Transcription
Tesla just reported Q2 earnings. We're going to get to some highlights from the call in a second. But first, let's look at the actual earnings report itself.
Starting with the most important and mind-blowing stat of the entire quarter. Make sure you're sitting down for this. FSD adoption grew in the quarter. No [ __ ] Sherlock. Including record net new subscriptions. Wait for it. We also achieved record FSD attach rates in North America with over 55% of new deliveries including FSD subscriptions. Holy [ __ ] bro. That is an enormous number.
Honestly, we could forget literally every other word in the entire report and every other number. I want to talk about why this is so important. North America is Tesla's most mature market. FSD supervised has been available in North America longer than anywhere else. And generally speaking, it's more capable in the US than anywhere else. This is an insight into what other markets are likely to look like in terms of FSD take rates when FSD supervised is equally capable and has had a little bit of time to permeate through word of mouth.
This is before FSD unsupervised at which point the value offered by the software immediately increases an order of magnitude or more. The difference between quote unquote having to supervise, which basically means doing nothing but at least still pay attention versus being confident enough and legally able to pay no attention while the vehicle drives you from point A to point B while you do literally anything and or anyone at the same time. At that point, the gap obviously is going to close between take rate at whatever it is and close to 100%. The fact that we're already now seeing in the US in a single quarter FSD take rates of over 55% is batshit insane. I was not expecting this. Eventually, over time, yes, but this is pre-supervised and we're already above 50% in the most mature market.
What the actual couple of big operational milestones a record Q2 for vehicle deliveries. We already knew about this. Robo Taxi launched in three cities in Florida. Pause. In July, pause. In the span of less than 3 weeks, pause. And began production of CyberCap. For the first time as a company, Tesla generated over hundred billion in revenue on a trailing 12-month basis, only a 10x away from a trillion. and quote, "Progress also continued on battery capacity expansion, the main limiting factor to near-term vehicle production volume increase." Sorry, what? Tesla battery constrained again. What? And importantly, construction at our Fremont factory for Optimus began after we decommissioned the model SNX lines with anticipated production later this year.
Won't get too buried in the numbers here. If you guys want to go through the whole report, you know where to find it. If not, link in description. But one massive outlier here, Tesla services and other revenue up 50% year-over-year. What the [ __ ] is going on there? Also worth mentioning, automotive revenue up obviously quarter over quarter by 26% but also best quarter in the past 12 months despite not only deleting Model SNX which typically sold for two plus times what a Model 3 or Y goes for massively skewing up automotive revenue per vehicle but also the introduction of the entrylevel Model 3 and Y the standard the lowest starting price of any Tesla vehicles explained pretty well on the next slide. Model 3 and Y deliveries year-over-year increased 25%. Might explain why Tesla is currently battery constrained.
And just to remind everybody, Tesla's active FSD subscriptions, we heard earlier, 55% of all new US Tesla vehicle deliveries in the quarter included an FSD subscription. They now have almost 1 and a half million active FSD subscribers globally despite FSD not being available still as I record this in most countries around the world. Bottom line, year-over-year increase in FSD subscriptions, active FSD subscriptions, 56%, bro. Holy [ __ ]
A nice little tidbit here. One interesting note for the number nerds. Capital expenditures for the quarter almost $6 billion yet. Tesla's cash on hand. Cash and cash equivalents went down by about 1.2 billion bucks. That's it. Essentially spent nearly $6 billion, but their cash pile went down by just 1.2. Not bad.
Under AI inference and compute, we continue to make progress on construction and equipment procurement for our semiconductor FAB in Austin. Very weird thing for just a car company to be building a chip fab. While still in its early stages, this project is critical to building our own chipm capabilities to ensure reliable long-term supply of essential logic and memory chips for our products. Something something terra.
Taking a quick peek at cumulative paid robo taxi miles, we can notice that there's a slight softening in the quarter. Not major, but worth paying attention to. Great signs for cyber cab production, of course, as we know began in the quarter. It's hard to hide a lot of this, but Tesla letting everybody know that somehow missed all the posts on X, people spotting them. Engineering test drives of production cyber cabs on public roads also began during the quarter, and we began offering employee rides in cyber cabs on our Gigafactory Texas campus in July. Both important precursors to cyber cab deployment in our robo taxi fleets. For the record, as I've said recently, I'll be pretty surprised if we don't see the first cyber cabs in operation this quarter, as in Q3. I won't be shocked, but I'll be pretty surprised.
On to the earnings call highlights. Starting with Elon's opening remarks. A few boring bits we already knew. Record Q2 deliveries. Model Y still absolutely killing them. We knew this. On to the very high FSD take rate. Musk disclosed something very interesting. He's hearing about customers actually going to Tesla and asking, I want to get FSD with whatever car I can get it with. Essentially, people are buying FSD supervised with a car attached. So FSD, the software itself, is becoming an increasingly significant demand driver for Tesla vehicles. The energy business continues to grow incredibly fast and in today's episode of obvious things are obvious will be crucial for AI data centers. Tes are investing a lot in growing the business.
What's new? Musk flagging big capital expenditure this year but expecting it to yield incredible returns. Maybe the best ever capex returns Tesla has ever seen. Lots of stuff happening or about to. Cyber cap production commenced. Optimus production soon. Semi production underway. Mega Pac 3 production soon. We've got the lithium refinery, cathode refinery, battery cell production, scaling, and Tesla also flagging an intent to massively increase production of solar cells.
And for the 12 people who didn't already know this regarding robo taxi, Tesla has very ambitious goals, but needs to be very careful and cautious about causing accidents. The negative press factor, it'll be global headline news if there's a single serious accident involving Tesar taxi for months. And it's not that Tesla does not care or should not care about negative press. Who cares, bro? Just roll out the taxi. Like, whatever, bro. The issue is that the result of unreasonably and unfairly negative headlines when it comes to a Tesla robo taxi will scare people, NPCs who don't have any context or understanding of big numbers, relative risk away from using these products. End result, more people dead, injured, and permanently disabled because they got scared by negative and unreasonable coverage of an incident involving a Tesla robo taxi and then didn't want to get in one end up driving themselves around using something else, getting wiped out. So, this is not a self-interested, oh, we don't want bad press cuz bad press is bad position from Tesla. It's a bad press will cause people to die who get scared of our extremely safe robo taxis. So, we need to be ultra cautious. In addition, of course, it's not just that, but in his own words, Musk says that Tesla's moving as fast as humanly possible, but making sure we do so without harming anyone at all. Literal quote, as fast as possible without harm to anyone. As I've said before, it's Tesla limiting how quickly they're scaling robo taxi in cities. Now, obviously, well, they haven't launched yet. Different story, but they're currently in seven US cities and counting at relatively small scale. They could be operating many more robo taxis in those cities today.
For updates on robo taxi. As most people will know, Tesla posts live updates on X. Every time there's a new launch, you'll hear about it on X. Musk also adding something quite interesting which addresses something I said earlier to keep an eye on. Cumulative robo taxi miles expecting to scale rapidly. Duh. And perhaps an increase of 10% in miles per week. I think I heard that correctly and maybe I imagined it. Let me know in the comments below. I'm pretty sure that's what he said. And presumably that's a 10% week-over week increase in miles traveled. So let's just say one week it was hypothetically a million miles and next week it would be 1.1 million miles and so on. That grows compounds very quick.
On Optimus Elon making sure the Elon is far too optimistic crowd. I could never trust him. Folks don't have too much to complain about with the Optimus ramp explaining it's a super difficult problem to get the design right and also to scale production. It needs to function well obviously but you also need to consider durability and wear. He adds that this will be the hardest product to scale manufacturing that Tesla has ever done. Literally everything is new. There's no supply chain for any of the parts. Whereas automotive, I mean there's existing supply chains for tires, glass, windows, like there's lots of stuff in vehicles that already has established supply chain. That is not the case for Optimus. Everything is new. a Musk warning explicitly and extremely clearly that Optimus ramp will follow an S-curve, but the initial portion of the S-curve will be quite flat and long due to the newness of parts. So, if anyone, any investor has anything to say, any complaints about how slowly Optimus ramps in the future, refer back to these clear statements from Musk. It's going to be extremely difficult. It's the hardest thing they've ever done, and the beginning of the production ramp will be very flat and slow on Terapab.
Lots of juicy nuggets here too. Tesar expects to announce a location and more details, a full launch announcement, not just on a news call, but like a full presentation blah blah blah soon. And it will be a very big announcement. Without Terrafab, Tesla will be limited in their ability to scale Optimus production. Musk then adding that they must solve memory, logic, and packaging, which will be done by Terapab to scale Optimus. So this is a very clear signal that they're not [ __ ] around when they're talking about millions, tens of millions, eventually hundreds of millions, then a billion plus Optimus per year. The emphasis terapab is necessary to scale Optimus. Not terapab is necessary to scale robo taxis for Optimus. So read between the lines. And we also learned that Tesla has placed equipment orders sometime during the quarter presumably for their development fab in Austin on digital Optimus. Shout out to the macro hard project. Just some clarification. It's already been discussed previously, but a little bit more color on this. Driving a computer screen is essentially what digital optimus will do, which is comparable to driving a car. Excuse the typos. Essentially, whether you're driving a car or a computer screen, it's photons and or pixels in and controls out. Very similar problem. So Tesla very confident they can adapt self-driving tech for vehicles, cars to self-driving tech for computers. And importantly, the AI4 computer can handle this work. Now, this is a big insight for those with AI4 hardware in their Tesla vehicles today. In the future, you may have the opportunity of opting in for your vehicle while it's not in use to be doing work driving a computer and get paid for the privilege. And obviously, digital Optimus being able to control a computer is also useful for physical Optimus because if physical Optimus needs to interact with computer during the workday, it needs to be able to use it.
onto Megapod, something that's not getting a lot of discussion, but a massive opportunity. Musk gave some more detail. AO4 chips, an x86 computer insight, essentially digital Optimus in a box. So, these are basically plugandplay scaled down AI data centers. Now the thinking here because these boxes can be placed anywhere, Tesla can essentially really simply drop a bunch of mega pods strategically at particular supercharger locations where they have approximately 7 GW of power available globally today and then Tesla now has distributed AI with distributed energy at their fingertips at existing sites that they already own. This may not sound like a big deal, although Musk believes it will be. But think about this. Imagine just a really random number, right? Let's just imagine one megapod inside contains the equivalent of 5,000 digital optimists. 5,000 productive workers. How many companies on Earth today employ less than 5,000 people, the majority of which work on computers all day? This really could be a big deal.
Tesla's CFO noting their biggest automotive backlog since 2023. As a result, they are increasing production to meet demand. Of course, they are still currently struggling to get their hands on enough battery cells and a few other parts and components as well, but they're working on it. And again, in case you missed it, expect enormous capital expansion this year and over the medium-term. Tesla investing aggressively in their future.
After Tesla CFO made a few remarks, this was unexpected. So unexpected in fact that [ __ ] the notes, we can hear direct from Ashok himself. First of all, I'd like to state that the robot taxi program has been operating extremely well, especially in terms of safety. The program has had an impeccable safety record. We have driven more than 380,000 mi of unsupervised robot taxi now across six cities in two different states. We have had zero notable incidents. Any reports have been of other actors impacting us when we were stationary. Again, I'd like to emphasize how safe the operation has been so far. zero notable incidents over 380,000 miles. Historically, the so-called experts have always claimed that you need lightars, radars, HD maps, and the entire kitchen sink to drive safely. Here we show that such is not true. You can have safe, comfortable, and affordable autonomy with just cameras.
>> Absolutely amazing, bro. I swear I I literally thought for a second there he was going to say that they also needed radar, lighter, gear, sonar, smoke signals. This is important for two reasons, obviously. Zero major incidents. 380,000 mi. Safe as [ __ ] PS. All the experts said this was impossible. Now what, [ __ ] This was incredible.
By the way, I've almost finished my brand new 20-year Tesla valuation model and 20-year SpaceX valuation model. They're launching together very soon on Patreon. When they go live, prices for all new memberships will increase significantly. Join today and you can lock in current prices for as long as you remain a member. and existing members who want upgraded access. Best to do that now as well before the price hikes to join and to see exactly what's included at every supporter level. Head to patreon.com/solvingthemoneyproblem or just click the link below.
After the mic drop moment, Ashrock wasn't entirely done. He had more to say, pointing out that Robo Taxi is scaling at a rapid rate and that the end point of scaling will be entire states, not just cities. Also emphasizing the ridiculous safety of robo taxi. again reiterating that it is scaling rapidly disclosing interestingly that cyber cab production is roughly matching miles driven in their robo taxi fleet and here's a direct quote for the people who I think we featured a clip of them on yesterday's video if you know you know complaining about how quote unquote slowly and disappointing robo taxi is scaling quote robo taxi growth has been literally exponential while keeping an impeccable safety record. Also adding that the technology bets are paying off. Eg no lightar radar gad helping continue exponential growth for the foreseeable future.
Now I'm not great at speaking engineer. It's not my first language. But did ashock just say cry harder [ __ ] Robbit is still scaling exponentially. No, he didn't say that. I'm just kidding. But this is important context. Things are going very well. Exponential growth looks extremely slow at first. And to be clear, Roboaxi is scaling slower than I expected. launched a little bit later than I expected and is scaling a little bit slower, but it's still super early. We're still really early in the S-curve. And as we heard multiple times, there hasn't been a single safety critical incident, yet they're still scaling now in seven cities, 380,000 mi. No serious incidents. Hello.
Investor questions. Somebody asking, "How exactly is Optimus going to collect data for training, improving, getting better?" There's a few different methods here. Optimus can learn by observing humans. For example, the tens of thousands of Tesla humans operating in Tesla factories. Optimus can watch and learn. Tesla also has a fairly small dedicated data collection team essentially doing tasks to supply relatively small but very high quality data to Optimus. And then there's the entire internet can watch videos of humans doing things. Plus the Optimus Academy, which we've already heard about to close the gap on skills.
Endless questions. Someone asking about regulation, roll out of robo taxi. Tesla pointing out that they have great federal government support, although a little bit disappointed in a recent decision in a recent location in the United States that seems intent on self-destructing. Uh, let's move on. One of the analysts asked about robo taxi milestones to watch. I mean, the March of 9 and also whether Tesla would consider using third party distribution, he partner with Uber Musk saying that he expects Tesla to be vertically integrated. So reading between the lines there, the answer to that question is no.
Asked about a Tesla Space X merger, I personally expected Musk would say, "Bro, can't talk about that on a Tesla earnings call. There's a procedure in place, you know, but we did get something and then nothing and then more. So let me explain." Musk immediately says, "Well, you know, there's lots of overlap, especially the Terafab project." Then he says, as expected, I can't really talk about that on a no call. You know, it's not not the appropriate place. Then he starts up again after saying he can't really talk about it. Goes and there's also Grock in the car. Grock for driving Optimus. We have Starink in Cyber Cabs and Starlink will be integrated in all future vehicles as well. Tesla CFO also chiming in. God the typos are no joke here. Adding that Starink in Cyber Cabs is a big deal for those who aren't aware. Stalin connection in every cyber cab. You want to watch 4K video, get some work done, no problem.
One analyst asks, why not scale robo taxis up in a single city? for example, Austin instead of new cities. So, it's amazing to me, by the way, that an analyst asked this question. I've kind of explained in detail previously, but Tesla themselves making it pretty clear. The whole point is so they can gain a foothold in new cities by making sure, first of all, to get the approval and second of all, it's actually working in a new city. A direct quote here on trying to scale is that Tesla's attempting to gain a broader footprint. The idea is once you have regulatory approval in a city to solve things in a small fleet, in a controlled manner in that city before scaling more broadly. This is the entire point. I've talked about this. Tesla's planting seeds everywhere they can. The 0ero to1 moment is getting regulatory approval in a city and being confident enough your software can operate safely in that city, launching in that city, and discovering yes, in fact, it is safe enough to launch and operate here. That's the part that really matters. After then, you just run around, you water all your little seedlings, and they'll sprout and germinate. eventually got a forest [ __ ] everywhere. But this is the whole point. The intelligent way for Tesla to be scaling is to seek regulatory approval strategically in cities generally speaking based on population. Higher population equals better starting point. Get regulatory approval which is not necessarily easy. There's different rules and regulations for every city. In many cases, it's not just one state broadly approves and you're off to the races. Many cities have their own rules. So get approval. That's a massive hurdle. Then do your testing and validation required for every new city. Then launch at a small scale, make sure everything's working, collect a little bit more data, and now you have a foothold in that city for future massive scale. I don't want to be too hard on this, an endless, but how do people not understand this? If you want to grow a forest of robo taxis throughout the United States and the world, you don't plant one seed in one city, water it really hard, and then plant more seeds in the same city and water it really hard, and think you're suddenly going to have an autonomous robo taxi fleet operating throughout the entire country. You plant seeds all over the place. They take a while to grow. You got to get them in the ground.
Musk also pointing out that Tesla currently having to accumulate miles in cyber cabs just to calibrate the cameras with a slightly different placement on the cyber cab. Lars also jumping in. So we had the whole team making the exact same [ __ ] point. By the way, Lars points out that regulations are different by cities. So they need to meet the varied regulatory requirements per city, hence launching in multiple cities so they can meet those requirements as opposed to just going hard in a single city.
And one final comment from Musk asked about Tesla's aggressive capital expenditures. He also just pointed out something that's worth contemplating. There is a balancing act, a decision to be made between being efficient with your capital expenditure versus moving fast. There are some cases where it's okay to be less capital efficient if it means you can move faster. So, expect Tesla to spend tens and tens and tens and tens and tens of billions of dollars in the coming years as efficiently as possible with the caveat that in some situations they may need to spend a little more lavishly to move fast.
Honestly, one of my favorite earnings calls. Great opening comments from Musk. Members of the team jumping in here. Quite a handful of them, actually. As I said earlier, though, the big standout, 55% FSD subscription take rate on new US Tesla sales in the second quarter. Wow.
Don't forget Patreon price hike when the new valuation model drops. You can join now locking current prices. Or don't, doesn't bother me, but it might bother you if you want to see the 20-year model after it drops and you have to pay a shitload more for the exact same thing. Just saying. Thanks for watching. If you appreciate the content and want to support the channel, check out the links below where you can get early access to daily videos, unlock hundreds of exclusive pieces of content, and plenty more. AG1 has been part of my daily health protocol for almost half a decade. It's packed full of vitamins and minerals, and helps me fill in nutritional gaps while supporting energy, digestion, and immune function. It also has prebiotics and probiotics to promote gut health. I take my health very seriously, and so should you. Try AJ1 today by visiting drinkag1.com/smr or click the link in the pin comment and enjoy a free welcome kit with vitamin D3 plus K2 and AG1 travel packs. That's drinkagg.com/smr.