Elon Musk *Critical* Tesla Stock Flip | FSD & Robotaxi

hey everyone me Kevin here so Elon Musk just dropped the price of full self-driving to $99 per month he left the full price option where you just pay once and then you don't have a subscription fee at $122,000 which really is probably a sign that they're not really converting many of those $112,000 individuals and so they're purposefully leaving that price at 12K because now you're not insulting that many people who paid 12K recently although there are going to be some uh and what you're doing is you're trying to drive this value for the $99 version by showing people look it's going to take you like 10 years to break even if you spend 12K so why not take this reasonably priced offer of $99 so now the question is what is this actually going to do for the fundamentals of Tesla does this actually change the thesis for the stock so what I did is I put together a spreadsheet and we'll put it up on screen and it is basically three different like impacts that this does for the stock uh now I purposefully went out to 2026 for this rather than going out to you know 2030 or 2040 or something crazy because I I want to see what the near-term impact for the stock could be and really the only way to do that is going out um honestly 2 and A2 years is already quite a bit the end of 2026 I think if you go out uh any more than that you're you're not going to see that immediacy in the stock movement and I think that's what most people care about so uh let's talk about this as we uh look at that Florida beach this why I love Florida I grew up in Florida uh okay so the first thing we're going to do is we're going to look at what I consider my base case scenario and there's a big problem that a lot of people have a lot of people are not realizing that there are costs associated with full self- driving a lot of people think Kevin full self-driving is going to be 100% 100% to the bottom line it's the best margin driver ever there's no capex and it's just going to be pure profit so there's realistic and then there's unrealistic and unrealistic is no cost there's a lot of cost you have to keep in mind they just spent probably somewhere around $3 to $500 million buying 10,000 h100s and guess what those are now the old school version of chicks Blackwell which will be available towards the end of the year you're probably able to train FSD four times as fast some of the newest chips so you're consistently probably to have the latest and greatest chips going to have to spend just on chips alone probably and and I should just say graphic chips which that's literally what FSD is trained off of right Graphics video graphics like you might be able to move some of these chat-based neural Nets off of some of these gpus in the longer term but you know maybe to like A6 or something more more specific to just Tech space but video video neural Nets it's always going to need the latest and greatest stuff to really get better because it's not just a matter of training at once like getting to the version you know 1234 that we have now that's great but really to iron out the rest of these edge cases we're just going to need a lot more training a lot more compute by using a lot more data like quite frankly the fact that 1233 and 1234 is making left and right turns from center lanes totally dodging the left and right turn Lanes is weird there should be plenty of data to tell the systems this is not what it's supposed to do now I'm not saying those things aren't solvable in fact we could bandage them tomorrow just by writing a line of code but that defeats the purpose we're trying to get away from writing lines of code so we need this data we need the training frankly we need better chips to really advance I think one of the reasons we got to 12 versus 12 frankly the anti- neuronet is because the h100s came out and so obviously that's a mode for NVIDIA uh but that's just going to get as the chips get better and better FSD will get better and better and these edge cases will go away but that takes cap X and that's what a lot of people forget is you're probably throwing 500 to a billion dollars at chips alone once you consider that we don't just need gpus but we need CPUs we need memory we need server racks uh and and then we need infrastructure uh and then we got to pay for that infrastructure the electricity uh the the power the the people who are maintaining the servers the people who are upgrading the servers whatever and then of course you need the R&D folks who actually uh implement this into the systems on Tesla vehicles the software updates to push this to people uh there's there's a lot that goes into it so I think initially we're probably looking at spending quite a bit uh of fsds Revenue on R&D so your AI training expense uh your chips whatever the things I just mentioned so I think if we go with my base case scenario we're probably looking in 2024 if we had a 12 1 12% take rate keep in mind that's already adjusting for the fact that some cars are never going to have F because if you have a fleet right now of about 82 million vehicles uh worldwide by the end of the year so that's assuming you know another 2 million of production we're probably going to and that might be generous there's a percentage that are in other countries where we don't do FSD there is a percentage that uh are scrapped vehicles that just got destroyed I mean obviously and there are a percentage of people who are just never actually going to want to buy a FSC uh because some people don't drive a lot you know if I'm uh if if I'm somebody who uh you know if all I do is I drop my kids off at school 4 minutes away and I pick them up and that's all I do twice a day 5 days a week do I re really need FSD for 10 minutes of driving a day no of course not so there's always going to be a percentage of people who won't get it and so hence when we Factor all of these things in the other countries scrap Vehicles people who just refuse to get it uh and people who have already paid for FSD so you're not getting that recurring monthly Revenue you're going to be looking at probably a pretty low take rate initially hopefully it's in excess of 12 1/2% but let's assume that it's 12 1/2% right now uh and then what we're going to do is we're going to multiply that obviously by uh uh the9 $99 a month time 12 right okay so let's go ahead and take a look at the base case scenario and explain how I'm doing this so if you look at the chart you'll see 8.5 million yeah that's always loud 8 half million Fleet size multipli by a 122% take rate you get the value how much money that could generate 1.2 billion take out how much you're probably spending on AI expense I think $800 uh million is quite frankly low uh I think they'll probably spend more than that but let's just say it's 70% uh divide that by shares outstanding you're only adding about 12 cents to EPS which adds about 4% to EPS it basically does nothing uh for EPS at some point in the longer term honestly I think you'll probably want to just include FSD uh in the actual price of the vehicle but then margin will go up so whether you calculate it here on take rate or you calculate it as part of vehicle margin you have to put the value of FSD somewhere so for 2024 on my base case I don't really see a big move in this $99 feature it's it's not as valuable as as I think we hope now if we go out to 20126 I think we'll be able to move that take rate to nearly a quarter of vehicles that'll be about 22 1.2% of vehicles and I think our AI expense will still be around a billion 1.3 billion is what I wrote down about 40% and uh that add 63 cents to our EPS which if we add a 40x multiple on that because hopefully we're growing at you know we're growing EPS at 25% a year a 1.6 Peg right we multiply that out we get a 40 multiple that adds about $25 of Share value to the stock in 2026 which I think is fair to look at now in fact I think frankly if we didn't have version 12 FSD Tesla stock would probably be around $140 maybe $145 so if you look at the last 30 days the last 30 days have been pretty wild for Tesla think about it for a moment you went from version 11 in the last 30 days to version 12 121 122 123 12 31 32 33 34 you got half uh off on monthly FSD Revenue you got an August 8th Robo taxi announcement and the potential cancellation of the $225,000 car which I know a lot of ultra Tesa you're like why is everyone so concentrated on a $25,000 car well I'll explain that in just a moment but the last 30 days have been crazy and so one of the reasons Tesla stock is the fact that it's flat over the last 30 days with the horrible q1 deliveries number is actually a sign that the market is adding back in some of the value of FSD or Robo taxi or or whatever otherwise the stock would be down substantially more of the delivery numbers that we had so what does this mean well this means we're we're probably already factoring in the base case into Tesla stocks value now and I think what we do is we can look at Bull cases and Ultra bull cases in my bull Case by 2026 I think we have 35% adoption and only 20% in expenses uh and that's going to drive about $52 in Share value I do not think that's being priced in right now uh and then in the ultra bull case which I personally think is unrealistic in the ultra B case we're only adding about $80 a share in value by 2026 which we would start pricing in relatively soon now that is pretty good especially since I think the vehicles for the company you're probably adding about $140 to the share price maybe $120 to the share price so 80 bucks should put us in the low 200s in the ultra bll case uh the problem with all of this is we can't really understand what the take rates are for Tesla because they don't give that data to us so without having that Clarity we're kind of at a loss if we don't have the clarity on what take rates are now then how can we know if the things we're changing like pricing or otherwise are actually being respected by the markets we don't and we don't really see a lot of FSD Revenue recognition in each quarter so it makes it very difficult as an investor to have Clarity this lack of clarity has always made me excited for the icing on the cake at Tesla but I've always underwritten Tesla as a car company that's growing rapidly and so for me when we're able to grow EPS by 30% a year well I can give Tesla a 50p ratio and when we can get to 5 million vehicles at some point in the future you know this is a $400 $500 stock for me just underwriting on cars but that thesis has falling apart now because Elon is shifting from something that is very clear to underwrite vehicle deliveries and a cheaper car to we're going all in on Robo taxi but Robo taxi comes with this massive level of investment that first of all we're not near Robo taxi now we're probably 10 years away from robotaxi and not only are we probably 10 years away from Robo taxi when Robo taxi first starts you're still competing with Uber and I know people think to us how the hell are we competing with Uber we're going to bankrupt Uber they have Labor right but you're trying to get Uber's Revenue so if I'm going to spend 20 bucks for an Uber for you know a 10 or 15minute drive or 20-minute drive one of the reasons I'm willing to pay for that is because I value my time and I'm not going to sit at a corner you know waiting for an Uber for 30 minutes or a robo taxi for 30 minutes if I could get an Uber in 2 minutes you need a lot of vehicles a lot of vehicles to actually get people using the robo taxi Network you need FSD perfected Vision based FSD perfected we're not close to that then you need people saying wow the robo taxi is better I don't have to talk to a driver it's cheaper which that lowers your margin anyway uh and I don't have to tip a driver this is great but they still have to manufacture the car you know even if you create a $30,000 Robo taxi 80% of that car so $24,000 of that car is still going to be cost for Tesla we're not actually convinced they can manufacture a car for $24,000 yet so it's not like these cars are free and you know Uber is doing millions of rides a day so in order for us to pull that off I mean we're going to need hundreds of thousands of vehicles so we're going to need one to two quars of of robotaxi vehicle production just to be able to start competing with Uber in in certain markets first where you get your delivery times down low enough at a reasonable price where people are willing to take a Tesla Robo taxi versus an Uber or a lift a lot of competition there so yes even though there is a human there and even though you know Uber spends somewhere around oh boy they uh they're they've got about $37 billion of Revenue they spent about $22 billion just on the labor so more than half is just going to the drivers right that's all future potential margin for Tesla which is hopium because the price is just going to go down the price of providing the service is going to go down you know if you take $22 billion of cost away somebody else is able to operate Uber basically for uh in that case that'd be about $15 billion so they can drive the prices down by 50% so the $20 ride is now a $10 ride and no tip which is great but again Des has to provide the actual cars somebody has to charge the cars the cars have to be maintained who cleans the cars between each ride the person who throw up in the back right uh and that's all assuming we're at level five autonomy so I I it it's great but like it's great hope but as an investor trying to trying to price it in I can't price Robo taxi in yet I I can only price in the hope of it but what it can price in is some reasonable level of speculation on FSD which that's why yes I think there is an ultra bull case where the expenses are virtually nothing for FSD and the take rates are like you know upwards of 45 to 65% but I think it's heavily unrealistic I think right now we're maybe adding 25 bucks a share now the difference is if we can get EPS growth from this and we start proving wow we're really growing from this then guess what happens if instead of valuing Tesla at a 40 multiple you start pricing it in at an 80 multiple on future growth because now you're getting EPS growth from software which eventually will be higher margin they could prove and this is the real hopium method they could prove to be the most profitable software AI play that exists I actually think that there's very little profitable software right now uh that's AI may maybe maybe maybe paler let's get windy uh P tiers an option uh I think most of the companies like uh you know the sales forces and the msfs you know they're they're trying and I think they can they can probably boost their sales a little bit on Office 365 oh get it with AI it's $5 more a month whatever problem is that's going to become so so Baseline that I'm not actually convinced you're going to see such Revenue explosion uh at at at some of the core software companies now that have run as much as they have on AI so I think those are a little bubbly I actually like Nvidia more than I like some of those plays but realistically Tesla's one of the most profitable AI plays or will be one of the most profitable AI plays the question is and that's from a revenue side the question is how much are they going to spend to get there on uh on cap backs and when are we actually going to get transparency on how much EPS growth we can actually expect from these things if Tesla can lay out a road map for here's how much we think we can do in earnings per share and here's how much revenue we think we're going to make based on a certain take rate that's when this stock can actually finally unleash and take off because you're going to get away from just focusing on vehicle deliveries which we know are in the toilet right now and you can actually start underwriting okay rather than just a hm point of view we can actually underwrite how much can we expect to show up in earnings for the stock okay we're going to I mean elon's done this before he just tells us look we're going to Target uh 60% take rate on new vehicles and a fleet take rate of 30% which we want to grow to 40% over time uh and then what we're going to do is uh you know once we get this going we're going to f we're still going to focus on a compact car you know it's not going to be called a model 2 or whatever we're going to call it who knows whatever but you're going to be able to get it as roboox incapable this we just get a vision then we can actually price the Stu so that's my take hopefully that's useful and if it is let me know in the comments down below I made some money on uh trading some uh Tesla calls yesterday it was really good Tesla was just like straight up and down uh yesterday it was down 2% on the day but I bought it was down 2% so I was arving the difference between that 2% and like 1755 and uh we were killing it on some calls yesterday so if you want all those Buy sell alerts and trade alerts make sure to check the link down below stocks and psychology

Channel: Meet Kevin

Views: 51,379

Keywords: tesla, tsla, fsd

Id: m6QWDRIGsAw

Length: 17min 54sec (1074 seconds)

Published: Sat Apr 13 2024