Welcome to Electrified, it's your host, Dilunumus. On yesterday's video, when it comes to the Canada plan to delay those tariffs, as some of you point out, turns out most of that deal was already in place as of late last year under Biden. So it does appear like the threat of tariffs on Canada has accelerated that deal, if anything, but again, most of it was already in place, so apologies for that one. ARK Invest put out its big ideas report for 2025, and as you would expect, Tesla is mentioned quite frequently. And look, I'm not advocating for ARK's investments in the short term, but personally, I find value in its research over the long term when it comes to tech disruption. Not that their research is bulletproof, but they've clearly been one of the few to understand the Tesla story for a long time. For this chart, they have Waymo's intervention rate around that 17,000 miles per intervention number, and ARK is expecting Tesla's FSD 13.4, maybe released in quarter one, to exceed that rate. And the human driver accident rate is closer to 1 million miles between interventions, and they're expecting FSD 13.6 in the second half of this year to be on par with that level. That would be made possible by a near doubling of compute at Cortex in Austin. I'm not sure where ARK is getting its compute figures from, though.
The last update we got from Tesla, they were supposed to be at about 85,000 H100 equivalent GPUs at the end of 2024. This from the quarter three slide deck. Then in the quarter four deck, Tesla said they increased AI training compute by over 400% in 2024. So nothing official has been reported, but given Tesla's plans to increase its CapEx on AI compute, I think a doubling for 2025 for H100 equivalent GPUs is definitely reasonable and on the table. This was the pretty self-explanatory chart. It was ARK's 2014 estimates of when different brands would release their autonomous vehicles. Fast forward a decade and only Tesla and Google's Waymo are still in the mix. And listen, props to ARK for being transparent like this because they're reminding everybody that in 2014 they thought Tesla would launch its RoboTaxi fleet sometime before 2018.
When it comes to this page showing some of the companies working on RoboTaxi's and how the US and China are really leading the way, in case you're wondering, wait a second, why wouldn't Huawei's ADS be on the list for China? It's because so far, Huawei and Avatar have not announced any plans to make their vehicles RoboTaxi's in the future, they're really strictly focused on driver assistance right now. When it comes to an annualized commercial ride run rate, Waymo right now is leading the way at about 9 million rides per year, followed by Baidu's Apollo go at about 3 million rides per year. But when it comes to autonomous miles and an annualized run rate, Tesla is at about 3.5 billion miles per year and then there's everybody else. ARK said by 2030, their research suggests the number of RoboTaxi's will have scaled to about 50 million globally with Tesla at a 50% market share. For what it's worth personally, I think that 50 million number is a bit high by 2030. If it was me, I'd be expecting between 10 and 20 million, but of course I'd be happy to be wrong on that one.
Based on Tesla's expected efficiency of the CyberCab of at least 5.5 miles per kilowatt hour, paired with battery cost declines that could mean a battery could cost as little as $2,300 for a RoboTaxi. Given the drivetrain of the vehicle is about 20% of the vehicle cost, they say a CyberCab could be profitable at a price of $15,000 for below. This would be the chart to screenshot because we're always arguing that Tesla's FSD suite is going to be much more cost effective than any of the other competitors, especially the ones relying on LiDAR. Because of that, ARK thinks Tesla's cost per mile will be 30-40% lower than Waymos. I would just point out that is at scale for both Waymo and the CyberCab.
Something people forget about when it comes to autonomy, these deadhead miles, ARK calls them the number of miles driven without passengers which account for 45% of ride-hail miles today. Autonomous vehicles could reduce deadhead miles by removing the human in the loop and harnessing AI tools to predict demand and traffic patterns. With a human ride-hail cost per mile of about $2 today, the RoboTaxi has an opportunity at scale to take that cost per mile down to $0.25. The chart for Global Stationary Battery Storage and the purple is Tesla's share by year. A pretty impressive stat, Tesla accounts for about 19% of global energy storage today. As we know, with a Shanghai factory getting underway and a third Megapack factory planned, energy is set to continually be a bigger part of the Tesla story.
On the humanoid front, ARK did some basic calculations and found the net present value of a humanoid robot to be about $550,000 based on a 10-year working period, aka if a humanoid robot can replace 100% of the work of your average US worker. Over a 10-year period. I'm always hesitant with calculations like this because there are still so many unknowns for one the longevity of these humanoids, the battery packs, replacement cycles, but what ARK is saying if a humanoid can replace 100% of that worker's work over 10 years, someone should be willing to pay in theory about $550,000.
And as we know long-term, the expectation for the retail cost of Optimus is expected to be well below $100,000. There's a lot in the report, but that's most of the Tesla-specific data. I'll have the full thing linked below. CNBC shared a video from Consumer Reports about the driver assistance systems and where all the top automakers currently rank. Because you're all dying to know, here are the results. Ford's Blue Cruise is ranked number one, followed by GM's Super Cruise and then Mercedes is in third with Tesla's autopilot and FSD combo in eighth. There were two main reasons that Tesla got dinged.
翻译如下:
“众所周知,从长期来看,Optimus的零售成本预计会远低于10万美元。报告中有很多内容,但这部分是特斯拉的数据重点。我会把完整的内容链接在下面。CNBC分享了一段来自消费者报告的视频,关于各大顶级汽车制造商目前的驾驶辅助系统排名。想知道结果的人有福了,这里是排名情况。福特的 Blue Cruise 排名第一,其次是通用汽车的 Super Cruise,而奔驰排名第三,特斯拉的自动驾驶和全自动驾驶组合排名第八。特斯拉被扣分的主要有两个原因。”
The first was not offering a collaborative experience between the human and the car, meaning if you wanted to avoid a pothole using Tesla's FSD, if you try to move the steering wheel too much, of course it disengages the system. Whereas other automakers like Mercedes offer a more collaborative experience. The second reason Tesla lost points was more controversial, she was saying that if you cover up the interior camera of the Tesla that the system will still work. Obviously we know that's not true with FSD, but it may actually be possible with Tesla's basic autopilot.
I can't test it today personally because both of our cars have FSD. And given the video did not explain those nuances, I did find that to be quite misleading. As many of us know, Tesla's driver monitoring is maybe even too aggressive. You can clip parts of this video and make it seem worse than it really was, but she did speak highly of Tesla in some regards. In terms of the capabilities, it's among the top. Mercedes does beat it out just a little bit, but it's in the league with BMW and others, so if you are looking for a high performer, Tesla's up there.
And they did actually touch on Tesla's ability to stop at stop signs and stop lights and do all of those things that the other competition definitely cannot. But they also left out key details about Mercedes system like how it's super limited and it has to be the perfect conditions for it to work. For me, the weirdest thing about the whole video was that Ford took the number one spot, but they barely mentioned Ford throughout the entire video. If you're new to the channel, this is a quick explanation of why I believe that's the case.
On the Consumer Reports Philanthropic Partnerships page, you'll find the Ford Foundation at the bottom of the page. Then, reading CR's annual reports, you'll find the Ford Foundation at a giving tier of over $100,000. And then on CR's media page, they had a digital safety initiative and scrolling all the way down, you'll find the Ford Foundation has contributed $1.5 million for this effort. And yes, it's true, the Ford Foundation is separate from Ford the company, but obviously there's plenty of overlap and connection. And just FYI, you can find similar information on the Ford Foundation's grant webpage.
As always, do with that information, what you will. We have what I think is a pretty big shakeup in the humanoid race. As Brett from Figure said today, I've made the decision to leave our collaboration agreement with OpenAI. Figure made a major breakthrough on fully end to end robot AI built entirely in house. We're excited to show you in the next 30 days something no one has ever seen on a humanoid. The figure agreement with OpenAI was supposed to use OpenAI's language and reasoning capabilities paired with Figure's robots aimed at enhancing the robot's ability to understand natural language and to process visual data and ultimately to be able to do certain things autonomously.
So I gave you the Apple because it's the only edible item I could provide you with from the table. Great. So based on the scene right now, where do you think the dish is in front of you go next? The dishes on the table like that plate and cup are likely to go into the drying rack next. And Figure's collaboration with OpenAI was part of the reason that Figure was able to hit a $2.6 billion post funding valuation after raising that $675 million round. That all happened less than one year ago and now Brett is saying, well, we don't really need OpenAI anymore because we have something brewing in house. The problem is we know very little about what Figure has going on behind the scenes specifically when it comes to things like a compute cluster.
There's really not much out there but Figure has had job postings in the past for roles like AI training infrastructure engineer. Responsibilities designed deploy and maintain Figure's training clusters. Through an NVIDIA blog post, we learned that Figure's O2 conversational humanoid robot taps into the NVIDIA Omniverse and NVIDIA GPUs for fully autonomous tasks. According to NVIDIA, Figure was using the NVIDIA DGX for AI training, NVIDIA Omniverse for simulation, and NVIDIA's Jetson in the actual robot. So my point here is I think Brett's being a little fast and loose using the term in house. Remember outside of simulated data, the only thing Figure really has in house would be those robots that they deployed to BMW where they're now collecting some real world data. I really don't want to come off as a skeptic. Let's just say I have my reservations about Brett and everything that he says. Alarm bells also go off when people start throwing around buzzwords that Tesla has made famous like fully end to end and built entirely in house. But I'm here for all of the competition so let's see what they have.
I was sneaking around earlier today trying to capture my wife in her natural habitat but it did not take her long to spot me. She of course is fully on board with ensuring she's not missing any key nutrients in her diet as she's a big proponent of nice hair skin and nails. AG1 is the sponsor of this video but let's just stick to the studies for a second and talk about some of the overlooked ingredients in AG1, the first being spirulina. Studies have shown it can enhance athletic performance and reduce oxidative stress. Not only that but it's antimicrobial, anti-cancer, anti-inflammatory, and it's among the world's finest sources of organic nutrients. Moving to ashwagandha, it's been linked to stress in cortisol reduction, lowering cortisol by 27 to 30% in stressed adults. It's also been linked to better cognitive function which enhances memory, focus, and sleep quality in clinical trials. And please do me a favor, AG1 aside, do not sleep on vitamin D especially in the winter as over 50% of people globally are reported to be deficient in vitamin D. With the welcome kit from AG1, you get the vitamin D3K2 dropper. And just one more, over 12% of Americans have an iron deficiency which is often linked to B12 deficiencies. And AG1 has a host of B vitamins including B12. There are certain things in my life I'm willing to pay for, one of those being the quality of ingredients that I put into my body as this is the only one I'm gonna get. So if you'd like to try it out for yourself and grab that free welcome kit from AG1, my link is below, it's drinkag1.com slash electrified, or you can use the QR code on the screen. Enjoy.
Finally, after weeks of unknown delays, the Cybertruck all-wheel drive variant is officially qualifying for the $7,500 tax credit. So with that, maybe a military discount and a referral credit, you can get a Cybertruck all-wheel drive around $70,000. The past few weeks, there have been more anecdotes about slowdowns regarding the Cybertruck, so hopefully this will be a slight shift to the narrative and we'll see how quarter one comes in. It does seem like ultimately though, for Tesla to be doing over 100,000 Cybertrucks per year, they need to get that cost down to between 50 and 60,000. There's a lot of new technology, it's going to take some time, but it's been in production now for a year, and typically year two should be a lot more smooth and most of the problems should be worked out by now.
For the Cybertruck in Canada, however, things are going the other direction when it comes to the foundation series pricing. As the price has increased for the all-wheel drive and the cyber beast by 11.5 and 9.6% respectively. But the non-foundation series Cybertruck is now also available for purchase in Canada, the all-wheel drive just under 115,000 and the cyber beast just under 143,000. First deliveries expected this month. The non-foundation series is also now available in Mexico. For the tax credit for the Cybertruck in the US, it should be noted it's only for 2025 model year vehicles.
Ferrari is gearing up to launch its first fully electric EV on October 9th later this year. We know very little about this vehicle, but they've said it'll be launched in a unique and innovative way. Ferrari does already have some hybrids for sale and a fun fact, hybrids made up 51% of Ferrari's car sales last year. The spy shots do have plenty of camouflage and misdirection going on, so be careful with these photos, but the rumor is this vehicle may start over $500,000. The Model 3 is now part of the referral program in Canada. Ford put out its US sales data for January this year and EVs were up 21.2% year over year, coming in at over 5,600 vehicles for the month. F-150 lightning volume is still very low down over 15% year over year under 2,000 units for the month, and you have to remember January of last year they had those production shutdowns.
The Mach-E for Ford really is the needle-mover right now, up over 172% year over year, just over 3,500 units for the month. And Ford's commercial eTransit was actually down over 79% year over year with only 230 units for the month. I have not seen Ford make any strategic announcements about changes with the eTransit, so we'll see if this is a one-off or if the trend continues. Today, China announced some retaliatory tariffs on the US, 15% on certain things, and then 10% on some vehicles, all starting next Monday. But Trump and Xi Jinping are set to talk later this week, so we'll see how that negotiation goes. Potentially Tesla's Model S and X that they export from the US into China could be impacted, but the numbers are quite low and there are no guarantees.
GM has officially cut 50% of the workforce at Cruise, and they've also acquired full ownership of that Cruise division, bringing those workers and NEIP back under the GM umbrella. Norway's new car sales data for January is in came in 96% were fully electric. January of last year, that number was 92%. Then, when including hybrids, the number was just under 97%. Elon's involved in a new little spat with the head of Norway's $1.7 trillion sovereign wealth fund. The spat is with Nikolai Tanyan, who voted two times against Elon's pay package. A while back, Tanyan invited Elon to a dinner. Elon said he could not go, and then Elon sent some text messages that Tanyan somehow made public and Elon is upset about that. Elon said, when I ask you for a favor, which I very rarely do, and you decline, then you should not ask me for one until you've done something above nothing to make amends. Friends are as friends do.
Also getting mixed up in this whole thing, Europe's largest pension fund actually sold all of its Tesla stock in the third quarter of last year, citing concerns over the pay package. Which means they sold all of their Tesla at share prices between $199 and about $270. AKA right before the run up. Tesla stock closed the day at $392.21 up 2.22% while the NASDAQ was up 1.35%. The volume was 26% below the average. Don't forget, check out AG1 linked below, grab that free Vitamin D3K2, especially if you're in a place like me, where natural sunlight just is not an option certain months of the year. 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.