I Ran the Numbers on Terafab: Uh Oh / Tesla's New $3B Order / Tesla Semi Details Revealed ⚡️

Kind: captions Language: en Welcome to Electrified. It's your host, Dylan Loomis. Quick shout-out to my newest patron, Peter KS. Thank you very much for choosing to support the channel. Well, well, well, what a weekend it was. Now, listen, if you're expecting a deep dive video on everything that the Terafab will mean to Tesla and SpaceX for the next 10 years, this is not that video. Primarily because it would be impossible for me to cover everything I want to in one episode. So, for today, I think our time will be best spent going over a few key data points that will help to frame this entire situation and really set the expectations going forward. I know, right now, the Tesla community is as torn as ever, and there are plenty of people that share the sentiment strongly that Ramy just shared. He said, "I'm sick and tired of hearing a million robotaxis, a billion humanoid bots, age of abundance, order of magnitude." And yet, we look at Tesla's financials and what they're actually shipping, and we're not there yet. And it's a fact, Tesla's bottom line has been stagnant, if not regressing, now for 3 plus years. So, as inspiring, motivational, and exciting as the Terafab event was, I think for today, we need to go over some things and make sure that people do not get ahead of their skis. This photo right here is just one EUV lithography machine from ASML. Just this month, Dylan Patel said that a great rule of thumb right now to produce 1 GW of chip output would require about three of these EUV machines. As I've said before, last year, ASML, the only company right now making these commercially at any type of scale, made about 50 for the entire year, and the expectation is that they'll make somewhere in the neighborhood of 60 to 70 for this year. And at the event, Elon said the Terafab plans to do 1 TW of compute every year, which would be 1,000 gigawatts. Meaning SpaceX and Tesla would need somewhere in the neighborhood of 3,000 of these EUV machines to reach that scale. Bear in mind that three EUV machines to reach 1 gigawatt of annual chip output is just a rule of thumb. It is not a hard and fast rule. It will depend on the chip and the specifics and many other factors. However, for current leading-edge chips, that is the benchmark. Another fact for you, Dylan Patel said that by 2030, globally, the expected fleet of these EUV machines deployed is expected to be around 700. And that number is cumulative. AKA, in 2030, adding up all of the EUV machines on Earth, the number is expected to be around 700. And for SpaceX and Tesla to hit the numbers mentioned using that benchmark, they would need around 3,000 themselves. And that's only one data point of dozens, if not hundreds, that I could share to get across the magnitude and, frankly, absurdity of what Elon's SpaceX and Tesla are trying to do here. Now, before I carry on, I do want to make it clear that I am incredibly excited about this opportunity. I have been fascinated by this space, and as someone who plans to invest in SpaceX and Tesla for a long time, these types of paradigm-altering decisions are exactly what I, as an investor, am looking for. I also happen to believe that because of physics, this is going to take longer than most people are expecting. There are good reasons why, to date, most of these companies have been separating each of these steps into different buildings. Those being the logic, memory, and the packaging. Isolating these three environments keeps any contaminants isolated to one part of the process. So, if a company like TSMC has a dust problem, we'll call it, it would only impact one part of the process because it's isolated. But, if terrafab has a circulation or really any type of issue, then all parts of the process are going to be impacted. And believe it or not, these three different individual steps are actually very different from one another. And bringing those three under one roof brings with it the risk of cross-contamination. In this process, it really goes down to an atomic level, and there are all kinds of vapors, and chemicals, and particles, and residues that could atomically disperse throughout the factory impacting any part of the process. And that's why Samsung, one of the only big players in the space to do both logic and memory, still does them in separate fabs with separate teams. Now, Elon did say terrafab will technically be two fabs, each making only one chip design. This greatly simplifies process flow and allows more linear adjacent movement of the FOUP, which is a front opening universal pod. Just think of it like a plastic container that would house these silicon wafers as it moves from one step to another. So, in fairness to SpaceX and Tesla, they're touting doing things very differently than they've been done so far. Elon even went as far as to say that they would be pushing the limits of physics. Really looking to get into some new physical designs that have never been done before. Elon said having new iterations of a chip design be produced every day in the research fab with under a 7-day lag means being able to try out many high-risk, high-return ideas. So, to be clear, the actual location for this massive terrafab is TBD. They have not decided where it will be, but it will not be at Giga Texas. There's simply not enough room for it. However, at Giga Texas, they are planning to build what they're calling the Advanced Technology Fab, which can just be thought of as an R&D lab. And Elon did say that building will have all of the equipment necessary to make a chip of any kind, logic or memory, and they'll be able to make those lithography masks or the reticles that we just talked about in a single building. Which Elon is saying will allow them to do that recursive learning faster than anybody on planet Earth because no one else is doing all of this under one roof. Which, of course, sounds awesome on paper, but to be clear, this has never been done before. This type of quick iteration of different chip designs. So, if the teams can figure this out sometime in the 2030s, it will pay off very handsomely for everybody involved. I just want to make the point that success here really is not guaranteed. And just one simple example of this, for the past three to four years, Samsung has been notoriously struggling with its yields on its three nanometer tech. They've largely been stuck between 10 to 40% yields, whereas TSMC on its three nanometer tech is up around 70 to 90% yields. So, Samsung, one of the leading players in the world in this industry with experience and talent, is still struggling with some of these most advanced chips. And remember, they're not trying to do everything under one roof. And for me, I'm really not expecting any Tesla stock appreciation based on this announcement for a long time. Not only because of the difficulty and uncertainty, but even if successful, it's not like Tesla's going to have a product it'll be selling into the market here. All of the output, as far as we know, will be used in-house by SpaceX and Tesla. So, it's not technically a revenue driver as much as it is a margin saver, meaning Tesla won't have to pay those margins to other companies for their chips. And just my own expectation here, I think best-case scenario, we're looking at 2030 before Tesla has even one chip that it can use from this fab. Elon said 100 million square feet is the right order of magnitude to think about the size and scale of the Terafab. For comparison, Giga Texas is 10 million square feet. And Nick shared this infographic to show the scale of that comparison. And bear in mind, it sounds like it'll be two separate buildings because Elon said there will be two different chips, one for inference in the vehicles and Optimus, and then one optimized for space. And then that third building will be in Giga Austin, which is the R&D lab. But, my point here is it took 1 year for Tesla to build Giga Texas. And I can assure you that the techniques, the equipment, and the precision required to build a Terafab are going to be much more difficult than auto manufacturing lines. So, let's say they pick a location and actually break ground in Q3 of this year. Best-case scenario, 2 years to actually get the foundations of the building. That takes us to Q3 2028, and then 2 years at least of testing and iteration and working on the yields. And that right there has us already in late 2030. So, realistically, baking in some delays and challenges and talent acquisition problems, it really would not surprise me one bit to see that all pushed back until closer to 2035. On that specialized talent, it's important to note that roughly 20% of all of the jobs at a foundry like this are actually very specialized roles. So, 80% you can hire pretty much whoever you need to and get them trained up in a reasonable amount of time. But, that remaining 20% there are really only a few thousand people on planet Earth that are qualified for these roles right now. And both Samsung and TSMC in the States have both delayed their fabs by over a year due to this lack of specialized labor. And these companies were forced to fly in workers from Taiwan and Korea. Elon said all chip output on Earth right now is around 20 gigawatts per year and the expected demand from SpaceX and Tesla would be at 1 terawatt, meaning global output right now is only about 2% of what SpaceX and Tesla are expecting to need. But, pair that with somewhere in the neighborhood of a 100 million square-foot factory for the Terafab, and of course that location will require thousands of these specialized workers. Now listen, if any company can attract the top talent required and find a way to get them qualified quickly, I'm very confident it's SpaceX and Tesla. But, all I'm saying, at a minimum, this should be a part of the conversation. You know those times when life feels like it's pushing you a certain direction? Well, last fall I had my wife sprinting outside panicked and ultimately she was yelling at me to get off the grass as I was in my bare feet. Reason being, I completely forgot we had just had our lawn treated. Then, just a few days later, I happened to watch an Andy Sly video and I saw him talking about Sunday, which is a custom lawn service that uses no harsh unwanted chemicals and uses ingredients you can trust that are planet friendly, in line with Tesla's original mission, and designed with your family in mind. And fast forward to today, and I'm happy to announce that Sunday is now the sponsor of this video. So, finally a new one for you all to hopefully try. And I think Ashley would be upset if I didn't make it clear that Sunday has committed to never using glyphosate aka Roundup in any of its products. The process is fast and easy. You just go to their website, type in your address, and then Sunday uses satellite imagery of your actual yard plus local soil and climate data to build a custom lawn plan specific to your property. They ship everything to your door for free timed to when your lawn actually needs it based on your growing season. >> Next step, soil samples. They give you the tool. We'll get it from a couple different spots, put it all together, ship it out. >> Your first box will come with a soil test kit that you send back to their lab and they use those results to dial in your next shipments even further. So, every product you put down is backed by real data on what your soil is actually missing. You'll get these custom formulated nutrient pouches that connect right to your hose, so there's no mixing or measuring. They use ingredients like seaweed, iron, soy protein, molasses. Again, no harsh chemicals which, for those of you with kids or pets using the yard all summer, it's a big deal. >> I'm actually pretty eager to see the before and after here over the next few weeks cuz I'm not going to lie, right now my lawn's looking pretty rough. >> And you never have to worry about timing. The shipments show up when it's time to apply. You just hook up the sprayer, walk your yard, and you're done in about 15 minutes. So, if you'd like to try it out and support the work I do here, you can head to get sunday.com/ electrified linked below or use code electrified 25 at checkout to get 25% off your first lawn care plan. Again, just trying to set realistic expectations here so people do not end up disappointed. At Tesla's Battery Day in 2020, they said the plan was 100 gigawatt-hours of cells by 2022 and that would be supplemental to what they're getting from suppliers. Elon also said they would be at 3 terawatt-hours by 2030. Well, the last update we got in 2024, the run rate was in the neighborhood of 17 gigawatt hours per year of 4680 production. Let's be generous, round that up to 20, that's still only 20% of the stated goal and that was 2 years after the deadline. So, given I believe chip fabrication will be orders of magnitude harder than making battery cells, I do think it's wise to put some sort of discount on the numbers and timelines Elon's talking about here. So, I believe this 1 terawatt number really shouldn't even be part of the conversation or the expectations until well into the 2030s. And there are various reasons for that, not the least of which is the ASML EUV machine example we covered earlier. And now yes, on that point there are companies out there working on breakthroughs in this arena. Steve Jurvetson, who was a former board member at Tesla, has now been talking about LACE lithography, touting a novel approach to chip making that can extend Moore's law 10 times beyond what's possible with light to atomic resolution. He highlighted the dominance of ASML and said LACE has developed a new approach, instead of light, LACE's engineers have made a form of lithography that uses a helium atom beam. With that, this Norwegian company will be able to create chip designs that are 10 times as small as what is currently possible. But once again, that is future technology at least from a commercial viability standpoint. And on this point, the serendipity of this one would be pretty wild if true. Omar said he lives in the town of ASML. Some insiders are saying that ASML promoted a lot of strong EUV people into architect roles and management tracks. Then McKinsey came in and basically told them to cut the management roles. So now some of those EUV experts who got promoted are the ones on the block. About 3,400 roles targeted, mostly management. Half will be reassigned, but the rest will be gone, aka free agents. Joshua shared the news tagging Elon saying please hire them as McKinsey was nuking ASML. Elon replied asking is this for real? Then Jordan from The Limiting Factor said I bet Elon gets into lithography machine manufacturing before long. It's a big limiting factor for a type two civilization. Sharing that news we just went over and Elon replied to that saying true. Maybe Elon said true agreeing that it's a limiting factor or maybe he said true that they'll also get into EUV machine manufacturing. So, it's fair to ask if there's a potential 1,500 uh specialized workers being laid off right now from ASML, the one company on planet Earth making these EUV machines at scale, could that force Elon's hand a bit faster to try to get involved in EUV manufacturing. Not that it's something that would happen quickly, but just starting that process and making that decision. And to everyone saying that this is just a pump for the SpaceX IPO, I would say okay. Any CEO worth their salt is going to do anything in their power to maximize the money raised in an IPO because, you know, it's the main reason you IPO in the first place. But we've already learned that Wall Street will make you prove it before they give it a value. It's the case with FSD, with Optimus, and it'll certainly be the case with the Terafab. And look, I know this segment will be perceived by some that I'm bearish or negative on this venture. That's certainly not the case at all. Again, my primary mission for today was to set some proper expectations, so we don't end up with even more people disillusioned with Tesla in the future. But this really is a topic that we will be uncovering and digging into for months and years to come because it does go that deep. And I haven't yet said a word about space or Starship and how all of that is now going to factor in. And I have to say, after this weekend, an eventual SpaceX acquisition of Tesla finally has become much more likely than not in my opinion. It likely wouldn't happen for at least another year, but this is all going to cost a lot of money. We're talking upwards of $30 so it really will be an all hands on deck type of venture for SpaceX and Tesla. But given the fact that the tie-ups, synergies, and integrations between these two companies seemingly grows with every passing month, I would start to begin wrapping your head around that future outcome. And today I'm not even going to provide a summary type take on this Terafab event because I didn't even scratch the surface of getting into it. As I said, this will be something that we continually dive into in the months and years ahead. Jay Leno put out a new video with the updated Tesla Semi and some words from both Franz and Dan Priestley. There are a few hundred Tesla Semis on the road now that have driven a cumulative 13 and 1/2 million miles and they do have one lead truck that's sitting at 440,000 miles. They did hint that Tesla is working on wireless charging eventually for the Semi and the batteries are designed to last 1 million miles. For the standard range variant of the truck that goes 325 miles, it can pull the same load as the long range truck. And really one of the only differences between the two trucks is that the long range has three batteries in parallel. Man, they're referring to packs of course, whereas on the standard range they remove one of those battery packs and then shrink the wheelbase accordingly. They decided to remove the tilting side windows because it was a little awkward getting clipboards and things through that window, and they instead moved to a more traditional drop glass, kind of like what you see on cars. Windows that just drop straight down. One of the cool things about the Semi, they're using a lot of the same parts from the rest of the Tesla lineup. And one example of that, the Semi is using a lot of the Cybertruck's power share equipment to do what they're calling an electric power take off or an EPTO, which is just like an outlet that can supply around 25 kilowatts of power to whatever is attached. And one example they gave, usually semi trailers have a diesel engine to run the cooling system in the trailer itself. It's like a mini generator. But Tesla will be able to power its trailers and do that refrigeration directly from the truck's battery by reusing those same electronics from the Cybertruck. They said the Semi was 50% cheaper to run on a per mile basis in California compared to diesel, and it's 20% cheaper across the country, which includes things like maintenance and everything else. Priestley said the customers that have demoed the Semi are saying the worst thing about it is having to give the truck back after the demos are done. The Semi does have air suspension, and it has air brakes, which are just standard air friction brakes. And they're using a brake by wire system that is actually moving some valves, but it's electronically controlled. And Dan said that all trailers still have air brakes, so Tesla needed to be able to interface with those. And they said the turning radius in the standard range Semi is roughly the same as it is in the Model Y and the Model 3. >> Yeah, there's nothing special or unique that's for this truck. So we can use off-the-shelf tires. We have a recommended set that help you extend your range because the rolling resistance is a large piece of it, but you can put any set of tires on that work for your operation. >> Does this weigh more than the equivalent of gasoline truck? >> Yeah, so since the last time we were here, we cut about 1,000 lb out of the truck. And this long-range truck, the 500-mile version, is about on par and in part because it's got a 2,000-lb weight exemption that electric vehicles have. So, payload parity is very possible for most applications. What we hear from customers is they usually like to haul about a 45,000-lb payload, and we can definitely do that with this truck. It's outdoing it every day. No, cuz these are air brakes. So, >> Oh, yeah. >> it's on them is really low, and the service team has been exceptional with our the fleet time is usually up 95% uptime. Almost 75-80% of breakdowns, if you have one, are returned back to the customer in less than 24 hours. And it's actually almost half are back to a customer within an hour. So, we're right there with rapid response. >> And believe it or not, the Wall Street Journal recently put out an article highlighting how everyone that's been demoing the Tesla Semi has effectively been loving it. One person said, "It's just easier on your body. It's less stressful because you're not really having to engage the clutch and the stick shift." Fleet operators are getting excited about the reduced operating costs, specifically maintenance costs. An executive was excited about the potential of fewer service personnel being needed. And while we do not have official pricing, the Wall Street Journal said reports are suggesting it could come in roughly $100,000 lower than comparable battery electric trucks. Another CEO said the Tesla changes everything. It opens up a whole different type of delivery that I can make. And this really just continues the theme of seemingly nothing but good news coming out about the Tesla Semi. Joe Tegtmeyer shared these images from Giga Texas of what appears to be a new vehicle body structure that was just shipped to Giga Texas. At first glance, it does indeed look like a Model YL, and it would make sense if it was shipped in from China. And if you've been following along, you may have seen that video where I heard the rumor was that the Model YL was indeed coming to the United States. Now, I will acknowledge that if you look at the rear part of this body frame, it does seem to have a higher slope in the back than the Model YL. So, is there a world where this could be a new vehicle form factor, like a Model YXL, something bigger for the United States? Listen, I think anything is possible, but given everything Tesla already has going on, I would not think that's the most likely outcome. And given the rumors and the sales figures, the Model YL coming to the US does make a lot of sense. And we've said for years that the images from these drone videos can be very deceiving when it comes to perspective. This one largely went overlooked, but there's reporting that Tesla's looking to buy equipment worth about 2.9 billion dollars in order to manufacture solar panels and cells from Chinese suppliers. The company, Suzhou Maxwell Technologies, the world's biggest producer of screen printing equipment used to make solar cells. The word is this company has been seeking export approval from China's Commerce Ministry, but seemingly there are some other potential suppliers from China in the mix. It's not immediately clear how much of this equipment would require approval or how long it would take. But the Chinese companies were told to deliver the equipment before this fall, with two saying it would be shipped to Texas. They said Elon plans to build the solar capacity mainly for use by Tesla, although some will be used to power SpaceX satellites. And last month, Tesla visited several solar companies in China. And as we know, the US solar market is heavily protected by tariffs aimed at curbing imports of cheaper panels and cells from China and Southeast Asia. However, solar manufacturing equipment was excluded from tariffs by the Biden admin in 2024 at the urging of US solar panel makers who argued they had nowhere else to buy the machines needed to set up domestic factories. That exemption has been extended by the Trump administration and the US has been pushing to create its own solar supply chain to reduce its dependence on Chinese companies. So, two things. The screen printing part of the solar cell manufacturing process is the part that actually makes the bits that collect and carry the electricity. And second, I thought it was interesting they said this equipment would be shipped to Texas and not Buffalo. We know to date Tesla's been doing a lot of its solar manufacturing in Buffalo, but could they be looking to expand in Texas instead? And as Joe Tegtmeyer shared today, there's already plenty going on at Giga Texas as he shared some videos of what we believe to be the site for the future Optimus production to the tune of 10 million units per year. And again, just using industry benchmarks, that $3 billion purchase could unlock somewhere in the neighborhood of 40 gigawatts of cell capacity, which would be about 40% of the eventual goal that Tesla's trying to get to of 100 gigawatts. But, it sounds like that equipment may be delivered as soon as Q3 of this year and the groundwork both literally and figuratively that Tesla is laying this year truly is pretty special. So, it might feel like a slow year of more transition, but they are clearly moving and spending very aggressively. I really don't like relying on the RoboTaxi tracker tool, the third-party option, but it's all we have and I don't want to gatekeep. So, right now they're saying the total fleet is 525 vehicles, eight of which are unsupervised, 431 in the Bay Area, and 94 in Austin. But, as I always say, apparently people just completely disregard that most of those are actually inactive. So, right now in Austin, we're still at 42 active vehicles. And in the Bay Area, it's 431 active. So, there might be some actual expansion happening up from one unsupervised vehicle to eight in Austin. And as Benjamin just shared, the Cybercab was just spotted yesterday testing in Oklahoma City. When it comes to supervised FSD approval from the RDW, Tesla did say the new expectation is April 10th. And the RDW replied saying that Tesla is indeed in the final phase of the assessment process. The joint intensive program began about 18 months ago. So, yes, we're coming up on 2 years to approve supervised FSD. And all they said, after the rest of testing, a decision will be made. But again, the new date to mark on your calendar, April 10th. Homar said it's the most important 3 months in Tesla's history, and the market doesn't get it yet. Elon replied saying they really don't, and I've given them all of the information. And Elon shared this from Jesse saying, "When an item's on sale, everybody loves it. But when a stock is on sale, everybody loses their minds." Implying that he believes Tesla stock is on sale. Today, Tesla stock closed the day at $380.85, up 3.5% while the NDX was up 1.22%. The volume was 25% above the average. And don't forget, if you're interested in trying out a custom lawn plan from Sunday, I'll have all of the links below. And just in case you're new, trying out these sponsors, even for a short time, really is a massive help to the channel. So, thank you in advance. Hope you all have a wonderful day, and a huge thank you to all of my Patreon supporters. >> Mhm.