xAi has the premier model for deep though.

So "failing to produce exceptional results" doesn't seem to add up.

They are smart about shutting down things that won't work. That is smart

> xAi has the premier model for deep though.

When was that consensus reached

On the bot infested subreddit for xAI

I do have confidence that xai will be the premier hype generation AI platform

Boring company's las vegas tunnel is horrible. It's a one-way tunnel that barely provides transit to anybody. [1]

[1] https://www.youtube.com/watch?v=VPjODKUxV5g

Vegas Loop has transported over 2 million riders (barely anybody apparently)

https://thebulkheadseat.com/the-las-vegas-loop-just-surpasse...

thats like one bus route number of passangers.

and most of those were doing the literal loop just for the novelty.

They are doing 30k passengers a day, not just for the novelty. A bus line will do between 5k and 20k passengers a day, and that is with many, expensive busses that have to wait in traffic.

The loop has basically no traffic, and carries a lot of customers and costs a small fraction of what a subway does. A subway mile can cost a billion per mile. This cost 50 million for 1.7 miles.

It's a huge win for people who don't want to walk that distance and for the city streets have which have less traffic.

The math ain't mathing. If they are doing 20 rides per hour per segment, assuming 100% load factor and 24 hour service, which it isn't, that's 480 rides per day, per segment. In reality, probably half that or less.

If there were five separate segments, with the above parameters, that would be 2500 rides per day. Let's further assume two passengers per ride that's 5000 passengers, maxed out all day and all night.

Is there a Vegas betting book on this? What's the over under?

The math isn't matching. That's true. Where are you getting 20 riders and hour? That's ridiculously low

Neuralink's first patient was turned into a kind of Cronenberg-esque horror. xAI claims to have a first-tier AI but has no revenue, not that anyone else has enough revenue to make back the over-investment. National security launches are probably cash flow positive on an individual launch basis. But Starship has been a predictable disaster. It's not plausible that SpaceX as a whole is profitable and evidence for that is in that they require big new investments every year.

Tesla is a particularly tragic story. The vertical integration they had to do because of unavailability of suitable suppliers and car industry corruption and supply chain cartels is an under-appreciated success. Throwing that and all the other goodwill away by going mask off fascist is an amazing act of brand destruction.

The first patient is now using it nearly 10 hours a day and says that Neurolink has changed his life and is now able to do things he could only imagine doing before. https://x.com/ModdedQuad/status/1953274433025786031

You don't know what sort of revenue xAI has because it's not public. Starship is the largest, highest capacity reusable rocket in history, something which no government or company has even attempted. Predictably of course it will take time to get it right. Do you know how many Rockets the US blew up before it got things working?

Most, if not all SpaceX launches are likely profitable, as they are now putting up most of the world's payloads. Starlink is now profitable as well. It's unlikely they are losing money outside of Starship, but that's how companies work, they invest in the future.

Tesla hasn't been hurt nearly as much as you think. They still have the #1 and #2 best selling cars in Euro pe. https://cleantechnica.com/2025/08/02/europe-ev-sales-report-...

Elon motivated people to do things, it was one of steve jobs superpowers.

The naxi salutes had to have killed that. Yeah he'll find zealots and mouthbreathers, but talented ones?

Those companies did it with a passion. Not from the CEO, but the employees.

The packaging they were planning to use was very interesting: https://www.tomshardware.com/tech-industry/tsmc-to-go-3d-wit...

https://archive.today/i1EAP

"Tesla builds Dojo Supercomputer" - stock goes up

"Tesla disbands Dojo Supercomputer" - stock goes up

now that just sums it all up beautifully. over and over again.

I want to write a browser extension that analyzes past HN archives and decorates the usernames of posts and comments with a "Fell For It Again" badge in the case of people who seemed to believe that Dojo existed.

Maybe you should provide a list of things you were wrong about so we can tag you accordingly.

Look, if you're talking about making the HN-sentiment-badge extension general and customizable, I am in.

What a h8er! Next, you're gonna tell me there's no actual Cybercab and those were left over prototypes of a failed attempt at a low cost vehicle, as if this was some kind of pattern of behavior,

Just for fun, here is a great read from a "leading Tesla analyst" 2 years ago "Tesla jumps as analyst predicts $600 billion value boost from Dojo":

https://www.reuters.com/business/autos-transportation/tesla-...

Also worth noting that Elon said 2 weeks (!!!) ago on investors call that they expect to have Dojo operating sometime next year at scale equivalent to 100k h100... How people still don't see that this guy is just lying all the time is beyond me.

The ship is sinking very fast folks.

Given that AI is core to Tesla's robotic and self driving future, it seems like theres a lot more to this story than what has been leaked thus far.

Maybe an intentional troll, given the profile's project.the.snowball moniker. /shrug

There are other hardware vendors.

Yeah, there is more to the story - it's just pure fantasy !

Tesla robotic future ? They are years behind competitors... they are very good at faking autonomous robots though - people are still falling for their Twitter posts, Robotaxi unveil with teleoperated robots and Tesla Diner teleoperated robot serving popcorn (which last something like 2 days... then it was gone already). It's all PR stuff, nothing is really happening behind the scenes.

Well. My Tesla has been driving me around for a couple years now. it's not perfect but it's certainly not a fantasy to eliminate the human driver observation components and just let it rip. I dont know where Tesla stands in terms of progress for robotics, but a major difference maker will be Elon's willingness to risk billions to get there.

Merging together the technological paths of autonomous behavior of a robot and all the little actuators, power supplies and sensors suites does not seem like a big leap either. Challenging but not fantasy.

Part of Elon's process is demonstrating the potential need / business case, while also testing implementation in real world, high stakes environments. A bartender/service roll is a fun place to test since they can deliver a service in a public setting, potentially tethered to the wall.

"but it's certainly not a fantasy to eliminate the human driver observation components and just let it rip"

Yes it is fantasy. This is the same thing like showing a mockup and saying - okay, that's it, now we 'just' need to implement and make it work.

FSD as an ADAS system is fantastic. But it's nowhere near a fully autonomous driving.

Two years ago Elon actually sounded a warning about Dojo. I recall him mentioning this in an earnings call:

“And we’re pursuing the dual path of Nvidia and Dojo. But I would think of Dojo as a long shot. It’s a long shot worth taking because the payoff is potentially very high. But it’s not something that is a high probability. It’s not like a sure thing at all. It’s a high‑risk, high‑payoff program. Dojo is working, and it is doing training jobs, and we are scaling it up, and we have plans for Dojo 1.5, Dojo 2, Dojo 3, and whatnot. So I think it’s got potential, but I can’t emphasize enough: high risk, high payoff.”

https://thetranscript.net/transcript/5550/tesla-q4-2023-earn...

In other words, he admitted it’s a long shot to replace Nvidia with their own chip awhile ago. It’s interesting because Google hasn’t given up on their TPUs, but Elon being a hardware guy mostly, has.

I doubt Elon had the time or the inclination to build yet another hardware stack, and he probably saw something that showed him he couldn’t quite catch up to Nvidia?

The scale of Nvidia manufacturing and RnD should be staggering atm. Google built the entire software and hardware stack for TPUs before the AI boom. I can’t even imagine how hard it would be to hire GPU hardware talent right now.

Really hard to do training hardware as the software stack is quite complex and needs to be general and flexible. Inference on the other hand is easier, which is what a lot of new hardware will be. I doubt if it makes sense for Tesla to build hardware for datacenter inference. Xai on the other hand could benefit from it.

> it’s a long shot to replace Nvidia with their own chip awhile ago

It's also worth mentioning that having their own ambitious effort in secret development was a valuable BATNA (negotiating tactic) for long-term contracts with outside suppliers. The value of that recent >$16B deal with Samsung for AI chips thru 2033 may have benefited from the prospect of Dojo as a potential in-house alternative.

It’s amazing how people are operating with such immunity. If I lied to my boss like that I suppose I’d just be fired and let go without pay…or would I get promoted ?

I’m not into fucking society over like these sociopaths, I guess that’s “my problem”.

This is what people said in 2018 when the stock was 16$. In case you aren't keeping up Tesla signed a $17billion agreement with Samsung to produce their AI6 chip family which is designed to replace the dojo architecture. It is a harmonized design that can be scaled to training clusters or deployed in edge inference.

Another Karpathy initiative that turned out to be utter trash ...

https://www.hpcwire.com/2021/06/22/ahead-of-dojo-tesla-revea...

I don't understand co.ments like this. Clearly they realized that it wasn't worth doing and cut costs accordingly. This seems smart. Companies constantly place bets and adapt when they aren't working out.

As far as the ship sinking that's pretty silly. They continue to be profitable with some sectors like energy growing very quickly.

Virtually all EV makers have faced a couple of tough years. Many like Rivian and lucid are getting clobbered.

Tesla has a 37 billion dollar bank account and low debt. There is no scenario that the ship is sinking

What I'm pointing to is the absurdity of the situation. It was supposed to be a huge project, valued by some analysts at $600 billions, and years in the making. The CEO himself self recently they are on track to bring it online next year at a huge scale. 2 weeks ago he was still saying it, now it's axed.

It was clearly nowhere near where they said it was, or they would have some significant sunk costs and write downs. None of which exist in their financial statements for Dojo.

As for the ship sinking - I know it seems like a huge exaggeration now. But it's a car maker, with a collapsing sales and which made virtually all the profits from direct and indirect subsides - which just got axed too. That alone would be very worrying. But it's also a company which sold billions of $ of software that is still not working as advertised and lawsuit are only starting to come in. By their own admission the software won't work on >50% of cars they already sold and they would need to upgrade them 'sometime in the future'. Oh and the CEO is the most hated man in the US currently.

In 2023 their profit was 15 billion dollars. Of that they got 1.79 billion in regulatory credits. So about 12%. They also made over 15 billion in 2024 of which regulatory credits were 2.7 billion, about %18%. So your claim seems false.

https://ycharts.com/indicators/tesla_inc_tsla_automotive_reg... https://finance.yahoo.com/quote/TSLA/financials/

Here is an overview of the impact of direct regulatory credits:

https://edition.cnn.com/2025/07/22/business/tesla-regulatory...

Overall, Tesla got >$10B of free profit from direct regulatory credit sales over its lifetime.

Indirect ones (7500$ subsides to car buyers) are going away end of September. While it's not a direct benefit to Tesla, it was a huge boost to sales and profits.

Tons of industries and companies get huge boosts from the government, and that's true for all the other EV manufacturers as well.

You said Tesla gets "Virtually all their profits" from regulatory credits. Tesla has made 45 billion in profit since 2021. Even if they took in 10 billion in regulatory credits that is less than 1/4th of their profit just since 2021. So not exactly a truthful statement when you say "Virtually all their profits"

That's why I said 'direct and indirect'. Tesla benefited hugely from subsides for buying electric cars (I'm not disputing whether it's just or not - they were the only real ev maker in the US until recently) and those are getting cut off very soon. Let's see how profitable their car sales are in Q4.

No question they benefited from it, they wouldn't have grown nearly as quickly. Regardless though your statement that virtually all their profit being from subsidies is just factually incorrect.

I definitely think it will hurt US sales somewhat, however it's going to drive the other EV makers in the US into the tank. Most are still hemorrhaging money on their EV's. Tesla can just drop their production amount if sales fall off.

The US market is less than half their sales. We shall see.

"they continue to be profitable"

The credits are drying up, the sales crashed, the brand is dead, and there's only so many mark-to-market accounting you can do on cryptocurrency.

If the stock wasn't buoyed by some bizarre modern financial techno cult, I'd short it

Sales are down modestly compared to Rivian and Ford which are getting crushed YoY. Brand is fine, in June they were the best selling EV in Europe.

Energy continues to grow fast and there are a lot of exciting products in the pipeline

I wonder if this is to shift the gpus to X.AI?

I was thinking maybe this, but more likely X.AI would develop the models for Tesla, such as for FSD and Optimus.

Honestly, it might not be a bad move. X.AI has done some great stuff, Grok is rough around the edges but pretty much at the bleeding edge and it doesn't make a lot of sense for musk to be running two AI camps (Yes I know they are theoretically separate companies...(

Reply on news.ycombinator.com