TSMC on Tesla's Next-Gen Dojo / Amazon's Robotaxi / The Negative Free Cash Flow Convo ⚡️
ฝัง
- เผยแพร่เมื่อ 19 พ.ค. 2024
- 🧡 A massive thank you to my Patrons: / electrified
👍🏻 Leaving a like/comment and subscribing are great ways to support the channel for free.
Instagram: dillonloomi...
X (Twitter): / dillonloomis22
💰 Get discounts and/or freebies using these affiliate links:
🔹Athletic Greens (AG1): drinkag1.com/electrified
🔹Surfshark (VPN): surfshark.deals/electrified
🔹Ekster (Smart Wallets & More): shop.ekster.com/electrified
🔹Vessi (100% Waterproof Shoes): vessi.com/electrified
🔹SPAN (Smart Home Panel): www.span.io/?...
🔹 Ground News: ground.news/electrified
🔹EnergySage (free solar quotes, no phone number required) - www.energysage.com/p/electrified
🎥 Executive Producer: Halter Ferguson Financial
Podcast (audio only) version of Electrified:
Spotify: open.spotify.com/show/5yfYEUp...
Apple: podcasts.apple.com/us/podcast...
WARNING: I will never message you in the comment section asking you to message me via WhatsApp - these are scam bots that I'm doing my best to continually block.
===================================
00:00 - Next Gen Dojo
4:40 - Gurley on Tesla
5:10 - Megapacktory Shanghai
5:41 - Tegtmeyer on Austin Event
5:55 - Cybetruck
6:53 - Free Cash Flow Convo
10:41 - Gigacast Retreat
11:50 - Tesla HR Exec
13:08 - Autopark Spring Update
13:31 - Amazon's Robotaxi
14:10 - TSLA Stock Today
===================================
Some links may be affiliate links - you pay the same price, I get some credit. Great way to support creators you value 👊🏻
*Disclaimer: I was a financial advisor with Edward Jones for a number of years. That said, I am not your financial advisor so please always do your own research and do not base your investment decisions on my videos alone.
#TeslaDojo #TeslaNews #TSLAStockNews - ยานยนต์และพาหนะ
Don't. Sleep. On. Dojo. Have a great night!
Analysts have 89% buy rating on invidia but wtf do they know
but... I named my pillow "dojo"
Just that Tesla is competing in this area is fantastic! Ford? Regardless of if Dojo becomes better or slightly worse than the best in class, Tesla will learn enormously for the future, both for coming generations and regarding applications. No other autonomy innovator is close.
I appreciate that you don't do click bait.
actually shocked
i really try not to find a good balance between enticing to click and not clickbait bc as a consumer of yt i also dislike it. tough balance tho tbh lol
Reuters article is hilarious, Musk ALWAYS clearly stated there would be no single gigacasting for a car, and Reuters falsely made up stories that there would be, and now is pretending that their project they pretended Tesla was doing is now cancelled.
i have been in careers where i was media quotable and in 10 years i was never once quoted accurately including distortion by context. i have watched them lie and fabricate news live in conflict zones i was unfortunately in due to career (private sector not govnt job). if you trust ANY news source you are lost
There was talk of a single giga-casting undercarriage for the next generation vehicle, but, I think that was mostly conjecture from the Tesla community, not from Tesla.
@@surferdude4487 Yeah, there was speculation by Tesla fanbois, and baseless articles by Reuters, and Musk and Tesla explicitly said nope.
@@tribalypredisposed Two words: "Straw man"
You do an excellent job on your channel by making all the key points and even showing where you got them
Agree!
This may suprise some people but engineers are usually designing products 2-5 years ahead of time. If you have a 2 year product cycle this means you have both v2 and v3 in development when v1 goes in to production.
In my experience as an electrician, engineers copy and paste at the last minute and don’t like when you point out the flaws in their drawings.
Validating and testing new designs take significant amounts of time and money, as well as tooling (for hardware).
@@lifeistooshort649not quite the same situation...
That speaks to the big advantage of having design work wirh manufacturing. It makes so much sense.@@lifeistooshort649
@@lifeistooshort649🤣 I am an engineer. Accurate. 😂
Wu Wa's video title said Megapack "opening May" but based on the context of the description, I suspect this was a translation error. Probably meant "construction starting May" or "groundbreaking in May," which would be consistent with what we see in his video.
Rueters working overtime crapping on Tesla
Pure envy from a group of losers
I have though for a year now that Dojo is way ahead of where Tesla is telling us. I am waiting for the Leak to come out of its actual capacity.
A single casting isn’t necessarily better or cheaper. It certainly is possible that it is much easier to install other systems ONTO multiple castings, then join the castings, than it is to install those systems INTO a single casting.
The point I’m trying to make here is that it is hard to install stuff into an enclosed space, which would likely be the case with a single casting, rather than install stuff on the surface of multiple casting then joining them together to form the enclosure.
Thanks Dillon! Keeping to the facts and reporting thing based on KNOWLEDGE of how things ACTUALLY WORK sure does come in handy. We have too many people speaking about things they don't understand with very little humility or honesty to first admit they are NOT EXPERTS on the matter.
Dillon is a finance guy and probably never designed a chip. Have you? Do you have any idea how AI calculations are made. Seeing you’re a @dr-k maybe you do.
Get well soon my friend
One piece casting is incompatible with unboxed (Modular) and also structural batteries systems, and the savings from modular, (Optimus???) is higher as you eliminate the so much assembly through doors.
Tesla is the only company I have ever seen that fires people from the top down. Every other company fires from the bottom up. Tesla does things right.
Yes, remember the first Starlink management group and their exit. If something doesn't get done or someone says it can't be done, managers are replaced. On the other hand, worthy employees in that team are retained in the company. I've been there and it's brutal but effective from the company's point of view. TSLA Long.
One tip on the watery eye allergy.
Counterintuitively, steroid-based allergy relief may make things worse, instead, try hydrate it using artificial tear could ease the eye watering very well after a few minutes.
yea i usually dont take anything - thanks for the tip steve!
Allergies are caused by immune reaction or rather over reaction. Low vitamin D causes the immune response to be initially slow then ramp up too hard. At this time of year your Vitamin D will be low. I take 4000 IU with Vit K. Alternatively get blood titre tested and supplement accordingly.
Tesla is the king of NPU. Nvidia is GPU.
Go back to 2019 Autonomy day. It laid the ground for Tesla's expertise. Pete Bannon is the man! He was the first speaker. Begin at the 1:10:00 mark. Chip architecture.
Nvidia has been full on GPU. Jensen and Elon have a great respect for each other. The current Dojo / H100 marriage is Tesla's learning ground and the D2 is probably party of that already.
Like Sandy says, Tesla moves at the speed of thought.
And Tesla isn't looking to kill anybody. Elon wants everybody to partner together to save the freaking world, and American sociopathic 'investment advisers' still don't get it.
The ‘freaking world’ doesn’t need ‘saving’. The brainwashed youth that keep using that phrase need saving perhaps.
"And Tesla isn't looking to kill anybody. Elon wants everybody to partner together to save the freaking world, and American sociopathic 'investment advisers' still don't get it."
The world doesn't need saving.
I did vote all my 2600 shares towards Elon compensation package even if I am still at - 12% on my investment.
Elon has said in a some interview in the past that single cast body is not better than two pieces plus battery. Also this would probably go against unwrapped manufacturing.
Love the new intro background
Real talk
im glad bc it was a pain switching it over given my assets were on my old computer 😅
Another great one, thanks Dillon!!
I will be forever grateful to you, you changed my entire life and I will continue to preach on your behalf for the whole world to hear you saved me from huge financial debt with just a small investment, thank you Veronica Hoy.
I'm surprised that this name is being mentioned here, I stumbled upon one of her clients testimonies on CNBC news last week...
She is my family's personal Broker and also a personal Broker to many families in the United states, she is a licensed broker and a FINRA AGENT in the United States.
You trade with Veronica Hoy too? Wow that woman has been a blessing to me and my family.
I'm new at this, please how can I reach her?
The first time we had tried, we invested $1400 and after a week we received $5,230. That really helped us a lot to pay our bills.
Thanks for ur vids❤
Thanks for todays news. Greetings from Germany 🇩🇪
Amazon is bragging its achievement in Robotaxi! Not that easy!
Great stuff Dylan.
Concerning the eyes and allergies, I’d try Pataday. It’s once daily and over the counter, it used to be prescription only, but now it’s over the counter. My patients swear by it!
i usually try to avoid taking stuff for gut biome purposes but ill check it out thx!
Please reexamine the "orher" in front-rear-other unboxed assembly slide. Those three items are not focusing on the Giga Castings.
I also feel your allergy pain 😂
hang in there bruh lol
Dillon, thank you for your wisdom ;o)
I hope the allergies clear up soon and you feel better.
thank you ken!
Muddy field!!!! I see what you did there! 😏😂
I outdoor train also and I e been on a struggle bus with allergens. 🤧
I'll use the indoor treadmill with filtered air. It's bad enough already.
battery carage is titanium
Is the Dojo chip restricted to being used outside of China, like the H100?
Probably not. Nvidia and Amd seem to be specifically targeted rather then the technology itself. It doesn't ban exports of GPUs it bans exports of Nvidias A100 and H100. The restrictions are also for export (potentially to the Chinese military) and not internal use which Dojo would fall under.
Here is Nvidias SEC filing on the matter: www.sec.gov/ix?doc=/Archives/edgar/data/1045810/000104581022000146/nvda-20220826.htm
Probably
I think DOJO will be way way way better than what Nvidia can offer.
Nvidia had some strong offering a couple of months ago but they are still clinging to legacy/compatibility which tesla just skips over.
I smell Blackrock when I see Reuters articles that trash Tesla.
It's probably just a handful of "journalists" with Elon Derangement Syndrome.
Blackrock, General Motors, Big Oil
It seems likely to me that ant-Tesla rhetoric is orchestrated. Money talks, BS walks.
I’m so confused on what dojo does versus the Nvidia HM 100 chips. We have been hearing 85,100 chips by the end of 2024.
Dojo is Tesla's method of improving Nvidia H100 tech with faster communication between chips. Tesla licenses Nvidia H100 core design from Nvidia to use in their own wafer scale chip design. I think it is foolish to suggest that somehow Dojo competes with Nvidia, as Nvidia makes money via licensing.
Dojo and Nvidias H100 are chips focused on training neural networks and similar parallel workloads. Nvidias advantage is their industry standard programming interface (CUDA) and their disadvantage are their limited supply and high margins which decreases their performance per dollar.
Dojos advantage is a that it is potentially more efficient in terms of performance per dollar and performance per watt for Teslas workloads that they are designed for. Nvidias chips are designed to do anything parallel like weather modelling, physics simulations and neural networks. Think kitchen knife vs potato peeler, where the knife is more general but the peeler might be more efficient if you just want to peel potatoes.
The main disadvantage of Dojo is that you move away from CUDA which means that you have to put in more manhours to make a workload run on these chips. Making your own chips might be sligly better because you can customize it for your specific workload (video and neural networks in the case of Tesla). But it can also be risky and custom hardware needs large orders before they can start to compete with the scale benefits of general hardware.
In realty Dojo have a few advantages:
- Skirt sanctions. H100 can't be used in china and if Tesla is forced to set up a data center there they would have to use Dojo chips.
- Hedge against Nvidia price increases and supply constraints in the future.
- Possible performance benefits in terms of dollars or watts.
So they probably run most of their workloads like training neural nets on H100 and some workloads that haven't changed much over time like video encoding or auto labeling have been ported to run on Dojo. Tesla have ordered 85k H100 plus an unknown number of Dojo chips. They are probably using a mix of them for different workloads.
@@groet508 Thanks for the explanation, but I believe you have one detail wrong.
"Tesla have ordered 85k H100 plus an unknown number of Dojo chips."
What Elon said was "85k H100 equivalent."
Since Dojo licenses Nvidia H100 core tech, this 85k number includes Dojo.
@@royh6526 He said: "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."
The slide showed a graph of ~38k "H100 equivalent" and there was a fairly credible leak that said that they had 7.5k worth of D1 chips installed and the speculation seems to have landed that teslas compute is or at least was about 20% D1.
So it isn't entirely clear and a lot of it is just speculation. The numbers might even be outdated by now.
Luv you
The way Tesla seems to be letting people go. Seems to be a bit harsh. I feel they should be giving more notice to people letting people know that there were going to be cuts. If they keep laying people off the way they're doing it. They're going to encourage everybody to start unionizing. I see that as a problem.
If you tell people in 2 weeks, 1 in 10 or 20 will be gone work comes to a near standstill. Few can do the sort of thinking required when your brain or at least the subconscious is churning on the job issue. People jump ship and nothing works even in program that are not going to be cut. Spend the money on severance and do it like Tesla did. That way they can spend the 'notice' time actually looking for a job. EVen worse there is the risk of sabotage projects and stolen IP.
Termination will be immediate but pay will continue for notice period.
I don't think you have any actual understanding of corporate re-organizing at all.
@@TheEvilmooseofdoom it's possible I don't know a whole lot about how corporate structuring works. Personal feelings more than anything just seems unjust having zero notification of a layoff
Single cast is not possible with unboxed manufacturing
Single casting for a vehicle is not compatible with the unboxed process. Single casting just results in a traditional body with all its assembly problems. Being able to build up assemblies is the fastest, cheapest way to build a vehicle.
How does a single casting reproduce all the problems of traditional parts made from hundreds of smaller pieces?
👍
Elon should know better than use “never”!!!!!
Just reminding everyone. Tesla sells cars to customers, legacy OEM transfer inventory to dealers…
Next gen could be the same design but next gen process.
Sorry to hear about the allergy, but it is allegedly accelerated and being contributed by all the secret ingredients mixed in Athletic Greens, you're using and advertising.
Just a comment about BYD. They now have 99 new outlets in Thailand. You see their cars everywhere here. New BYD pickups are just around the corner. Teslas are not seen here.
BYD's capital budget is 20 billion and it appears they are here to stay. Not a threat to Tesla.
Single cast might not make financial sense. Going from several dozens of pieces to two should result in a huge reduction in manufacturing cost. Going from two to one is very incremental. Such an investment can wait.
I want to apologize for thinking you might have changed the font size. Should have realized TH-cam, for whatever reason is the culprit.
Dojo is interesting but it sounds like nothing has worked yet? they are still buying enormous nvidia capacity.
If they're able to use Dojo for training now why would you say nothing has worked yet?
@@TheEvilmooseofdoom are they? and if they are why are they still buying enormous amounts of nvidia products at great overpricing if their own is better.
Tesla appears to be transitioning from AI optimized supply chains, factories, energy products, transportation robots, and productivity robots to AI designed, built, and operated supply chains, factories, energy products, transportation robots, and productivity robots.
Tesla should go Intel fabs for their next dojo/hw5.
Why?
@@royh6526 monopoly hedging and Intel's A18 with backside power is supposed to be pretty good.
The 'back end' of the chip design process (the process of converting the logical design (written in Verilog) into a physical design that can be manufactured, is reliant on the fab's libraries. TSMC and Intel's libraries are completely different, especially when it comes to the innovative heat management, and serdes/Phy portions of the design. Porting Dojo to an Intel fab would require a massive redo of the back-end.
@@DG-wo8fx Is it harder to port a design to a new node from the same fab or to port one design from one node to a different node at another fab? For example, the same design is supposed to work on TSMC n7 and n6, but some redesign is required to go from n7 to n5. We have seen manufacturers like nvidia change fabs for new generations of chips (ie Samsung for Ampere and then TSMC for Lovelace). Is that just because they have the money/resources to port?
@tripzero0 That is an excellent question. Kudos for your knowledge of Nvidia' chip lineup.
I would say it depends on many factors. It is usually less work to stay with the same foundry as long as the new node supports all of the features required by the new design. But economics, scheduling, or a new feature requirement might make a different foundry more attractive. It could be cost (NRE, lower piece price, etc.), scheduling, or something else. For example, the manufacturing capacity of a given node at a particular fab may be fully booked, sometimes years in advance. Even Nvidia has to complete with other tier 1 customers for time on the machines.
These are just my random thoughts. I retired from the industry 8 years ago.
If Tesla continues to invest, they want to create a computer that will be around 40 times faster (in around 4 years, available later) than a chip from 2024. Edit: In Tesla's time calculation: a product that comes about 6 years later does not work in the chip industry.
An if like ya report U HAVE BEEN getting shout outs on other channel comments section 👍GO AHEAD U ROK!!! don’t change gak!!!
Does anyone else see the contradiction in seeking to manage the key-man risk in Tesla by making the company's board more independent of that Key Man ! Musk's vision and dogged perseverance has created the Tesla we see today so anything that reduces his influence acts entirely contrary to the interests of the company and its stockholders.
😊💯%
I have a limited understanding of all this, but I think curbing some of Musk's behavior could help Tesla. For instance, I think people's opinions of Tesla would be higher if Musk didn't continually promise full self driving "this year" repeatedly. Because Musk has done some stupid things very publicly that have caused people to not like him, some people don't like Tesla as a result. In addition, I think that sometimes Musk forces the company to do things just because he thinks it would be cost-saving in the short term (like not using any systems besides vision for FSD) or cool (installing yokes; removing gear stalks) and not because it's necessarily the best idea for the company.
@renee3461 I disagree with most of what you said, but especially the assertion that Elon is sticking with cameras for short-sighted cost savings reasons. Nothing could be farther from reality. Video trained AI is a generation ahead of the tinker toy Lidar based systems that companies with lesser capabilities are developing. Also, Elon has been putting a full suite inference processor into every car that has been manufactured for several years. (HW3 and HW4) This capability adds several thousand dollars to the cost to manufacture each car. Making cars today that will be able to support FSD in the future is not short-sighted. It is genius.
Hi Dillon Loomis 22 from Trisha 🙂💕
👋👍
I, for one, am NOT impressed with the new Cybertruck OTA update. So what if it won't smash a banana- that's easy! Show me that it won't smash a banana or chop a carrot, but WILL pop the top off a beer bottle 😂
Disclaimer: I do not endorse drinking and driving. Or drinking. Or beer. 😂
Sarcasm ?
not sure why you think you need be on screen just takes up screen space lol
Seems like Tesla has not set aside the 2018 comp package and the fleecing of the company is how they’re scraping it together.
Just watched an autonomous car get stopped by a guy wearing a stop sign t-shirt, you people are delusional.
Yeah, that was old. You are not terribly bright and will always be an easy mark. Nice fail.
😂😂😂
Did you know the computers that run TikToks algorithm are 8x more powerful than Dojo? And Google and Amazon's dwarf that.
There is more than one metrics that measure computer performance. The computer need to be matched to the task. A simple example is a chip optimized for floating point is a bad choice for an application that is primarily integer. The Other computers may be very good at what they do but could suck at training a NN.
You don't know beans about AI if you compare TicToc with real world AI. Real world means that AI will do useful, productive work that was formerly done by humans.
In other words, the TicToc algorithm has no practical use in the real world, unlike DOJO.
@@snookmeister55 I'm just telling you the joke processing power DOJO has compared to other players. Dojo is even using 2 generation old chips. Musk is a liar. His AI is nothing compared to the rest.
@@dynomotivedyno9194 Sure I believe you
first
HR - gone, PR/relations - gone, Battery tech - gone, Investor relations - gone, supercharger department - gone, vehicle programs - gone, AI- gone (a while ago) - who is next?
Probably “reduced” instead of “gone”
Likely “gone from location A” is more accurate
Only 10 or 15% of Tesla owners buy FSD. The next group of potential EV owners care more about a comprehensive EV charging network than technical wizardry! So Elon fires all the people improving the Supercharging network. After convincing Ford, GM and Mercedes to convert, he fires all the people who can support them working the bugs out. After th utility companies start making more energy transformers and wiring being planned,all contacts at Tesla are gone!
Is battery tech gone? I don't think so. Simply having others build 4680s,
@@danharold3087Tesla are still ramping 4680, they are fully supplying Cybertruck ramp. It's not economic to use 4680 s in other vehicles at the moment when the other suppliers are now cheaper than 2 years ago.
EV's are dying on the vine 😮
Nonsense.