DarkandStormy
Active Member
Reports of Giga Berlin pausing production for five days in June.
You can install our site as a web app on your iOS device by utilizing the Add to Home Screen feature in Safari. Please see this thread for more details on this.
Note: This feature may not be available in some browsers.
... that's the potatoe/potato thing.
How is this not considered theft?
Regarding AI compute, over the past few months, we've been actively working on expanding Tesla's core AI infrastructure. For a while there, we were training constrained in our progress.
We are, at this point, no longer training-constrained, and so we're making rapid progress. We've installed and commissioned, meaning they're actually working 35,000 H100 computers or GPUs. GPU is wrong word. They need a new word.
I always feel like a wince when I say GPU because it's not. GPU stand -- G stands for graphics, and it doesn't do graphics. But you know, roughly 35,000 H100S are active, and we expect that to be probably 85,000 or thereabouts by the end of this year and training, just for training. We are making sure that we're being as efficient as possible in our training.
It's not just about the number of H100s, but how efficiently they're used.
I don't care if it's a called a lie or not since that's the potatoe/potato thing.
I was just sharing with Webeevdrivers that it was not 10 million since he mentioned it, but the articles from 2022 were all about Elon saying they will make (assume sell) 20 million a year. Look at the quote specifically and look at my response (and I why even posted in the first place since I'm trying to post less now to lower the noise/clutter).
Yes, forecasts do change as you say and all the other companies do change projections, but my point again is that we can assume that Tesla is not assuming they will hit 20 million anymore by 2030 (or so).
Again, it's not something I was specifically fixated on, but to share that it wasn't even 10 million by 2030, but 20 million and as they were looking to sell double of Toyota per year.
"I think it's aggressive, but not impossible, that we could do 20 million cars in 10 years. And that would be a good number because there's 2 billion cars and trucks in the world that are in active use, so 20 million would be then 1 percent of the global fleet per year."
Because, unlike you and CNBC, we paid attention during the Q1 2024 Earnings Call, which took place a full 5 months after this memo that CNBC is now reporting on:
This is another reason why the Tesla semi is so exciting. It will be the first Tesla vehicle aimed squarely at corporate customers. Spacex gets pretty much everyone's business, because when rocket A launches at $80million and rocket B launches at $8million, you have to be insane, corrupt, or the worlds biggest virtue signaling organisation to give money to rocket A. With a personal car purchase, people already select a vehicle partly based on styling and superficial differences, so being influenced by an angry FUD stream to choose another EV is way easier.
When the semi hits mass production, there will definitely be a bunch of companies loudly proclaiming their orders and taking pride in them, and then a huge bunch of companies who buy them anyway, but don't want to jump into the culture wars over it.
And obviously a LOT of trucking is done by pretty no-brand haulage companies that nobody has heard of.
By the time the semi factory is complete, I think the first few thousand will already be pre sold to pespi, walmart etc, based on their experience with the hand built ones. The minute a 3rd party haulage firm orders some, it will snowball, because its a cut-throat low-margin business and anybody with ICE trucks will be screwed.
Tesla swapped 12k of Decemberish orders with 12k of January&June orders. These units are going into the not yet completed data center at GigaAustin.
How is this not considered theft?
DarkandStormy - please read and understand articles you post rather than just harvesting negative headlines. Otherwise your contributions here are unwelcome and unhelpful.
Yah maybe. But it’s just how tesla rolls. I don’t have a problem with it. If they think it’s the wrong time to build a semi factory they are probably right. When they have the bugs worked out and they can sell it at a profit they’ll build it.
But with the exception of the model Y they are typically 4 to 6 years slow on product delivery. In the case of FSD it will be more like 10 or 12 years. But that’s another story.
I don’t buy into any of this 10 million cars a year by 2030 nonsense either. They might make 3 million and that’s a stretch. They are not entering any new markets or developing any new models. They can only sell so many of what they produce into existing markets. The supercharger advantage is gone and there are a ton of new models from other manufacturers hitting the market. None that I would want but companies like Hyundai are moving up the ladder fast. Once they are supercharger compatible they will take off like a bullet. And with a lot more varied models than Tesla. As soon as tesla makes them supercharger compatible Tesla will lose sales like crazy.
Jmho.
How is this not considered theft?
The piece is dumb. All that happened is the delivery dates were swapped."But emails written by Nvidia senior staff and widely shared inside the company suggest that Musk presented an exaggerated picture of Tesla’s procurement to shareholders. Correspondence from Nvidia staffers also indicates that Musk diverted a sizable shipment of AI processors that had been reserved for Tesla to his social media company X, formerly known as Twitter."
"X and xAI are tightly intertwined. In a post on X in November, Musk wrote, “X Corp investors will own 25% of xAI.” Additionally, xAI uses some capacity in X data centers to run some of its training and inference for the large language models behind its chatbot, called Grok, CNBC has learned."
No matter the compensation vote, I think anything AI related that's not FSD would probably be developed/migrated to xAi. It just makes too much sense and it is far easier to hire/recruit AI people for xAi rather than Tesla due to upside/mission. I suppose this could help recruit at Twitter/X actually. This said, maybe this isn't a bad piece since the datacenter that Tesla is building isn't ready yet so they don't need the chips.
How is this not considered theft?
That is probably off topic but is also an absurd statement. In many parts fo the world flat refs are the norm. Where concrete construction is the norm flat roofs usually are also the norm.There are two types of flat roofs — ones that leak and ones that are going to leak.
The goal was 20M per year. Always was."But emails written by Nvidia senior staff and widely shared inside the company suggest that Musk presented an exaggerated picture of Tesla’s procurement to shareholders. Correspondence from Nvidia staffers also indicates that Musk diverted a sizable shipment of AI processors that had been reserved for Tesla to his social media company X, formerly known as Twitter."
"X and xAI are tightly intertwined. In a post on X in November, Musk wrote, “X Corp investors will own 25% of xAI.” Additionally, xAI uses some capacity in X data centers to run some of its training and inference for the large language models behind its chatbot, called Grok, CNBC has learned."
No matter the compensation vote, I think anything AI related that's not FSD would probably be developed/migrated to xAi. It just makes too much sense and it is far easier to hire/recruit AI people for xAi rather than Tesla due to upside/mission. I suppose this could help recruit at Twitter/X actually. This said, maybe this isn't a bad piece since the datacenter that Tesla is building isn't ready yet so they don't need the chips.
Some responses on the 20 million vehicles was saying that was in aggregate and they will hit 3 million by 2030. Of course they probably will hit 3 million by 2030 when they hit 1.8 million last year.
But it was always a 20m goal per year:
![]()
Elon Musk says Tesla will probably be producing 20 million cars per year by 2030
That's... optimistic.mashable.com
From Elon himself (referencing 30 million EVs/year and Tesla can do 20...probably...I guess the execution has not been excellent. Maybe RT will get them there now?):
"That’s total market, not all Tesla. We do see Tesla reaching 20M vehicles/year probably before 2030, but that requires consistently excellent execution," he tweeted.
Yup. Dumb. Tesla didn’t have the space for them.The piece is dumb. All that happened is the delivery dates were swapped.
Elon's investment number likely includes vehicle inference (current production and new chip versions), data centers, Dojo, staff, and H100 purchases. Nvidia only has visibility on one of those buckets.