Welcome to Electrified, it's your host Dylan Loomis, so in the title when I was saying that we were wrong I was referring to myself and Chris from Dirty Tesla. Yesterday we were talking about how actually SmartSummon hit a curb while Chris was testing it, but there's also this. It comes up and it almost hits this curb, coming straight out right away, but it stops prematurely. And throughout all of my tests you'll notice in other videos like Dirty Tesla has a video where he says his car hit a curb. But it clearly does not hit curbs and it's evidenced by zooming in on the camera footage. However, what is awfully frightening and it even fooled Dirty Tesla when he was inside of his car is that the car makes this grinding metal noise whenever it comes to a stop. And I did do a close up just to see if I did have any damage on my wheel and I did not. There you have it, SmartSummon may not actually be hitting the curbs, but doing quite a good job at simulating it. I'm sure many of you have seen the new Tesla AI roadmap by now, but if you haven't and for my audio listeners, for the remainder of September we can expect FSD version 12.5.2, which should have a 3 times improvement when it comes to miles between interventions. They do specifically say necessary interventions, but we can't really do much with that since they haven't released any MPI data at all. We're supposed to get 12.5.2 on AI3 with unified models for AI3 and AI4.
To date, the model for AI3 has been that quantized version of FSD but moving to unified this month. We've known actually SmartSummon is supposed to go wide next week, auto-park for the Cybertruck supposed to come this weekend, eye-tracking with sunglasses, we've known about that one. The end-to-end network on the highway supposed to come this month and FSD for Cybertruck also this month. In October we get some new updates we can expect on-park, park, and reverse in FSD along with FSD version 13 that's supposed to have a 6x improvement for miles between interventions. And then the real needle movers we have FSD expected in Europe and China in quarter one of 2025, but pending regulatory approval. It should go without saying these latter two updates are massive as that would be millions of customers now having access to FSD being able to use it, talk about it, share it with their circle, and likely be a lot more willing to pay for it since they can actually have the technology deployed. Based on everything I've been reading this year, it actually sounds like regulatory approval in China is a bit more likely or has an easier road ahead than it does across the EU.
There has been some progress this year for automated driving features across the EU, but one of the biggest things we'll need to see changed for this to be a smooth rollout for Tesla is one clear legislation or set of policies across Europe. That's because most progress right now is achieved on a country level. Responding to the roadmap, Elon added hopefully right hand drive markets in late quarter one early quarter two, again pending regulatory approval. It may not technically be a global release for FSD, but in the next 10 months for all intents and purposes, it could be. I feel like I need to remind everybody what Elon said in May, he said 12.4 was going to be a 5 to 10x improvement in that miles between intervention metric, but 12.4 never made it to a wide release and I know not everybody has seen a 5 to 10x improvement even with the release of 12.5. I've been saying I think it's time for Tesla to release this MPI data, especially if they're going to continually talk about these improvement numbers with every new release. Some people out there are saying that this is all enabled by Tesla's new compute cluster cortex, but the way I see it, by and large, that won't be truly up and running at max capacity until closer to the end of this year. That doesn't mean that cortex won't be contributing anything until the end of this year, but even so, I think that makes these releases even more impressive.
Overall, I love the transparency with this list, even though it may give some of the detractors, another window to argue that Tesla is missing some timelines, if that is indeed how this plays out, at least for some. And as I titled a video a few weeks back, in the year ahead, it's looking like things are going to be quite exciting at Tesla yet again after a period that we can classify as rather quiet, at least for Tesla's standards. Tesla does have close to 2 million local owners in China, so easy math if just 10% of them, which would be 200,000, actually pay around $8,000 for FSD, that would be $1.6 billion, which as we know is largely mostly margin for Tesla. Obviously, it's not likely that everyone would pay the full price up front as it's likely that Tesla will offer a subscription model as well.
George Powerhouse secured a battery and equipment supply agreement with Tesla for a 2000 megawatt hour or 2 gigawatt hour portfolio of projects across four different projects. They're seeking expedited approval of this portfolio to address some 2026 and 2027 winter resource shortfalls. All four projects will use Tesla's latest 2XL megapacks from Lathrop. Here's a table with these specific counties where these projects will be and they all have a commercial operation date between May and November 2026. The smallest project is 198 megawatt hours and the largest one will be 1.06 gigawatt hours. As always, we don't know specific contract details, but this could be another contract worth about $500 million for Tesla.
Morgan Stanley released another Tesla stock note saying Tesla remains our top pick in US autos. They said we forecast Tesla's gap operating profit to come in this year at $5.6 billion, but when backing out their forecast for EV credits, which are around $2.3 billion, and their estimate for Tesla energy operating profit around $1.1 billion, that drops the core auto operating profit number to $2.2 billion. Then they said they can actually account for most of that $2.2 billion from things like recurring revenue streams like connectivity, charging, software services, upgrades, FSD, etc. Implying remaining operations mostly the real core EV business to be lost making this year. However, operating profit is after operating expenses and the last line is very important. They said we note a large portion of Tesla's opx burden is non-auto, basically compute, which this algebraic exercise does not adjust for.
Now, the reason I think that was actually worth sharing is because much of Wall Street is in this camp and that they're waiting to see where Tesla's core EV business is going to bottom out in terms of margin and profit. That's why so many analysts are so laser focused on metrics like gross margin X credits. In fairness, there are still question marks like how much the fed cutting interest rates will one actually impact the auto loan rates that people can get and how much of an impact that will have on Tesla sales because they've already been offering lower interest rate incentives now for a majority of the year. Granted, a lot of that was in global markets, but the point still stands. And given that Elon has said the cost cutting for the current lineup is basically at the bottom, meaning there's not that many pennies left to be removed, we may be waiting for the next gen platform or more from autonomy for Tesla to actually boost its overall operating profit. And we know Tesla energy is looking great, but I'm just talking about from the auto side. So my point with all of this is that until that part of the business can show some signs of life again, I think many analysts in the big money on Wall Street is still going to remain quite skeptical.
And lastly, Jonas said Tesla has sent out invitations to the October 10th event, but as far as I know, that is not accurate. On LinkedIn, Tesla posted Tesla arrives in Lithuania. The Model Y and the Model 3 are now available, double checking the Tesla configurator for Lithuania. The vehicles are available. It's not just for test drives. As you may expect, the Lithuanian auto market is not huge. The past few years, it's fluctuated between 26,000 and about 35,000 cars per year.
Heinrich Zane shared an update on X saying that there are 26 brand new Tesla semis fresh off the prototype production line from Nevada. So far, Heinrich's really has been spot on a few months ago. He was saying that toward the end of this year, it's likely that Tesla will start ramping up production of the prototype line, and that is apparently exactly what they're doing. There's a total of 26 on this side of the building, 15 on the west side and another 11 on the east side next to the park next to the building. These are brand new Tesla semis.
Heinrich Zane 在 X 上分享了一条更新消息,说有 26 辆全新的特斯拉半挂卡车刚刚从内华达州的原型生产线上下线。几个月前,Heinrich 确实很准确地预测到了这一点。他当时说,今年年底前特斯拉很可能会开始加快原型生产线的生产,而现在看起来他们确实在这么做。这 26 辆卡车中,有 15 辆在建筑物的西侧,另外 11 辆在东侧靠近公园的地方。所有这些都是全新的特斯拉半挂卡车。
There's quite a bit of activity going on here. As far as I know, these semis are being delivered to PepsiCo until they get their 100 and also being used with in demonstrations for other customers. And that's usually short term, a couple weeks to a month for them to get a feel for it and figure out how it will work with a narrow logistics operation. And then also, many of them are continuing to be delivered to Tesla to the point where someday soon they may be using all Tesla semis for their deliveries of batteries, drive motors and powerwalls.
Many people are understandably up in arms about this time 100 AI list. It's their list of the 100 most influential people in AI for this year. Their first category is leaders where you'll find people like Satya, Altman, Jensen, Zuckerberg. For innovators, they have people like Lisa Sue, Brett Adcock from Figure. For thinkers, they have Ray Kurzweil, Ilya, Shoutout to Andrei Carpathi, and even MKBHD is somehow a big time AI thinker. For shapers, they have Teary Bretton, Scott Weiner, and Scarlett Johansson, clearly a critical AI figure. I think we can wrap it up there. You get where I was going.
Elon Musk is nowhere on this list. Of course, this list really means nothing, but the fact that they left Elon off the list is one of the more absurd things that I've seen this year. Elon did respond saying magazines don't establish relevance history does. But between Cortex and Colossus and Tesla's FSD and Optimus, and all of the public advocating that Elon's done for AI safety and the conversations given and his role in open AI in the early days, all you can do is shake your head and move on. And understand which publications don't deserve any relevance.
Sawyer said Tesla's referral program has launched in Germany and the UK. Some pictures were shared on Reddit of a test vehicle for Tesla with the LiDAR unit. But this time around, there was also a front bumper camera. And looking at the side profile, we have that camera pushed back from the regular placing on the side pillar, which means this is likely testing for the cybercab. 10-10 can't come soon enough. There was also a video of FSD being used in Italy. Now, I don't know exactly how this is happening.
It could be Tesla actually doing some testing themselves, or it could be somebody hacking the system. I said the latter because you may remember, a few years back, there was news that somebody hacked a Tesla in Ukraine and got FSD beta to work. Which scenario this is? I'm not sure, but I'll have the video linked below if you're interested in seeing the performance.
BMW is aiming to bring its first hydrogen powered vehicle to the market in 2028, using fuel cell tech developed in partnership with Toyota. The company said the vehicle would be an existing model with a hydrogen fuel cell option without giving any other detail. BMW said it assumed hydrogen charging infrastructure would have developed significantly by 2028 to accommodate its plans. And I think we all know what they say about assuming. Some more for Comedy Corner, apparently the Vice President of Li Auto just made a bold statement saying that their new end-to-end and vision language model solution has surpassed Tesla's FSD. He emphasized that their breakthrough, which they're calling one model, which they're also saying is really just an end-to-end model, aka they're trying to copy what Tesla is doing. But that breakthrough allows Li Auto to challenge and surpass Tesla. They just launched their trial program inviting 10,000 users to experience the new end-to-end system.
BMW 计划在2028年推出其首款氢能源汽车,该汽车将采用与丰田合作开发的燃料电池技术。公司表示,这辆车将是一款现有车型的氢燃料电池版本,但没有透露其他细节。BMW 预计,到2028年氢能充电基础设施将会有显著发展,以配合其计划。另外,一个有趣的消息,理想汽车的副总裁最近声称他们的新端到端和视觉语言模型解决方案已经超过了特斯拉的全自动驾驶(FSD)。他强调,他们所谓的“单一模型”的突破实际上就是一个端到端模型,即他们试图模仿特斯拉的做法。但这项突破使理想汽车能够挑战并超越特斯拉。他们刚刚推出了试验计划,邀请了1万名用户体验这一新的端到端系统。
Li Auto is spending about 2 billion yuan, which is only about $282 million this year on compute for its model. Li Auto spending hundreds of millions and Tesla spending multi-billions. Apparently not everyone in China is rightly acknowledging Tesla's lead with FSD. There are still some Gordon Johnson and Guide House solutions types in China too. Intel is considering options for its stake in struggling automated driving systems provider Mobileye. Intel could offload some of its 88% holding in Mobileye on the public market or via a sale to a third party. However, none of this is confirmed, they're just considering what they want to do. Mobileye's stock is down about 71% this year and at least in part it's because many automakers have been pulling back on their production plans, which means less sales for Mobileye.
理想汽车今年在计算投入上花费了大约20亿元人民币(约2.82亿美元)。与此相比,理想汽车只花费数亿,而特斯拉则投入了数十亿。显然,中国并不是所有人都认可特斯拉在全自动驾驶(FSD)方面的领先地位。中国也有一些像戈登·约翰逊和Guide House Solutions这样的人存在。
英特尔正在考虑其在陷入困境的自动驾驶系统供应商Mobileye中的持股选项。英特尔可能会在公开市场或者通过出售给第三方的方式,减持其在Mobileye中的88%股份。不过,这一切都还没有确定,他们只是还在考虑该怎么做。今年Mobileye的股价下跌了大约71%,部分原因是很多汽车制造商缩减了生产计划,这意味着Mobileye的销售额减少了。
Tesla stock, on the other hand, had another solid day, finishing at $230.17 up 4.9% while the Nasdaq was up 0.25%. It was a higher volume day for Tesla trading about 40 million shares above the average volume the past 30 days. It's not really Tesla related, but Trump did say at the suggestion of Elon, I'll create a government efficiency commission tasked with conducting a complete financial and performance audit of the entire federal government and making recommendations for drastic reforms. I think Friedberg would agree something like this is desperately needed for our country. Hope you guys have a wonderful day. Please like the video if you did. You can find me on the next links below and a huge thank you to all of my Patreon supporters.