원문 보기
but you know Carl my problem is is that when you do it when you put a 500 billion dollar business uh when you say it's going to be that we're not talking about 2023 24 25 and the market works and discounts out about a year but Jonas does it out about three years he's out of sync with both the economy and the stock market so I didn't like this call at all I like the hostess twink you call more frankly I I hear you and listen it's a wow Jim Cramer bags on Tesla suggesting the market only prices out a year and that this guy's going three years out Jimbo is not realizing the massive lack of institutional allocation to Tesla and that you need to start allocating now buddy to be part of oh my gosh what what kind of poop is this Jim I tweeted Jim Cramer says he does not like slash agree with the Morgan Stanley price Target hike so I wrote thank you Jim Cramer this is extremely Tesla because holy smokes Morgan Stanley didn't they finally released a piece on Tesla that's something that we've actually been pointing out as many have in the Tesla community that institutions are just totally blind to and let's be clear my opinion is that retail probably already has a really good allocation to Tesla with the exception of some of the retail bearers who are keeping some cash on the sidelines which is fine but I think it's institutions who are substantially underweight Tesla and they move from institutions to properly waiting Tesla in their portfolio would lead to a violent upside in Tesla share price now we're going to go through what Morgan Stanley just said and uh we'll briefly touch on the United Auto Workers strike basically uh if it goes through this week it's probably going to be good for Tesla we can talk more about that really later but take a look at what I tweeted uh just last week so uh last week I tweeted the following uh on the the eighth so about three days ago this was a Friday and I tweeted the following there's a reason 25 of my ETF which is quite literally the legal maximum for my type of fund is in Tesla I strongly believe that Elon Musk is great not perfect but means great now the Cyber truck will dominate and become the most popular truck globally and ever and that Tesla and pool employees are smart cutting the X and S prices was brilliant they listened to our feedback it's something a lot of us have been complaining about that the S and X's are just quite frankly too high the value proposition difference isn't isn't that high uh and uh that Tesla this right here is a big one Tesla is the the only highly profitable vision-based AI solution to exist I made the argument that I believe 99 of artificial intelligence companies sorry 90 of artificial intelligence companies will be gone in two years and 99 will be gone in 10 years and that Tesla will be a Survivor here's that argument that I made about institutional allocation to Tesla being the lowest of all Mega caps maybe that's because Elon is a bit nutty but it's kind of my kind of nutty is what I wrote uh and of course do I agree with everything no but I think they're going to do the best they can solar and batteries are about to explode under the inflation reduction uh credits and all the bearish narratives fail to worry me uh and somebody replied and said any negatives I said rates beside that none worthy I'll still believe this is true well now on September 10th so uh yesterday Morgan Stanley released this massive piece uh and the the main writer of this is someone who's written about Tesla at length before uh they are actually pretty good researchers like this this is this is not just like uh you know licking the finger and kind of seeing which way the wind blows there's actually some real logic in this uh and what I really like is they go deeper into one of the core ideas that I was talking about in my tweet which was this idea that a Tesla is the only highly profitable vision-based AI solution to exist and this is really profound how they break this down so uh on a high level point of view uh they simply argue hey look we think we're going to move Tesla from our previous 250 dollar price Target to 400 dollars uh now this is just an Institutional price Target but it's worth noting that's about a 60 increase in Tesla share price Tesla's stock just because this analyst is also respected at least in pre-market is moving up six percent how much of this will survive into the coming weeks and otherwise will really depend on how much institutional allocation changes because of this of course there'll be some algorithmic trading that goes into this some retail trading that goes into this in terms of a pre-market number but what you're really waiting for is this this explosion in Tesla shares that could come from institutions finally waking up and saying okay why are we underway Tesla this is much more than a car company and that's what they start off with they talk about look you you know how Amazon web services ended up driving 70 of the revenue at Amazon we think something like that could happen at Tesla as well and what's worth noting about that is Tesla's revenues from AI are really low right now probably you know somewhere around less than uh 20 certainly under 20 of Tesla's total revenues coming from actual artificial intelligence and recognizing uh uh you know FSD revenues and it also of course depends on which quarter you're kind of looking at because sometimes they recognize a little more like I think q1 of this year they were recognizing some more uh which which maybe represented a larger percentage but on an annual basis they're still recognizing a very low percentage of their revenues coming from FSD and we know that FSC and autonomy are extremely high margin and this really creates what Morgan Stanley starts talking about this this flywheel of profit basically for Tesla and so we'll look at that in just a moment but take a look at this Tesla's management has talked about how they need as much compute power as they can't currently get their Hands-On because they can't physically secure enough chips to train the artificial intelligence that they need and this isn't to say that when when they go out and buy as many Nvidia gpus as they can it's not to say that their own in-house designed application specific integrated circuit Asic AI Asic is not good or better than Nvidia it's just Tesla has so much demand for computing uh Vision based artificial intelligence data if they need everything they can get their hands on now to really get the release of fsc12 out fsd12 by the way worth remembering is really special because fsd12 is really getting rid of the old way of doing artificial intelligence which was uh let's let's start with uh you know Vision based systems that are labeled by humans and then let's try to get the computer to take the efforts that humans have done saying this is a stop sign this is is a red light this is what this sign means you know this is the stop line on the road whatever instead of going you know instead of taking this sort of a auto labeling approach they call it again humans sort of gave a little heads up and then the car is like I think this is a stop sign right and then we can sort of confirm it uh going into this basically retrained version which we've talked about previously as well it's completely retrained bottom-up trained version of artificial intelligence for FSD using instead of this Auto labeling technique a technique that says just go out there and learn how to drive which initially sounds really scary it's like wait a minute that sounds terrible that sounds like it's going to take a long time but Morgan Stanley actually makes the argument that it's going to happen surprisingly quickly because of not only a the number of vehicles on the road that Tesla can learn from whether or not they're using FSC Tesla can learn from all this what Morgan Stanley calls light years of data they are collecting Tesla can collect all of this and then learn from that and teach fsd12 so in other words the drivers themselves are not Auto labeling they are robots literally teaching the robot in other words every Tesla driver is the robot doing the teaching work and a lot of people were freaking out when it was either earlier this year or last year Tesla was actually laying off people in their labeling Division and now people are starting to go oh my gosh artificial intelligence has replaced artificial intelligence labeling jobs yeah that's how far along Tesla's artificial intelligence is and this this whole FSD thing uh is is so undervalued uh by the market it's really incredible uh but that's probably also in part because a lot of institutional investors and fund managers I hate to say it but there is an age element here okay you get older people we're like I ain't trust in the damn car I've been driving I've been driving 50 game years usually I'm gonna trust the computer to do it hell no I can't even get a computer to turn on without having to restart it you know like like you're gonna have an some element of that right uh and so there's gonna be a come to Jesus moment here at some point uh but anyway let's keep going through uh through this particular piece uh since uh since we're gonna be going through some numbers here I do want to just bring your attention to uh this date right here uh today's obviously 9 11.
uh it's a day that'll really live in my memory forever I remember being in a child and getting ripped out of school because of that uh uh that day I wasn't fourth grade great at that time but one thing that I want to remind you of is that on 9 15 we're having one of the biggest expirations ever of a coupon code for the programs I'm building your wealth if you like my perspective my research that I do on fundamental analysis earnings calls real estate analysis you want to do q a with me join me in those course member live streams learn from me learn my perspective in those courses on building your wealth most popular still right now is the uh stocks and Psych followed by uh surprisingly getting more popular here uh maybe unsurprisingly the zero to Mill real estate investing course followed of course by the wealth course and a lot of people are bundling either these two or these three we have some other Niche products as well like the do-it-yourself Property Management uh which we finally got our Builder relationship back which is really cool gives massive discounts for using a big box store that I will not mention here but it's kind of like Home Depot we get some massive discounts there for all course members we're releasing that this week as well after that expiration so anyway take a look at that linked down below okay so what do we have over here so the forces that have driven uh 70 of Amazon's ebit can work to Tesla as well okay great what's so powerful well really it's the D1 chip in the dojo ecosystem that'll actually potentially be six times as powerful as the Nvidia gpus now Morgan Stanley makes it clear here that they're more evaluating Tesla 6X efficiency claim based on on what Tesla's saying that Morgan Stanley did not independently verify that 6X claim so they're basically saying hey look if it is true that Dojo is six times as good quite frankly even if it's just like three times as good Tesla's ability to highly vertically integrate here by capturing the data with the robots basically driving the cars us humans being those robots combining the chip Talent with vertical integration the software that they have unlimited access to Capital and how this is all vertically integrated in one company is absolutely an obscene opportunity for Tesla to generate massive amounts of Revenue and this is what they call the flywheel effect where on one side the lower they can get the cost for manufacturing cars the more they can get cars on the road which creates not only more cameras that are picking up data but remember you got to remember the most important part Tesla should be paying its drivers basically because of the value they're giving Tesla or just by Tesla stock but anyway the drivers the more people the more this is why elon's like we could sell the cars for no margin he needs butts and seats training the damn car even when you have FSD on yeah Tesla drivers will know this even when you have full self-driving on and you're driving you know how sometimes Tesla is kind of like slow off the start of the stop sign off the light or it's like all right come on turn right already like you stopped turn right let's go what you can do is while it's on full shelf driving or or not and it's going to learn both ways you could tap the gas a little bit and kind of like okay go so you can accelerate without disengaging the system right and what's remarkable about this is all of this is just absolutely massive training so again we are the robots training the robot it's really bizarre it's kind of like we're training our for our own demise uh anyway oh that I don't know maybe I don't know why I'm laughing at that but anyway uh they argue here that with Dojo that went online this summer June July that uh Dojo all of a sudden makes Tesla a multi-industry potential AI player and that they can get AI training down from months to weeks it's efficient it can generate Revenue now uh you know what I think is absolutely fascinating as well is their talk about right here where which Industries they could actually be useful in that it's not just going to be cars but that Tesla is AI their Vision Ai and this is again going back to my tweet what why I said there's a reason uh you know my you know we're like 25 plus deep in Tesla there was a reason for that it is the best artificial intelligence play that exists there is no better AI play this Vision AI is a game changer uh and it's not just a game changer for cars but it's a game changer for manufacturing for robotics for Aviation for health care trains utilities security cameras security facial recognition you name it this is absolutely some incredible training also I mean think about facial recognition for a moment the training that's being done inside of the vehicle to understand what you're doing just by looking at your face as a way of determining are you on your phone are you paying attention to the road are you paying attention to the radio for too long are you dozing off these are some incredible tools that could end up expanding into much more or greater opportunities especially in the realm of security it's also kind of weird from like a whole privacy POV but uh we'll save that discussion for a different day so uh then uh what I think is so neat here is that the the problem for artificial intelligence really becoming like you know driving full autonomy are the edge case scenarios you know this is where again you know the problem that I regularly have with my FSD and I'm gonna draw it out here because some people just don't understand it uh you know the problem that I have with my Edge case scenario is I've got you know a road that kind of looks like this and then here we're gonna have a yield style of a road this is going to be our yield over here and uh then of course you could go straight over here this way okay great the problem is there's actually a a neighborhood entrance uh right here and then what you end up getting is a right turn lane here that ends like this then you have a bike lane that's sort of right here and then you have you know your your actual right turn over here and so generally what you do imagine this bike lane is sort of like a dotted line here as it's supposed to be so you can get over uh what what Tesla loves to do is it loves to go from the road over here and then go into this Lane and basically break this line over here which is what you're not supposed to do this this line I shouldn't even draw it with dots it's supposed to be solid so so the theory is that when you want to turn right into this neighborhood you turn your car right into the neighborhood first uh and uh and then you're gone you're out of the flow of traffic the reason you're out of the flow of traffic is so that the vehicles that are turning from the main road here are going this way can actually go in if you want to look this intersection up exactly it's um Johnson and North Bank Drive in Ventura but the point is uh that the car does this and then it's like oh no I shouldn't be here and then it kind of tries to Zig out and it's just a disaster every time so I pretty much always disengage over here but even me disengaging and then driving it the way it should be driven uh is is a way of training it bu called an edge case scenario and Morgan Stanley argues that Dojo is really going to be able to destroy these Edge case scenarios and finally iron these suckers out and everybody who drives FSD around their Town knows there's some areas where there are Edge casing areas they just think that fsd12 is going to be able to kill these Edge case scenarios quickly thanks to the uh the dojo exopods which big deal for them right here uh they they make some comparisons between Nvidia and how Nvidia server Stacks compare to Dojo how important they are today but really how it's going to cost you about six times the amount of money probably to set up these these uh these uh Nvidia server racks rather than just using Dojo because Dojo is the application specific chip designed not for broad-based vision-based AI but for what Tesla needs and what's really incredible here too is listen to this line right here for comparison if Nvidia is expecting to ship 200 000 to 250 000 h100s Tesla is expecting to ship 40 to 50 000 Dojo chips now it takes more than one Dojo chip to end up being the equivalent of an eight h100 anyway these h100s are pretty beastly uh but still what's remarkable is when you line it up like this you're like oh my gosh wow Tesla is actually manufacturing chips at this level that to some extent you know with respect to your uh what a quarter uh it's almost matching a quarter of nvidia's h100 production uh and it just it's designed not to show you that the revenue is somewhat 25 equivalent but that Tesla's really producing a massive amount of these chips and it's not just making the chips it's using the chips in artificial intelligence I can apply uh to various different Industries around the world that is expected to to dominate uh artificial intelligence uh and vision Based training which is very exciting but uh anyway so uh yeah you can see here one uh 25 D ones is expected to be 30 times faster than 24 gpus not directly comparing to the h100s but making some I think a little bit weaker comparisons over here just because we're not talking specifically about which chip but anyway Morgan Stanley is very bullish in addition to this sort of one of the big last points they make is they actually this is something we've been talking about for years by the way well I mean ever since Elon started talking about acquiring Twitter so I guess that's been about a year and a half Morgan Stanley is finally saying it oh my gosh the vast amount of data that's starlink and Twitter have access to in terms of language learning in terms of uh you know basically I mean everybody's bandwidth going through Starling uh learning about humans all of this data is so useful and eventually likely to be very profitable for Tesla that we haven't even seen the beginning yet of Tesla autonomy or Tesla uh artificial intelligence revenues yeah and again this is why it goes back to my belief that this is the best and likely most profitable AI company to exist with the best likelihood of surviving uh again not just because of the vertical integration but that flywheel effect which takes us from again the more robots and seats driving these cars teaching the robots that's us the more data we get the better we can make the platform the more users want to use it the better the product is the more uh different uh vertical Stacks you can get into and then all of a sudden Tesla actually does become this software company with sas-like margins really really incredible piece and again it all goes back to this idea right here that I believe that Tesla is the only highly profitable vision-based solution to exist and there you go Morgan Stanley has memorialized it two days after my tweet I think they did a really good job and I I agree with him I actually think honestly uh their price Target would have probably been higher uh if it weren't for the potential and I know this sounds crazy but if it weren't for the potential institutional backlash uh at Morgan Stanley like for example if Morgan Stanley came out and said we have all you know four thousand dollar uh uh price target for uh oh yeah my mic isn't playing this sorry there we go thank thank you for catching that with the microphone uh hopefully that wasn't too bad anyway uh you know if Morgan Stanley came out and they're like oh we have a you know four thousand dollar price target for Tesla people would just laugh this off it actually wouldn't have the real effect that Morgan Stanley's piece is having now Morgan Stanley suggesting oh 60 upside in Tesla share price this is just the beginning once these things actually start happening at Tesla the price Target will actually end up getting revised even more and more and more again think about it and and you know I don't want to use Kathy in in a negative light here because you know I like and respect Kathy but if if they pulled a Kathy and said it's going to 1500 and it's going to five thousand it's going to ten thousand even though they're probably right it's gonna sound like a YOLO so Morgan Stanley's kind of cooking the the lobster or the frog in the pot slowly and Morgan Stanley just went like oh I see where we're going with this let's turn up the heat a little bit now and then later we'll turn it up a little more you know we'll go from 400 to 600.
to 690 to 800 to a thousand right and slowly turn it up they'll actually have more respect than the we're going to the 1000. again even though that's probably just being more transparent and upfront Morgan Stanley is setting it up to actually be palpable by the well I trust the dumb computer laughs so anyway uh that's my take on uh the wildness of this Morgan Stanley piece and Tesla uh so thanks for watching that tank on Tesla make sure to remember mark your calendar for September 15th if you want to get into house hack we are starting house hack exclusively for course members in this fundraise Round And if we max out because there's an SEC limit in terms of how much we can raise that's it that means only course members will end up getting the one-to-one valuation of the Vanguard of real estate uh so it's another yet another reward for course members we also have course members who are able to Shadow me for a substantially uh discounted price which is really exciting pretty much every course member who comes Shadow me and is like oh my gosh this is what you guys actually do in a day they're just like like I thought you were just a YouTuber and then I come here and see this uh it's really cool so you could you could see all these things by going to meet kevin.com learn more about uh about what I do and how I provide value so check that out and I look forward to seeing you ther
Comments 0