Welcome to Electrified, it's your host Dylan Loomis. Just a heads up, I do have a sore throat today and I'm not feeling well, but there are some important things I wanted to clarify, so bear with me as I work through this one. The Department of Energy put this chart out showing that ice sales have indeed peaked as of 2017. That peak came in just over 80 million units per year and as of 2023, the last full year with data, that number was below 60 million. They did say the ice category includes traditional hybrid vehicles that cannot be plugged in.
One other trend to note, despite the overall recovery of these passenger vehicle sales since 2020, the overall portion of those sales made up of ice vehicles has been on a decline and it's really flatlined for the past three years. The official Tesla Optimus account shared this video of Optimus walking on some outside terrain in different scenarios. So despite this video somewhat looking like CGI, apparently this is real. Milan Kowats, an engineer at Tesla said these runs are on mulched ground where I've myself slipped before.
What's really crazy here is that for these, Optimus is actually blind, keeping its balance without video yet only using other onboard sensors consumed by a neural net running in around 2-3 milliseconds on its embedded computer. The team is working on adding vision so it can better plan ahead, making the gate look more natural on rough terrain, making it more responsive to velocity and direction commands, and learning how to fall to minimize damage when it's unavoidable, and the ability to stand back up. So to everyone saying that Optimus walks like Joe Biden, I would just say, why don't you go outside, close your eyes, and try to walk up and down some mulched hills. And be sure to have a friend or family member record it and send me that video. But seriously, the fact that Optimus is doing this without any video input and just using onboard sensors is quite impressive.
Giuliani Barz, another Tesla engineer said our goal is to make thousands of robots doing useful things before the end of 2025. And confirmed, the team is hiring across many different divisions for Optimus AI. Elon added Optimus can now walk on highly variable ground using neural nets to control its electric limbs. Then just FYI, a bunch of people online are saying this video is fake because there's no other movement in the video, but that's just not true if you look here, the flags are moving, the cars driving, these plants in the foreground are waving back and forth so that's just a lie.
And watch this video in slow motion where it slips and almost falls and I would add 9 out of 10 people in this scenario are definitely going down and or tearing their ACL or their quad. But Optimus slips, swings its arms back real quick, brings the knee back, finds some grounding, and then corrects itself, which is again just wild only sensors. Elon said I forgot to mention Optimus was doing this blindfolded we haven't yet made vision part of the control loop for uneven terrain. Optimus robots are regularly cruising around our workplace in Palo Alto where vision is used for static and dynamic object avoidance.
And Ashish another Tesla engineer said we're making Optimus walk increasingly more robust and ready for all terrains achieved by replacing a chunk of C++ code with nets. So Optimus is taking a page out of Tesla's FSD version 12 playbook removing C++ code. So the two things to keep in mind on that front, one how smooth FSD version 12 was compared to version 11, and two the rate of progress we saw after version 12 was released. Now obviously all of the C++ code for Optimus has not been removed, but it sounds like they're walking down that path.
And for what it's worth so far none of the engineers have said anything about teleoperation which in my opinion makes sense because to tele operate something like this I don't think is reasonable. Specifically when it comes to latency problems for balancing and the reflexes needed to actually catch itself. And a friendly reminder Elon posted this saying going for a walk with Optimus January 30th earlier this year. So as we've been saying all year the pace of progress for Optimus is pretty astounding.
In case you're wondering what Elon meant last night when he said Tesla shorts reciting litany against fear right now, that's actually a mantra that's in the science fiction novel Dune. It's basically a psychological tool to help people try to overcome their fears by reciting the mantra. Elon also said the Tesla cue went cue sharing a post from whole Mars. Whole Mars just explained that you add a cue to the end of a ticker symbol when that company goes bankrupt. So Elon saying Tesla cue went cue means those betting against Tesla went and are going bankrupt.
I'm sharing that though because there's plenty of reporting on this new model cue that Tesla plans to launch in the first half of next year. But this is all stemming from Deutsche Bank's report from its autonomous driving day on December 5th. Becky Peterson, a reporter for the Wall Street Journal said that Travis Axelrod, who works in IR at Tesla, met with Deutsche Bank analysts. But from there everything is up to interpretation because the report has not been shared and there's plenty of conflicting information.
And just know the model cue name that everybody was talking about that was actually a misinterpretation by a Chinese source. It turns out that's just the name that Deutsche Bank is using to refer to the upcoming more affordable Tesla vehicle. With rumors like this, just try to think logically, why would the head of Tesla's IR tell a Wall Street bank about Tesla's upcoming vehicle? The answer is he would not. He's not going to share any new information because it's not the proper outlet and they don't want to Osborne sales in Tesla's Q3 deck. They told us plans for new vehicles plural, including more affordable models, plural, remain on track for start of production in the first half of 2025.
They'll use some aspects of the next gen platform and some aspects of the current platform. But here are the details from the meeting, just know that model Q name is not official. But they're saying the new model will be priced below $30,000 after subsidies, which again we've already known. They're saying in the second half of next year, Tesla may launch other models, which checks out and they're saying it could be a three-row, long wheelbase model Y variant launched in China. I think the most solid thing we can lean on from this report is that Tesla is reiterating many of these targets, one of which being it's aiming for 20-30% sales growth next year.
他们会使用下一代平台的一些方面,以及当前平台的一些方面。不过,这里有会议的详细信息,但请注意,型号Q的名字还不是官方的。他们表示新型号在补贴后价格低于3万美元,这一点我们已经知道了。他们还说明年下半年,特斯拉可能会推出其他车型,这个说法是可信的,并且还提到可能会在中国推出一种三排座位、加长轴距的Model Y变体。我认为从这份报告中我们最可以依赖的信息是,特斯拉正在重申许多这些目标,其中之一是明年 aiming for 20-30%的销售增长。
But Tesla said achieving the high end of the growth range would require essentially flawless execution. The plans for Gigamexico will continue to depend on geopolitical dynamics and the tariff situation. Tesla said 2025 will be a year of product launches, and as expected, that would be a disruptor to profitability in the short term, thanks to fixed cost absorption being less efficient. Tesla still expects to launch unsupervised FSD in California and Texas next year using existing Model 3 and Y vehicles. The company is still planning to use an in-house developed ride-hailing app to control the value chain.
Tesla believes that regulation is the biggest obstacle to the widespread deployment of self-driving taxis, which is why moving regulations to the federal level would be such a big deal. And to start for unsupervised, Tesla is planning to use an entirely company-owned fleet, and eventually they'll adjust the supply dynamically based on customer demand and traffic patterns. The cybercab is still expected to use the unboxed method when it begins production in 2026, and any subsequent products released are expected also to use the box-less process. At full production, Tesla expects the cost of building a cybercab to be less than $30,000.
As the cybercab rolls out in 2026, companies will need to invest in service and cleaning and charging equipment, aka the wireless charging, with Texas and California obviously to be the first states to roll out these services. They confirmed FSD version 13 should go to a wide release about 2-3 weeks after the initial rollout if no issues are found. The report said Tesla's management does not see any real competition in the US or Europe from a cost and scale perspective, but in China, Tesla sees more entities taking a similar end-to-end vision-based approach.
The third-generation dojo chips are expected in 2028, as the first gen chips could not compete with Nvidia on cost and performance, and the second gen for 2026 still can't compete on performance, but should achieve cost parity. And just a quick aside from me, that's obviously a high bar and confirmation that Tesla is looking to compete with Nvidia both on cost and performance. And as Elon likes to say, the third iteration of any project is usually the sweet spot, so gen 3 dojo in 2028, keep that in the back of your mind. They said the intelligence of optimists will continue to improve and should at some point reflect the rapid improvements in FSD over the past year.
Tesla's strategy to sell to third-party customers is to sell the hardware and software together or to rent out the robot. And finally, megapack demand remains very strong, and the company is sticking to its guidance of more than 100% growth this year, implying around 27 gigawatt hours of production. And just know that number should be for only megapack production because Tesla's total energy storage deployed for this year is likely to be around 27 gigawatt hours. And I know that production and deployment are different, but it's a good enough proxy for this purpose. Just to show you through the first three quarters of this year, Tesla has already deployed 20.3 gigawatt hours of energy storage.
You have some other Chinese sources saying that Tesla is benchmarking the BYD Dolphin and the VW ID3 for this more affordable Model Q. They said the Model Q internally codenamed Redwood will be built on Tesla's latest vehicle platform, and compared to the Model 3, the new car will be 15% smaller and 30% lighter, and will use an LFP battery pack with 53 and 75 kilowatt hour capacities on both single and dual motor drive variants, with an estimated range of 500 kilometers or roughly 310 miles. This Chinese source went on to say the Model Q will be priced higher than BYD's Seagull and the ID3 starting at an estimated price of $19,200.
But even if this reporting was accurate, that would likely be for the Chinese market. But as part of its report, Deutsche Bank has named Tesla as one of its top picks for 2025, and said Tesla is at the forefront of the technological and scale curve in terms of reaching practical autonomy. Competitors are trailing behind in resources while Tesla is also ahead in eventual cost per mile and fleet size. With massive compute resources, the bank says Tesla can train several end-to-end models within weeks, a capability that places it ahead of competitors. Nothing we don't already know, just highlighting more of Wall Street waking up to what's going on, as Deutsche Bank makes Tesla stock a standout recommendation for 2025.
So to wrap this section up, just remember all of those reports are not really official from Tesla as far as we know. The reason being, we don't know exactly what Travis Axelrod did confirm, versus what Deutsche Bank is just speculating about based on their conversation. Either way though, this meeting has caused people like Gary Black to raise their Tesla price targets in Gary's case up to $380 from $300 before, based on increased conviction that Tesla's more affordable model will be launching in the first half of next year. And look, no shade to Gary here, but we've already already known this to be the case for a few months. It seems like Gary is just looking for a reason to increase his price target. But again, in fairness, if you want to use this report mainly as a reiteration of everything Tesla is working on, I do think that's fair.
And some of the new data like possible pack size and drivetrain, and starting cost in China, just treat those as rumors for now. As expected, Elon's pay package story has blown up. We now have over 470 different sources talking about the topic. And because this is likely a case for the history books, it's being covered by both sides of the political spectrum. We have 69 sources leaning left and 55 leaning right. But look at the difference in the headlines. One article on the right was titled Democrat appointed judge blocks Musk from receiving compensation package approved by shareholders. Whereas one on the left said Musk's huge Tesla payday is blocked again, but at least he's got Trump.
And what's pretty cool is for any news story, you can see these varying headlines and the political bias, thanks to the sponsor of this video, ground news. Ground news is a website and an app developed by a former NASA engineer on a mission to give readers an easy objective data driven way to stay informed on breaking news without any algorithms. And what's great for the holidays is now you can also give a subscription to ground news direct to anyone's inbox you'd like. And I like to think each subscription is a way to hold the media accountable. And as you know, it's a great way to support electrified right now. They're having their biggest sale of the year as they're offering electrified viewers 50% off their unlimited access vantage plan.
If you want to check it out, you can scan the QR code or use my link in the description to grab this discount for a limited time. And I know many of you are already enjoying ground news, ground news sources, all of the articles written on a topic and provides data about the political leaning, factuality, and ownership of the sources reporting, plus it's backed by three independent news monitoring organizations. And as I've been saying, the my news bias tool has been super useful for me to track my own news consumption data to ensure I'm not wasting time or being misled. So again, if you'd like to lock in that 50% discount, the QR code is on the screen and the link is below.
AI driver shared this video of FSD 13 taking the right away around this FedEx truck. You'll see this car flash the high beams. But I think FSD is really reacting to that vehicle stopping, maybe not so much the actual high beams. So I did want to make that small clarification, but either way, it handled the situation incredibly well. Holmars also shared this video of FSD version 13.2 getting out from around this UPS truck if you missed it, it's one you have to see. you Brian from Futuraza gave us some more details about the cybercap. The engineers are most excited about the efficiency of the vehicle and that 5.5 miles per kilowatt hour could be taken as a worst case scenario. And something that's incredibly exciting, he said the way it's designed is to be built entirely by robots. Now maybe that's not the case right out of the gate, but in the long term that seems to be the goal.
AI 驾驶员分享了这段视频,展示了FSD 13如何从FedEx卡车周围顺利通过。你会看到这辆车闪了大灯。但我认为FSD实际上是在对那辆车停下做出反应,而不完全是因为大灯。因此,我想做一个小说明,但无论如何,它都处理得非常出色。Holmars也分享了一个FSD 13.2版本绕过UPS卡车的视频,如果你错过了,那是一个值得一看的视频。Futuraza的Brian提供了有关cybercap的更多细节。工程师们对车辆的效率感到非常兴奋,5.5英里/千瓦时可以被认为是最差情况下的表现。他还提到一个非常令人激动的消息,那就是这个设计的目标是完全由机器人制造。虽然可能一开始达不到这个目标,但从长远来看,这似乎是终极目标。
We talked about the small cabin, which means you need less stereo equipment to get the same great sound. And that also means there's less room to actually heat and cool so the HVAC should be more efficient. Plus the cybercap uses the solid roof so less heat will escape and using a glass roof would be heavy and expensive. So the solid roof will mean the heat pump will work better too. On the rear hatch that opens up into the trunk, the underneath is formed plastic so the inner part is actually molded and formed straight to it. Which means it's very easy to make and then it's also cheap and easy to replace. There will only be five switches in the entire cybercap for the windows, the door opening and the rear hatch. The engineers did not want window switches at all, they wanted it through the screen but NHTSA would not allow it.
We know the vehicle has no side mirrors which reduces a number of parts in complexity and the short sidewall of the wheels does actually improve the rolling resistance and the front wheels are smaller than the back wheels because the arrow hit is bigger on the front end of the car. Which means you can have slightly bigger wheels on the back of the car and have less of an arrow impact because the back of the cybercap is tapered in. The model 3 had around 380 pieces when it comes to manufacturing, the model Y with structural pack was down to 180 and the cybercap will have 80. It sounds like to start the cybercap is going to be a single motor front wheel drive variant. This is why the front does not open, the hood space is already pretty small and what helps in accidents is having a crumple zone. So when you have less frontal surface area you need to use more of it to be safe in a collision.
Brian did confirm there will not be a human driven version, no wheel, no pedals, the cybercap is Robotaxi only, again this case should be closed as we've been saying for months. The engineers are hoping the cybercap is built with no rare earth elements in the motor and the cybercap will only need to handle speeds up to about 90 miles per hour as the highest speed limit at least in the US is 85. The back wall of the interior is indeed structural so that's not coming out. The cybercap will be battery agnostic so Tesla can put in any type of battery cell they want. The pack will only be semi-structural and it will be built around the world eventually. It should be around 1000 pounds lighter than a model 3 but he did get some conflicting answers on the specific weight reduction. The interior seats will be adjustable and when it comes to cleaning there are no door pockets, there's no induction phone charging because people may forget it, there is no glove box, there's only three pieces of glass, there's no gap in the seats where stuff can easily fall, so the cleaning of this vehicle should be about as easy as it gets. And the word is you will be able to bring your own entertainment into the cybercap but it's not clear how they'll do that. Again shout out to Brian for driving all of that way and getting us this information I hope you guys go over and show him some love, his youtube channel and X will be linked below. Elon did say there will ultimately be tens of millions of driver list Teslas throughout the world giving rides 24-7 and at this point I can confidently say it's going to happen way sooner than most people are expecting.
Joe Tettmeyer shared another video of Gigatexis showing the cybercap doing some more testing with a cyber truck with LiDAR following it around. Talking about the cybercap, Elon also said if the product is special enough no logo is needed. So obviously the best logo is no logo. A lot of people are missing this one, the CEO of CATL told Reuters in an interview he expects the business of developing and managing zero carbon electric grids aka battery storage to be 10 times larger than supplying EV batteries. So it's not just Elon that thinks the market for battery storage is massive. Robin Zhang also said CATL is aiming to build independent energy systems big enough to power a massive data center or even a city. We've talked about it many times this year but here's another anecdote on the point of the intersection of AI compute energy independence and battery storage and all three of those industries will be intersecting heavily with data centers. I'm guessing most of us did not have this on our bingo card for 2025 but the Wall Street Journal put out an opinion piece saying that Elon was actually underpaid when it comes to his 2018 pay package. Jeff Yas wrote, Judge McCormick based her ruling on a vague fairness standard that would effectively let one investor challenge a corporate transaction approved by a majority of shareholders. If so, Elon should countersue because at the time of the deal for his compensation plan in 2018, most option traders thought the pay package was worth around $3 billion. And as we know, the value of the pay package has since ballooned to well over $50 billion, reason being, Elon and the Tesla team have created that much value for shareholders. Jeff said we believe the board took advantage of Elon's overly optimistic enthusiasm for Tesla's future prospects and gave him less than he deserved. Since the deal in 2018, Tesla stock has appreciated 1700% a real long shot.
Joe Tettmeyer 分享了一段关于Gigatexis的视频,展示了带有激光雷达的电动卡车进行测试的情景。谈到这款"cybercap"产品,埃隆·马斯克表示,如果产品足够特别,就不需要标志,最好的标志就是没有标志。许多人对此未曾注意到,中国宁德时代的首席执行官在接受路透社采访时说,开发和管理零碳电网(即电池储能)业务的规模预计将是供应电动车电池的十倍。所以,不仅仅是埃隆认为电池储能市场巨大。Robin Zhang 还表示,宁德时代计划建设足够为大型数据中心甚至一个城市供电的独立能源系统。我们今年已经多次讨论过这个问题,人工智能计算、能源独立和电池储能三大行业将与数据中心紧密结合。可能大多数人没想到这一变化会在2025年出现在我们的预想中,但《华尔街日报》发表了一篇评论文章,说埃隆在2018年的薪酬方案中实际上被低估了。Jeff Yas 写道,法官 McCormick 基于一个含糊的公平性标准做出判决,这实际上允许单个投资者质疑多数股东通过的企业交易。如果是这样,埃隆应该反诉,因为在2018年酬金计划交易时,大多数期权交易者认为该薪酬方案价值约30亿美元。正如我们所知,该薪酬方案的价值后来暴涨至超过500亿美元,原因是埃隆和特斯拉团队为股东创造了巨大价值。Jeff 认为董事会利用了埃隆对特斯拉未来前景过于乐观的热情,并给予了他低于应得的薪水。自2018年交易以来,特斯拉的股票已经上涨了1700%,真是一次遥不可及的成功。
There was a new study published in Nature Energy today titled Dynamic Cycling Enhances Battery Lifetime. New research suggests these batteries once thought to have short lived inherently expendable shelf lives may actually last significantly longer than expected. In some cases, properly cared for EVs may even outlive their fossil fuel counterparts. EV batteries may actually last about a third longer than previous forecasts. The shocking disparity in battery life estimates stems from fundamentally unrealistic testing environments that became industry standard.
Basically, the industry has not been testing EV batteries the right way. They said to our surprise, real driving with frequent acceleration, breaking that charges the batteries a bit, stopping to pop into a store and letting the batteries rest for hours at a time, helps batteries last longer than we had thought based on industry standard lab tests. Which makes sense, in the lab effectively they were just taking batteries from 100 to 0, charging them up from 0 to 100 and repeating that cycle, but that's not how it plays out in the real world. The researchers tested 92 commercial lithium-ion EV batteries over two years across four different types of driving profiles.
So, doing some testing using these dynamic charging habits like we see in the real world resulted in an increased battery lifetime by up to 38%. And get this, though many drivers believe rapidly accelerating and breaking degrades EV batteries faster than steady driving, the researchers found a correlation in their data suggesting sharp short accelerations may actually lead to slower battery degradation. I'd love to dive further into this study, but I'll be honest, my throat is reaching its max capacity for the day, but I will have the study linked below.
Tesla stock closed the day at $389.79 up 0.15% while the Nasdaq was down 0.62%. The volume was about 2% higher than the average the past 30 days. You may have seen Rivian was up over 11%, at least part of that was because benchmark securities initiated coverage with an $18 price target saying Rivian has a massive market opportunity over the next few years. And Lucid was up over 10% on the day, likely because the gravity has entered production.
Don't forget, check out ground news links below, maybe gift it to a family member or a friend if you're interested, they are running their biggest sale of the year. Thank you all for bearing with me today in my low energy, hopefully I will feel better as the week goes on. Hope you have a wonderful day, please like the video, if you did, you can find me on X-linked below and a huge thank you to all of my Patreon supporters.