저자: Kevin · 외국인
요약 보기
테슬라가 하드웨어3용 FSD v14 라이트를 출시해 소비자 불만과 집단소송 우려가 완화될 수 있다는 주장. 테슬라 주가는 당일 8%대 하락 후 반등 시도 중이며, 단기 상승 여지를 보나 SpaceX 락업 등으로 1년 내 SpaceX 주가에는 비관적. 테슬라 밸류에이션은 자동차 기준으로 비싸다고 보며, 2027년 말 목표가 312달러를 제시하나 본인도 성장률·마진 달성 여부에 불확실성이 크다고 함. 모델3/Y 성장 둔화 전망과 Cybercab·Optimus 기대가 대비됨.
다듬은 글
Today was an important day for Tesla, and it comes in the same week that we will get delivery volume information from Tesla. This is all very important for stocks, and we need to talk about it . First, we'll talk about the 14 light hardware. Then we'll talk about how this could affect the stock and what the value differences could be for Tesla.
Why is this important? So, hello everyone, I'm Kevin. Let's get straight to the point. So, today Tesla shares saw a price drop of 8.46. We literally went from $379 a share at the market open all the way to rejecting my famous trend line at $414. We could see this deviation here, within 73 cents for those of you who like TA, all the way through the end of the day.
Uh, and we're stuck at about 411. So the big question is, why has Tesla moved so much ? Keep in mind that it's worth noting that the alpha report we published this morning indicated that if the trend is positive, then equipment manufacturing companies like Tesla, which have recently sold off their shares, could see short-term upside.
Also, in this morning's alpha report, we also talked about SpaceX and said that SpaceX shares were up about 2% right then. They said, “Look, SpaceX’s NASDAQ IPO date is July 100–7, and many lockups don’t happen until August, while people who made a quick buck on their $135 IPO purchases were able to sell the shares after their price rose to $220.
And who cares if they got a small fine, they weren't blocked. I think the selling pressure is over in the short term . We have a buzz around SpaceX right now that is really exciting. So you might have a few weeks where we really have an upward trend. However, over the next year, thanks to the lockdowns, I am a little more pessimistic about SpaceX's stock price.
And so in this video we'll focus on Tesla. Now remember, if you want to get tips like the ones I just gave, the market is open every day . The market is open almost every day , join us at mekevin.com. Our coupon for the "Pope" coupon expires tomorrow. Uh , and that's because the Pope gave... They're nodding at me, and I'm very happy about that.
Remember, you can potentially write this off. , short-term, medium-term, long-term trade alerts, reinvestment course, diversification, debt, trading, insurance, bonds, legal entities, real estate, finances for your children, anything . But now let's get to what's important, which is Tesla. So , Tesla, the 3 hardware is what people have been stuck with for about 15 months.
Er , well, or longer. I mean, people who had hardware 3 were stuck on version 126 FSD for about 15 months. That's a long time to sit back and wonder: will we ever get an update on FSD? And today Tesla announced the release of version 14 light for fully autonomous driving, and that was specifically for people with hardware 3.
So in other words, people who had older generation cars, like before 2024, those people...Just got a really important update. This also came about three weeks after Tesla was caught up in a failed class action lawsuit, where the person who prepared it claims that Elon's first-quarter earnings report on April 22 was effectively an admission that full autonomous driving would never come to Hardware 4 users.
This angered a lot of people, and many at that point became pessimistic about Tesla , thinking it was nonsense. You know, since 2017, many people have been given the idea that we will have fully autonomous computers. Some of the older computers have already been upgraded to newer versions, but not to Hardware 4, because, as Elon pointed out, this would require the creation of many factories, essentially in cities across the United States.
He said: " We will have to create micro-factories in large metropolises." “And Elon admitted that Hardware 3, as I wrote, simply does not have the ability to achieve full autonomous control without supervision. This admission likely led to this class action lawsuit. So now, in response to the class action lawsuit, it seems like maybe it's just convenient timing , but it looks like Tesla is finally saying, "Okay , we'll take care of Hardware 3, people." Otherwise, if you don't pay for it or get such an update, you'll likely be stuck waiting for some kind of hardware update for your car with Hardware 3, which can cost anywhere from $300 to $5,000.
And many people said: " Look, I already paid 5 thousand, or seven thousand, or 8 thousand, or 10 thousand, or whatever it was for FSD." Why don't I have FSD? Why am I behind on updates? »“And of course, that’s why this class action lawsuit was filed. So what is version 14 light and how does it actually fundamentally change the standard version?
Well , it really does matter. That's good. So, hardware 4 is more advanced hardware. We already know that 5- megapixel cameras are in hardware 4. Do you know how many megapixels are in hardware 3? 1.2. That is, part of the resolution . The 4 hardware is said to be about three and a half, some say up to five times faster in computing power when handling full autonomous driving workloads, simply because the chips are better.
But the point of full autonomous control is that it's not just about maximum power or how much computing you can have. Often it all comes down to memory. It seems like everything these days comes down to memory. And hardware 3 only has 1/8 the memory bandwidth of hardware 4. 1/8 is not great, right?
1 divided by 8 is approximately 12 1.5%. So what is a version? 14 light that was just released for hardware 3, people? Well, after 15 months of being stuck on version 12.6 and skipping version 13, they are getting version 14 light. And it's expected, and I've seen it quoted in a few different ways, to be around 14 to 15% of the hardware size for the neural network.
And this was done, essentially, by distillation. In other words, the model was compressed by distillation. Distillation is something like Tesla's deep search. In my opinion, this is quite brilliant. Distillation training has been going on for about ten years. So it's not something super-novel per se .
But the fact that Tesla is distilling its own advanced fully autonomous driving system to help manage old and outdated, sometimes, you know, 7-8 year old, vehicle hardware . Well, some of them have already upgraded their computers, like my dad's Model X, which is from 2017. That's what brought me to Tesla, which had a hardware update between...and now.
But the point is, now he's getting an upgrade for what's actually going on with the 9- year-old car, and he loves the fully autonomous driving of the 126, but now he's getting an upgrade in a teacher- to-student model through distillation. Let me explain this a little bit, because I know this whole " teacher to student" thing can seem a little complicated.
So, let's come up with a scenario to quickly explain distillation. Essentially, let's say you're driving through an area at 30 miles per hour. You're driving the speed limit , maybe it's 25, whatever. Uh, an experienced driver sees a row of cars clustered on the right that weren't there before. It's just a bunch of people parked on the right, and a ball is rolling across someone's front lawn.
An experienced driver says, “Okay, there’s a chance a child will jump out of this row of cars. Let's slow down. At least take your foot off the gas a little ." “That’s kind of the mindset of an experienced teacher in a certain situation, right? This is intuition. We have situational awareness. So we say, you know, we should probably do XYZ to minimize the risk of someone getting injured, and you just have a headache.
Good. " Hardware 3" is essentially learning. Good. Good. I don't understand all this nonsense you just said, Kevin, but I get it. If I see a bunch of cars parked in an area, I'll just take my foot off the gas. I'll slow down a bit. Got it. This is an expression of intuition. This is not about learning how to be so intuitive.
It's just saying, "Hey, dude, you know, in these 50,000 scenarios, different scenarios, you could have thought about doing this." “ Another quick example that I like to think of as a pilot flying. When I learned to fly a jet plane. These things are insanely fast . Uh, you know, there are times when I'm flying and I'm doing my checklist or checking the weather at the airport where we're supposed to land.
And then the other pilot says, “Hey, we’re at point XYZ. You know, something we should do here?” You know, they have maybe 10 years of flying more than I do, while I have one year. Uh, and I'm like, "Oh, yeah, sure." We should start the arrival procedure, right? Uh, for this area . So why don't we start, uh, with that, you know?
Set up navigation this way, however you want, right? This is a gross oversimplification, but the point is, the next time you go to that intersection, my instinct should be, “Oh yeah, that’s right, I have to do something right now. That’s right, I have to complete the arrival procedure at this moment .” This is one of the ways a teacher teaches, isn't it?
This is distillation. So it's not that good, but it's not bad enough because there are a lot of scenarios and data that people are talking about that Tesla has that other car manufacturers don't, and that, frankly, even Nvidia with its drivetrain platform would like Tesla to have, and now it can offer distillation to other cars in its fleet.
And that's great. This doesn't mean we're at a point where, " Hooray, we now have cyber taxis," but there may be a hint that cyber taxis are coming, and I'll show you that data in a moment. So this is version, uh, 14 light, and, uh, I, I think this person made, you know, a pretty good, uh, little first impressions video.
TechGeek Tesla. I'll just show them a little bit how it's done here. So, uh, they shared their first impressions. They seem impressed. Uh, I'm not going to play the whole video here, but here's just an example where he says that now he can, which he couldn't do before , pull up and the car will park itself in the driveway.
He tried to persuade her to pull into the garage, but she simply pulled into park mode. And he was a little disappointed that she just went into park mode. Uh, then there was, uh, a scenario where she would n't go into park, and to go in, he had to press the gas pedal, but then she stops about halfway.
And so, obviously, it's not perfect yet. It won't be something like 14. But honestly, for me, there's a version 14 in the Cybertruck. I love it. I'm a big fan of FSD. I think they are simply the best product out there for fully autonomous driving, definitely, full stop. And I don't use the FSE14 for parking.
Maybe I should. I usually just have this habit, maybe because I've been using FSD vehicles for 9 years. You know, I started back in 2017, when everything was on autopilot. And I have a habit of just taking control and parking myself. And I also don't have a driveway or a garage where I park. So, you know, maybe I'm not the best person to say that I don't think it's that important to a lot of people.
Maybe so. This is additional convenience and comfort. This is great. But I think it's actually important in places like unprotected left turns, you know, driving through parking lots, and it seems like the person here, at least , is pretty impressed with the, uh, 14-inch light. Now there are some people who are not impressed with the 14- inch light.
Uh, some people say the response time for a 14- inch light is too slow. I don't know if this is a problem with the 14-inch light or if it's just a problem related to a certain scenario. For example, here's someone on X who posted this video and said that a car's brake lights come on, and there's a delay between them coming on and the car reacting, and that this delay in reaction leads to an accident.
Personally, I disagree with this because just because a car's brake lights come on doesn't mean the car has physically started to slow down. Usually, if you just apply a little pressure on the brakes, you can turn on the lights without slowing down the car. It's like you activate the brake light actuator before the brakes even start to hit the brakes to slow the car down.
So since the brakes in vehicles are physical, I think that matters, and that might be a partial explanation for the 6, 7, you know, 67 second delay, but I don't know, it depends on how fast they did it. For example, if you're going to fly a jet, a lot of jets, well, the Phenom specifically has electronic braking.
Uh, it's brake-by- wire, which is really weird because when you press the brakes, there are ceramic brakes that get hot. It's like there 's a delay, and if you fiddle with the pedals to brake the car, because there are two brake pedals, a left and a right one. If you play with them , the plane will go like this when you stop.
It's like "footballing," as they call it. This is strange. In any case, the bottom line is that this is electronic type braking. The vehicles do not have electronic brakes. So there may be a gap in the time delay. I'm not as impressed as this post, but maybe Hardware 3 is slower. But that's about the jade-ness I've seen so far.
Uh, the comments I see are, hey, he can park himself. Steeply. Excited. I am personally glad to see my dad's comments on this . You know, again, this is a 9- year-old car. Some people say it's smoother motion control. Uh, and the deployment is expanding. Okay, cool. Perfectly. So, uh , does this move stocks ?
In my opinion, yes. So, first of all, uh, it reduces the potential need for micro-actors. I think micro-actors will be a huge distraction and extremely expensive for Tesla. If they pass the cost on to the consumer, the consumer will get angry , and then they will sue Tesla, as they already do. Uh, all this comes after Elon Musk said in the first quarter that Hardware 3's quote simply has no chance of achieving unmanaged FSD.
We will have to create micro-agency in large metropolitan areas and, in essence, offer quotes for retrofitting old vehicles . This is not good. Especially considering that there are many Tesla “bulls” circulating around the world who say : “I can’t wait to add my Tesla to the fleet so it can mine Bitcoin while I’m not using my car, drive itself and be my robo- taxi, and my car can be my slave .” I think we are quite far from that.
But let's just say, uh, this lightweight V14 release will probably significantly reduce people's concerns about not getting updates for the past 15 months, and it will save Tesla's EPS. Upgrade kits will likely not be available for version 14, nor will the actual version 14, until mid-2027. And honestly, I don't think they ever will.
I'm still waiting for my towbar that I ordered in 2017, which Tesla sold me as an option , and then it didn't come with the car, and then they said they would send it to me, but never did . I do n't want him anymore. Either way , someone at Tesla is watching. I do n't want him. Everything is fine. I got over it.
I got over it a long time ago . My dad still drives my black X now and loves it. Uh, it's actually incredible to see, you know, someone who's 40 years older than me, so excited about full autonomous driving and really passionate about it. But anyway, the website electric, which is usually a Tesla fan, claims that this is just a way to move the goalpost.
Personally, I think this is a great move, and it's actually quite impressive on Tesla's part. Uh, I argued again in the Alpha report this morning that Tesla had a good chance of breaking out higher today. Uh, and the fact that it moved 8% is great. I hope many of you made a lot of money from this. Uh, I'm not always perfect, but I do my best for my challenges, and that's what I've observed.
It was a kind of Tesla turn, if you will. Uh, now if we look at Tesla's valuation, they're pretty expensive. Uh, it's actually just a car valuation for the company, right? So, if you multiply them just by a factor of about 1.67, Tesla's valuation now looks pretty insulting. Uh, that's a fair value for an automotive and energy company, uh, and assuming a growth rate of $124 over the next four years .
This is very, very low. Obviously , if we take Optimus and Cybercap into account, that's where the extra $300 will come from. Now that we look at deliveries, perhaps this is where we can see the emergence of some CyberCabs. But look at this, uh, I took...Oops, that's not my spreadsheet. Let me bring my chart here because I want you to see the pace of change.
So, I have a consensus table on Tesla deliveries that Tesla published. And then I set growth targets in percentages. So you can see the Tesla consensus. You will see that after 27 years, the consensus is that Model 3 and Y deliveries will only grow by 8%. They will grow by only 7% in 2028, 10% in 2029, and 8% in 2030.
This is much lower than the historical 50% growth rates we had for cars. So, the situation is pretty bad on the Model 3 and Y side . However, look at this. For all other models, we could potentially see growth rates from 63 to 95, then from 45 to 39%, which would bring us to 362,000 cars by the end of the decade, which would be about 15% of all cars Tesla produces.
What will lead to this leap? Well, in 2028 you will have a big increase in cars. This is almost twice as much as all other models. This is where, I think, most analysts start writing some kind of releases regarding the cyber limit. Again, it's obvious that a semi-trailer could also fit here. We know the cybertruck isn't selling very well, but semi-trailers and cyber goods are probably already here.
You can also consider Optimus. Sure, energy is booming, but it's also not growing, at least according to estimates, at a growth rate of 50%. Previously, when Tesla lost its 50% growth rate in cars, they said, "Well, at least we have 50% growth rate in energy." No longer by consensus, it's expected to be about 34% between now and next year, then 31, then 26, and another 20.
So not fantastic. But, hey, there's also a lot of optimism that Elon expects to buy Tesla for SpaceX shares if, well, you know, if Tesla shares fall enough and SpaceX goes up. And this creates a certain floor for Tesla's price. Now, if I enter, let's see what the ratings are now. If I go to Wall Street estimates, and we can estimate this ourselves, we will consider a higher growth rate.
Let's take the reference 2.67 . Let's take the earnings per share at the end of 27 . Okay, let's look at the end of 2027. Let's be a little more optimistic. And then we take into account the growth rate of 3756.51 divided by 3, the growth rate is 48%. Okay, if we take these numbers, we can justify a much higher valuation for Tesla.
Okay, here it is. $312 would be a reasonable target price with higher service-based margins for Tesla. Assuming 48% growth, that would be, uh, the target price for the end of 2027 . Now if I stop at the producer about 1.6 67. Obviously we're going to see the price drop below $200, closer to what we've seen, there's such a bearish valuation, probably because that growth is closer to about 40-42%, somewhere there, right?
So, uh , you really need this growth to stay at that level. You also need them to be a higher margin manufacturer, which you know they are not right now. Now their pricing power has increased a bit, but only because margins are improving somewhat. Uh, and that was, uh, partly due to the rate cuts along with some additional FSD accounting.
So it is unclear whether the margin decline will persist. Let's say they are also limited now regarding battery packs . One more thing to know. And then, of course, you want to see wider implementation of FSD internationally. I mean, when will China approve FSD? Right? So there is still room for growth in the rest of the world.
Again , I'm an FSD fan. Personally, I'm not thrilled with the Tesla valuation or the space acceleration, but then again, the Elon Prize, you know, so you just need to know that starting with that. Well, that's my opinion, anyway. Thank you very much for watching. If you want to know my daily thoughts on short-term, medium-term, and long-term price targets, be sure to head over to Kevin.com, use the Pope coupon before it expires on June 30th, and I'll see you ther.
Comments 0