Tesla Optimus Delayed as China Holds the Magnets

Executive Summary

This week on Autonomy Signals, Grayson Brulte and Rob Grant discuss why the actual deployment of autonomous systems remains constrained by Asian supply chain dependencies, which limits near-term revenue realization for humanoid robotics programs. Concurrently, the European Union is delaying its high-risk classification for L4 vehicles under the EU AI Act, while Waymo sets its sights on Canadian expansion through strategic lobbying in Toronto and Vancouver.

Key Autonomy Signals Episode Questions Answered

Autonomy Signals Topics & Timestamps

[00:00] AUTNMY AI

Welcome to the debut of Autonomy Signals, a new show co-hosted by Grayson Brute and Rob Grant. The hosts set the stage for a weekly format designed to cut through the industry noise and deliver the most important signals in the autonomy economy.

[00:24] Signal 1: Potenial Tesla Optimus Gen 3 Delay

Tesla’s Optimus program is facing potential delays and rising costs due to a shifting regulatory environment in China. The Chinese Ministry of Industry and Information Technology (MIIT) is cracking down on dual-use technologies—innovations that serve both commercial and military purposes. Because China controls 90% of the rare earth materials needed for Optimus’s specialized magnets and actuators, the Chinese government is demanding that companies share technical specifications and secure permits to use these materials. This creates a massive structural choke point; moving production to America could raise the estimated cost of an Optimus unit from $46,000 to $133,000. Furthermore, the FTC is currently investigating Tesla for potentially overstating its “Made in America” branding given its heavy reliance on Chinese components.

[23:35] Signal 2: Europe Delays Classifying L4 Autonomous Vehicles as High Risk

The EU Parliamentary IMCO and LIBE committees voted 101 to 9 to delay classifying Level 4 autonomous vehicles as high risk under the EU AI Act. This pushes the compliance target back to August 2nd, 2028. The delay highlights the massive regulatory uncertainty in the EU, creating an environment that discourages autonomous vehicle companies from entering the market. Specific compliance requirements, like strict privacy by design and localized data storage, will create costly hurdles for companies that rely on cloud computing or preloaded maps. OMEGA’s analysis suggests the EU is deliberately weaponizing compliance to build a regulatory walled garden that protects domestic legacy incumbents, such as Bosch, from foreign software monopolies

[48:45] Signal 3: Waymo Eyes Canadian Expansion

Recent lobbying records reveal that Waymo is targeting Toronto and Vancouver for its expansion into Canada. Toronto (in the province of Ontario) represents a much smoother entry point, as the region has given a “soft embrace” to AV technology. In fact, Waymo’s Co-CEO has already met with Toronto council staff to discuss regulations for ride-hailing and goods delivery. Vancouver (in British Columbia), however, presents significant regulatory roadblocks, as the province currently bars any technology above Level 3 from its roads and requires separate insurance approvals from the ICBC. OMEGA notes that Waymo will have to navigate complex jurisdictional gridlock between provincial and federal authorities to secure exemptions before commercial operations can begin.

[51:29] Closing

The hosts wrap up their successful first episode by highlighting their proprietary algorithm, OMEGA, which powers their research. Listeners interested in learning more about the AI are invited to reach out via email to alpha@autnmy.ai. They sign off with the show’s official motto: The Future is Bright, The Future is Autonomous, The Future is Autonomy Signals.

Full Episode Transcript

Grayson Brulte: Welcome to Autonomy Signals, the new show from The Road to Autonomy, co-hosted by myself, Grayson Brute, and my partner Rob Grant. Each week Rob and I are gonna cut through the noise and deliver you the signals that matter the most. This week we’ve got three really interesting signals. One, there’s the possible delay of Tesla’s Optimus, Gen 3, due to the regulatory environment in China, China, and potential rising cost. Then there’s the EU who says, well, wait a second. We’re trying to regulate AI too fast. What’s that gonna look like with the impact on autonomous vehicles? And then Waymo, the leader in America, there’s potential. We’ve got lobbying records. They’re looking to go to Canada. Rob, let’s start with Tesla. What’s the delays potentially coming outta China? ’cause this new regulatory environment that’s taking place.

Rob Grant: Thanks Grayson. Uh, really excited for what we’ve got building here. Um, and. Look, OMEGA has provided us some increasingly important and relevant information about what is going on out of the CCP and Chinese regulations. In particular, the Ministry of Industry and Information Technology is cracking down on what they call dual use technologies. Now, most of our audience knows what that means. That means it’s a, uh, a technology that can be used for commercial and military purposes. And so the MIIT in China. Has put on a new series of restrictions about, um, needing a permit and to meet new standards over in China for the use of dual use technologies and where this hurts. Tesla and their Optimus program is that Tesla relies on specialized magnets, um, in order to, uh, use in Optimus particular for their hands and digits and actuators and things like that. And China controls 90% of this rare earth material. And so what the Chinese government has essentially said is, you must share your technology, uh, and how are you gonna use it? And some of how it operates, uh, and some of the technical specifications behind it, uh, with us in order to get a license, if we believe your technology is capable of dual use, um, as we all know. From the videos that we’ve seen out of China, they are really prioritizing, uh, development of humanoids, unitary, and exome and others are really making great headway there. And their fear, uh, is that the US is going to compete with them, uh, in particular, uh, Tesla. And so what the MIIT is saying here is if we believe your technology can be used for military purposes as well. You’re relying upon rare earth materials sourced and purchased from China through Chinese suppliers, um, then you are going to have to, uh, get our approval to do that. Uh, and that’s really gonna put, uh, a, a real kind of structural impediment to what. Tesla wants to do on the timeline that Tesla wants to do it in terms of developing, uh, mass producing Optimus, robots, and the revenue that comes along with it.

Grayson Brulte: There was a lot of revenue. But the one thing that Elon has posted on XA lot about, I’m moving my hand dexterity the special rail earth magnets to give it the 22 degree dexterity. They come out of China. So the government says, well wait a second, that can hold a weapon. We’re gonna put export controls on that. So if we look at the cost that OMEGA has uncovered from CCP documents. You’re looking at an estimated cost for Optimus today of 46,000, but if Optimus is to be produced in America, that costs rise to 133,000. And it’s not just the cost aspect of it is Tesla is highly reliant on the Chinese supply chain. Obviously, the CCP knows this is saying, Hey, well you got some superior technology. You know, by the way, you’re relying on us, so, eh, we’re gonna put, we’re gonna put some rules in place here.

Rob Grant: Yeah, I mean, I think, I think we’re talking, this is not like a supply chain risk in the traditional way that we think about it, right? This is a structural concern. These rare earth magnets are, are essentially monopolized by the Chinese government. Uh, so there are opportunities for Tesla to look elsewhere, including onshoring. Um, but that is a long-term process. Um, and in the short term, really the only place you can go to for, for consistent, reliable supply of these magnets is China. Um, you may be able to find some in other places in, in Asia, Japan, and Korea, or potentially in Australia. Um, but that is not something you can just turn to overnight. Um, these are, are, are structural risks right now, uh, for Tesla in terms of being reliant on these Chinese rare earth materials In addition. What, what we’ve uncovered through OMEGA’s research and its ability to search through public records as well, is the fact that the FTC here in the United States is now investigating, uh, Tesla, uh, for perhaps, um, overstating. Its made in America, uh, branding with the reliance on these Chinese, uh, suppliers, uh, and subsystems, as well as Chinese rare earth materials. FTC is investigating whether or not Tesla meets the requirements for the amount of materials that go into a product to say publicly that it’s made in America. And so not only are is Tesla’s reliance on Chinese materials, structurally putting an impediment because of the new rules that are going into effect in China. But also they’re causing issues here in the United States as well. And so, um, you know, what we potentially see is a delay in the realization of the revenue stream that Tesla is stating, uh, for its Optimus, uh, robot technology.

Grayson Brulte: There is the potential delay. And on the FTC note, this is according to the FTC and there’re probably documents that over 60% of Optimus is components come from China. That’s according to to the FTC. And if that’s not enough, well, OMEGA also uncovered this from the CCP that I want to talk about here on March 2nd, Rob, the MIIT released the humanoid robots and embedded in intelligence standardization system. Oh boy. Here. Here we go. So what does that mean? That means the standard suppliers that that provide the intelligence they have to privatizes Chinese humanoid projects like those from Xiaomi or Unitree that comply with the new national technical standard protocols. Tesla, as we know, these proprietary closed loop systems, so the way that this law is written in China for this new standard, those suppliers that are supplying Tesla have to deprioritize building for Optimus and prioritize building for Unitree . That is another risk that OMEGA uncovered.

Rob Grant: 100%. I mean, this is really. Where you’re starting to see the globalization in the global race between the US and China on human aids, uh, come to fruition. Uh, the Chinese have given as they have in other areas of, uh, emerging technology, their domestic companies, a leg up in terms of, um, the, the use of Chinese materials, uh, the direction from the Chinese government about how those materials need to be prioritized among clients. And, uh, what this is doing is, is saying that there is, um, a geopolitical importance to this technology that I think goes, um, underappreciated in the market, um, accounts of what Tesla and other US humanoid robots are building. Um, you saw it today, right at the White House. You had a figure ai, uh, robot walking down the, the red carpet with Melania Trump. Um. The US government recognizes the importance of the development of humanoid technology and the US no doubt is the leader in the cognitive AI layer of this technology. But in terms of the actual base infrastructure layer, um, and its physical execution layer. Uh, Chinese have a real grip on this technology. Um, and that is a structural risk that I think the US businesses, uh, uh, figure ai, Tesla, others are going to have to really wrap their arms around in terms of how do they continue to build this technology at scale? How do they get the cost? Down, uh, for building at scale. And, you know, it’s a real tough question given that the materials that they need to build this technology are not under their control. Um, they are, they are subject to the whims of the priorities and demands of the Chinese government right now.

Grayson Brulte: And that’s a scary place to be if you’re looking at it and having to rely on them. And I’m having flashbacks here. We’re not, we’re not going back to the future, but we kind of are gonna go back to the future here. Of electric vehicles, China became a giant choke point on that. It seems like up here we go again. History is repeating itself, not with autonomous vehicles this time, but with humanoid. With that saying, when do you think we start to see policies to put in place in the US that pro prioritize this type of manufacturing? ’cause I noticed today with the Figure AI at the White House, it had an American flag on his front chest. Very, very proud. That was a, to me, it was a very big symbol there.

Rob Grant: Yeah, I mean, look, the US. Uh, when it passed the CHIPS act a couple years ago, I think is starting to realize that, um, manufacturing of these, uh, rare earth materials, uh, refining of them, discovery of them is a real national security, national economic issue. And so I hope we don’t have to find ourselves in the position that we did with EVs and battery technology where we’re having to play a severe game of catch up, uh, and, and looking. You know, um, outside, uh, of the US for innovative technologies on battery. Uh, and so I think what we have here is a real clear signal early on in this humanoid, uh, race, so to speak, that the US needs to start producing its own rare earth materials and in greater volumes. And, you know, I don’t think. The companies that are playing in this space are ignorant of that fact. Uh, I think they understand it. Um, you know, we have Tesla building. It’s, it’s, it’s a new lab. Uh, it’s Terraform lab. Perhaps that is in response to some of the structural things that we’ve talked about today. Um, but I do think. You know, as we look at an increasingly complex world driven by geopolitical events, you know, everything from what’s going on in the Middle East to what may potentially happen around the world, um, onshoring and having the domestic capabilities for building this technology is, uh, has to be one of the US’ top priorities going forward.

Grayson Brulte: It has to be because for all practical purposes, China said up. Optimus, sorry. We’re gonna put a hold on you because Optimus requires approximately 3.5 kilograms of high performance N-D-F-E-B magnets per unit to service 40. Servo motors. So what does the MIIT say, oh, oh, guess what? You have to get a non-military certification. So what does that do from an economic standpoint? It puts Tesla $685 million order for linear actuators on hold, under review. Without that order going through, Optimus is gonna get delayed. And, and you, you alluded to it, but I’m sitting here saying, okay. To me is that why Terrafab is, is getting accelerated. And yesterday I was in Fort Worth with, uh, Mr. Perot Jr. And Governor Abbott came and talked about the factory where he used to say, and they’re accelerating it. Perhaps that’s why we’re seeing this massive acceleration. The Terrafab say, wait, we can’t be relying on this.

Rob Grant: it could be. I, I, I mean, I think from a strategic point of view, look Optimus, uh, long-term. Still a great bet to, to succeed and, and produce tremendous amount of, of volume, uh, and revenue for Tesla. I think the short, it’s the short term. That we’re looking at right now. The structural reliance on these materials from China is a real risk. And so having to get this approval from the M-A-I-M-T for, um, or convincing them that it’s non-dual use basically is what Tesla has to do. Um, there’s no timeline on that there. You look at what we know from that rule. Um, China doesn’t have to act until it wants to act. Um, and so that, that really puts a, a difficult, uh, um, kind of unknown into what Tesla is trying to do in the short term. Um, look, it’s not about is our technology is Tesla’s technology, um, advancing at the right clip? I think it, it, it is. I mean, everything that we see and everything that we, we understand of it, they’re not behind technology wise, right? They may even be ahead. In terms of the components that go into it, the physical execution, the, the, the, the structural risks, um, they’re real. Uh, and, and this is something that I think folks need to take account of as they determine what risks that they wanna place, uh, in, in the Optimus, uh, a product line.

Grayson Brulte: And if that risks that you described. Are not enough there. There’s another risk that’s even bigger. China’s implemented a Chinese version of the US foreign direct product rule. So what does that mean in the theory, China could say, no, you can’t export that product.

Rob Grant: Right. You can’t export it, even build it somewhere else. Right. That’s the thing. Um, and, and so I mean this, this is something the US has done, um, as well. So the, this type of rule kind of globe. Will reach in scope is not new. Um, the US has done this with certain, uh, products as well. Um, but it’s, it’s kind of new for, uh, China in this area. And what that says is even if you collect all the component parts in the right way from China and build them in the United States, the Chinese government still has an ability to reach into the US until, you know you can’t use those parts that you’ve acquired under the right rules in the right way. To build the thing that you wanna build.

Grayson Brulte: You, your right, even if you’re assembling in, in California or Texas, nope. You can’t do that. You can’t even get out of the country. So. It’s a choke point and it’s, it is a, it is a huge risk and it’s something that that’s what OMEGA’s so good at. She’s good at really uncovering these signals and one of the features of our proprietary OMEGA algorithm that gives OMEGAs take, and I want to go through each one of these OMEGAs takes with you, Rob, as it relates to the humanoid signal here, and this is to this is to quote OMEGA. The actual deployment of autonomous systems remain constrained by Asian supply chain dependencies. This friction limits near term revenue realization for humanoid robotics programs. I fully agree with that. I think she’s spot on with that one.

Rob Grant: I think, I think it is, it is a, a huge near term, um, structural disadvantage. As I say, I remain. Very positive in the long term because I think these are issues. Um, they’re not unknowns in terms of how to address them. It just takes time to address them.

Grayson Brulte: It does. And then here, here, here’s another take from OMEGA. The Optimus delays expose a critical structural divide in the global autonomy race. The United States dominates the con cognitive layer of physical AI, but Asia maintains an iron grip on the physical execution layer. We talked about it. That’s right. Does this take Secretary Lutnick and, and the Department of Commerce to say, okay, we have, for all practical purposes, a majority of the brains that were built here, and I say majority of the brains because let’s not forget the way that the DeepMind deal structure, they’re still in the uk. They’re a UK company, so you, but they are an ally. But the US is still ahead on the Frontier Lab research, but we don’t have the ability to take that digital ai, if you wanna call it the virtual AI, and put it into physical yet.

Rob Grant: Yet, I think that’s the key word. And I think you’re seeing this, right? You’re seeing this in different fields beyond humanoids where, um, you know, we’re trying to build our own, um, uh, chip making facilities. We’ve got government support for that. We’ve got Terra Fab, we’ve got other. Um, production of these types of not only raw materials, but finished materials that we’re trying to bring back into the United States in order to ensure that this structural divide that we’re seeing between the physical and the cognitive, uh, that that divide is, is closed and it’s closed in the favor of, of companies in the United States that are developing these technologies.

Grayson Brulte: You’re right, and that raises the labor cost. And here’s another take from OMEGA, until Western manufacturers can replicate the cost efficiencies of the Chinese industrial robotics ecosystem, the commercialization of humanoid robots were remain gated by geopolitical tensions in overseas tri logistics rather than neural network advancements. OMEGA tracks the global shipping lanes and we’re seeing what’s happening in the straight. And let’s not forget, I’m sorry, there’s a little issue with Taiwan. It just seems that there is a lot of risks here. And while Onshoring eliminates the, the geopolitical risk allows companies such as Tesla to build Optimus and figure to build their humanoid costs. Is this come down to advanced manufacturing and automation and factors? How do we get those costs down?

Rob Grant: I mean, the cost is gonna be the real, real issue here, right? I mean, I think what we see is, um, you know, from what Tesla says, it’s about what, $46,000 or so for building, uh, each of their robots. If, if, if, if some of these, uh, immediate effects of what the Chinese government has done. Um, uh, are taken into account. We’re looking at maybe a, a, a new cost of $130,000. It’s almost three x the amount, uh, to, to do the same thing with materials that aren’t from China right now. And so, uh, there are other places you can go, uh, for some of these materials, but that cost doesn’t come down to, to where the Chinese are producing it right now. Uh, you might get it down to 80,000 or $86,000. But at the current level of $46,000, there’s no other place to turn to other than China for materials that come at such a reduced cost.

Grayson Brulte: The one thing I think a lot about, and a friend of mine recently spent time there was in Perth, Western Australia, and when you speak with individuals that are politically connected in Australia, they said that getting individuals to relocate to Perth is near impossible. That’s why they’re, they’re looking at automation, but however you have the raw materials there, could we get to a point where. Some sort of global manufacturing goes into Perth with, with refining, and they try and build an ecosystem which would really help that western part of Australia’s economy.

Rob Grant: I mean, I think so, I think, I think you see that in Australia, you see some of these, uh, particularly magnet, uh, the, the type of materials that are going into the magnets that we’re talking about. You see them, uh, in Australia, you see them in Canadian deposits. Um, I think there are real opportunity there. Uh, but again, these are not things that you can simply just turn on, right? It, it, it is a, capital intensive, operationally intensive, um. Procedure in order to extract these materials and then to get these materials refined and put into the right types of, of chips and other things that can be used by the companies that we’re talking about. And so I do think there is opportunity there. Uh, I think both the governments and businesses see it. It just, you know, what we’re doing with again, is the constraint of time. Uh, and so, you know, short term, this could be a real kind of, Uh, Uh, str like we said, structural impediment, but long term you solve these issues and you become even stronger and you’re able to produce them, uh, without the risks of these geopolitical tensions. And, and you mentioned one that we haven’t even really touched on, which is potentially what happens with China and Taiwan. Uh, and we should probably reserve that for a whole nother conversation, but that, that is something that, that, that goes beyond, adds another layer of risk to you in what we’re talking about today.

Grayson Brulte: There’s a lot of risk in the OMEGA bottom line. ’cause after every research report OMEGA generates rush, she gives us a bottom line. The bottom line on the this signal this week is the rare earth choke point is not a risk. It is the current reality. The consensus view frames China, rare earth export restrictions. As a headwind, the Tesla’s Optimus timeline. Tesla’s Optimus program is not delayed by engineering complexity. It is being throttled by Beijing’s, deliberate weaponization of rare Earth and dual use technology controls OMEGA’s. Right? And I think it’s, it’s a signal that I’m happy we’re here to talk about today ’cause it hasn’t been covered a lot.

Rob Grant: Agreed. I, I think it’s one of the, the most underappreciated elements. The development of the technology, uh, its ability to, um, drive revenues, uh, drive that multiple, uh, Tesla that will get, uh, Elon to his trillionaire status. Um, but it’s also, you know, gonna have some, some direct impacts on when consumers are gonna be able to engage with this technology. Uh, and, uh, when, uh, this technology is, is going to kind of become real for most people, which is like, how will it help me in my day-to-day life? And so I think all this is a potential timeline delay. Uh, I don’t think it’s, it’s decades of delay, but I do think it will have, uh, a kind of delay both in terms of its financial impact on Tesla as well as eventually the consumer impact of when I can interact with this technology.

Grayson Brulte: I agree with you. It’s gonna be a delay. We don’t know how long the delay will be, but. If you read some of the studies that have come out of MIT, or actually some out of the Milken Institute of Aging, when you put a humanoid in a senior home or an individual that’s trying to live on their own, they can interact with, okay? Remember to take this pill today. Remember to take that pill today. Okay? Jeopardy’s on a five o’clock. Companionship. And then when you have, when they have that companionship, they have a higher quality of life. So it’s not just the talks around, oh, this is gonna be good for factories. Oh, this will be fun. It can, it can play soccer in the backyard with your child. No, it’s gonna have a really positive effect on humanity and, and society. And I think that’s something that we’ll continue to talk about. And hopefully we, we could see more signals or let’s move on to the. The next signal. Next signal has taken us to Europe. Where on March 18th, 2026. The EU Parliamentary IMCO and LIBE committees voted 1 0 1 to nine to delay. That’s right. Oh boy. To delay classifying L4 vehicles is high risk. Under the EU AI Act compliance target, it’s been delayed to August 2nd, 2028. I’m just gonna say it was the EU going a little too fast on regulation. Like, well wait a second, because when you’re voting 101-9, that, that’s not exactly a close vote.

Rob Grant: Yeah, it’s, it’s really interesting. So, you know, the EU AI Act, uh, was passed a couple of years ago. Uh, this is in 2024 and now we’re sitting here two years later, uh, a couple of months out from its effective date, which is August of 2026. Uh, in between. The EU had put upon itself kind of a directive. Let’s develop some of the guidance and procedures and requirements that folks, uh, who are designated, uh, to be participating in, in high risks activities with ai, uh, need to meet now. They haven’t put any of those. Policies or procedures or standards out. Uh, and so what you have is this great uncertainty, uh, about how one would comply with these things, uh, when they need to comply with it. And, and you know what, what? Things that you can do today to prepare yourself for what’s coming. Uh, and, and where this gets really interesting is, you know, right now, um, without further action. And there is one more step that needs to happen for this delay to take effect. This rule goes into effect in August. Um, and while all the, all this, um, new bill is saying as well. We think we’re not gonna hit the August timeline. Everybody knows we’re not gonna hit the August timeline, so they will go into effect in December of 2027. And for embedded ai, which is what goes into autonomous vehicles and things like that, um, the new rules won’t take effect until August of 2028. Basically, you know, six, seven months after most of the rest of the thing would take, uh, effect in December of 2027. What this really signals is, look, the eu, uh, again, um, is just providing regulatory uncertainty on top of regulatory uncertainty. It’s already difficult to operate, uh, there as a AI or AV company. Um, mostly because there’s the intersection between what they’ve got going on for the, what they call sectorial legislation, which is the legislation that kind of applies to how you might build a car, uh, and the standards that you have to do for that, as well as what the AV standards for operating and safety are for each particular jurisdiction. EU and now you have an additional layer of uncertainty, which is how will the intersection between this AI act and those sectorial legislations operate? And so really what it becomes is, is, is another reason for folks to say, Hey, you know what? I’m gonna go look. At deploying my technology, if I’m a Navy company in the Middle East, I’m gonna look at, uh, the Asia countries. I’m gonna look at London, who’s no longer a part of the EU Right? And this is why you’re not seeing, I believe many folks saying, we’re immediately gonna look to Target Germany, or immediately gonna look to Target Spain. Now they’re, they’re on the roadmap. They wanna be on the roadmap. But even if you’re. Filled with lawyers and regulatory attorneys as, as some of these places like Waymo can do, right? They can, they can have the smartest minds looking at this. There’s still so much uncertainty there and really it’s just hurting not only the technology coming there, but also the technology that wants to be built there. Um, and so, um, you know, again, I think this is another EU kind of self own, which is like. You wonder why we haven’t seen many autonomous vehicle companies out of the EU countries and, and this is an example of why.

Grayson Brulte: And as part of this legislation, I almost fell it on my chair because I’m like, how do you define this? Privacy by design for all the sensors, all the data has to be encrypted. Can’t have any identifiable information, must be kept in the EU Sensors. Privacy by design. We see Bosch is coming out with a new privacy sensor. If I’m looking at this from a technical standpoint, I basically, and this is more of your wheelhouse than mine. I’m like, oh God, I gotta have a whole separate different car with different sensor suites. To meet this regulatory environment. Am I reading and interpreting this right?

Rob Grant: it could be read that way. Yes. Uh, I think really what it does is it gives a prioritization to those companies that can meet this privacy by design or in some senses, what it does is it give a prioritization to companies that aren’t using. As much, uh, cloud computing, uh, technology. And so if you are somebody who’s not using, you know, maps, uh, you’re going map list or you’re relying purely on cameras where you have most of your data and, and data collection, uh, coming from the edge, right from the compute unit on the vehicle itself, um, you’re more easily going to be able to meet. What we think are gonna be the requirements under, uh, the EU AI Act. Um, now if you’ve got preloaded maps or you want to take your data, uh, and put it up in the cloud, uh, and train your model off of that, um, that’s gonna be a lot harder to do under this, right? These, these are basically localization requirements. Uh, and so, um, if you are in some kind of hybrid model, which some of these. Uh, companies are, which is use of preloaded maps, moving your training data out over the cloud. You know, my, my assumption is that Waymo does this. Um, you’re gonna have to look perhaps at a, a, either more localized. System or you’re going to have to look at different sensor suites. Um, and that could be very difficult if you’re having your car produced online with a certain sensor and that sensor is not EU compliant under this act. Um, you build a different line. Do you have to then kind of after, uh, manufacturing, go back and change each of the cars that you wanna put put into the eu? Uh, this all adds cost and time, right? And it’s another disincentive to come there. But if you are somebody like Tesla and you’re using camera only and you Edge computing is there and you don’t have preloaded maps and you’re, you’re able to kind of train your data locally. This may provide you an advantage. Um, so I don’t know if it’s as crystal clear as like everybody is penalized under this as it is that some people are, I would say if I’m, you incentivized to do something different, um, but it could, it could wind up being a cost difference for a Waymo to operate in the EU from what Tesla is doing in the EU So that may give an advantage to, to Tesla as well.

Grayson Brulte: It could, but then the question is, and this is all can be done with AI, you’re gonna have to blur all the faces. And then I think about this, I’m like, oh boy, is this GDPR all over again where you’re trying to put a defacto global standard? Is that where this is going?

Rob Grant: I think so. Yes. Yes, for sure. but but I mean, look, these are not new types of prescriptions. We see this in many places, right? If you try to operate a financial center, uh, or a data center in the Middle East. You have localization requirements. Data must be kept locally. China does this, right? Uh, all of the data from, um, your vehicles in China, if you’re, if you’re even allowed to operate, there are, all that information has to be kept locally within Chinese. So these kind of walled regulatory gardens, so to speak, about where you can keep data, what you can collect, what you can’t. They’re not new, but they’re also, um, in a sense. Anti-competitive, right? It doesn’t create an open market. Uh, and so if you are an external company, a US company trying to come in, um, assuming that you’re going to operate under similar rules as the us this may not be the case. Uh, and you’re gonna have to create a specific. Vehicle or a specific set of, um, components or even a specific data center, uh, where you train your EU model, so to speak. Um, and that’s, that’s another cost, right? And so that’s another reason to say, you know, when does this technology come there? It could be delayed. Uh, and I, I think the U seems to be fine with that.

Grayson Brulte: It’s sad really when you think about it. If you just look at the, the German automotive industry, once the, the, the crown jewel of the world for automotive, no longer the crown jewel. And as of today, we ride is operating in several European countries. Biden is operating in several European countries and they’re eventually gonna be on, on Lyft and Momento is gonna be there, and then Uber’s going to be there. There’s a lot of data that’s gonna go back and forth between the AV partner and the robotaxis platform provider. It seems like based on this, when August 2nd, 2028 rolls around, oh boy, somebody’s gonna file a lawsuit. Say, yeah, we’re not gonna comply with this. It just seems that this is not gonna be resolved by then unless I’m completely mis misreading this entire picture.

Rob Grant: Look, I think this is part of the reason they need a delay, right? Some of this stuff is supposed to be addressed, uh, in that was supposed to be addressed in the intervening years between 2024 and 2026. Uh, and look, this not only applies to avs, right? This we’re talking about embedded AI and avs for, for, for, for our audience. Uh, but this applies to all the large language models. All all the things like that. So if you’re a small, um. Company, right? And you wanna use, uh, AI technology to help you hire over in, uh, the EU countries. You’re not quite sure if you can do that, right? Uh, because you’re gonna have to prove if somebody sues you that the AI wasn’t biased in some sense. Um, and so there’s a lot of openness as to how you would prove the AI wasn’t biased. And this could go for the same for an av, right? You have to have this kind of traceability and decision making under this act that’s gonna be really interesting, um, for certain companies. Right? I talked about how Tesla may benefit under this in one sense, but they could also, uh, on the same side have a real challenge. How are they gonna have the traceability of the AI’s decision making as to what the vehicle did in an incident, uh, some of their neural networks and things like that. It is really hard. To go back through the black box and figure out exactly why this decision was made at this time. And the thing is, the next time the technology encounters the same thing, its decision making may be different. Uh, it may have learned or it may be taking something else into account. Uh, and so it, it could produce real challenges when it comes to kind of meeting the requirements. Now, right now, the biggest challenge is the requirements aren’t even out there, so you don’t even know what you need to meet. Uh, hence the reason that OMEGA picked up the signal is, hey, the U is signaling it needs more time. Um, and that, that’s an interesting signal to give out to the market, which is like, Hey, we’re not loving the market for, for AI providers in EU to begin with. Uh, and now you’re signaling that you’re even kind of confused as to what you need to do yourself. Uh, it’s not a great signal for the market.

Grayson Brulte: It’s not a great signal. This is signal’s not from OMEGA. This is you and I going back a decade, decade and a half together. When you, you were in industry and I was in a, in a appointed government role and certain cities, not Beverly Hills, by the way, we never asked for it. Certain cities asked for all the data, and then you had consortiums that showed up trying to get the data. And we all know where that went. The city’s never got the data. It just seems like EU is trying to get that same playbook to go all over again. And if so, it’s going to be dragged out and. One of the private sector companies is probably gonna throw the privacy element back at the e. Will you wanna preach privacy while sharing this data with you is a violation of privacy. It seems like there’s a lot that still has to be litigated here.

Rob Grant: I think that’s right. I think that’s right. I think, look, I mean, ideally they have a robust discussion, uh, as to what types of information needs to be shown ahead of time, right? They’ve already classified this. AI embedded in AVS as high risk, right? Meaning that it could cause harm to life or limb. Um, I don’t think that classification is, is wrong. I think that’s a correct classification. But what are the kind of rules of the road now that you’re in that classification? How do you, um, produce the evidence? What data needs to be submitted? How frequently, how are updates taken into account? Uh, things like that. Um, those. those. Are really difficult questions and the further they kick ’em down the line, I think the further they come to a point where the technology is just passing by and all the questions are wrong to begin with. And so, you know, I think the eu um, needs to really start to understand and interact more with the companies and understand the technology. Um, is advancing. And so some of it’s thinking as to what they require from the technology. Um, it’s not incorrect, right? I mean, I’m not arguing that they shouldn’t be regulating in these areas and be, shouldn’t be considering things like privacy. But, um, really what this is, is a white flag almost saying we don’t know what to do with it. Uh, so we haven’t done anything with it. And at some point you need to fill in the details. If you want people to really invest and bring that technology, um, which I think you do, I think you want to create an ecosystem there. You want to have innovation. You wanna see this technology because it is good, right? Um, I think there are a lot of benefits to bringing this technology to places all around the world and then having people look at it and say, you know what? I can build something better, uh, than that. Um, and you can’t even get inspired to do that if you can’t even put this stuff out in the road.

Grayson Brulte: No. And that brings us to OMEGA’s take ’cause OMEGA was spot on with this one and OMEGA said the European Unions is executing a strategic counter offensive, having lost the foundation model software war to the US and the hardware manufacturing war to China. Regulators are weaponized in compliance to turn the continent into a regulatory walled garden. This explicitly subsidizes their legacy industrial base protecting domestic incumbents such as Bosch and Continental from foreign software monopolies. Why single out Bosch? Because they have an AI act compliant LIDAR and sensor suite that you need to deploy. This is European Union protections of 1 0 1. There’s no other way for me to say this to you.

Rob Grant: I look, I think OMEGA is spot on. It’s spot on. But, uh, you know, that protectionism, I think is real detriment to the development of their own kind of advanced technology ecosystem.

Grayson Brulte: that’s Europe. Now let’s go to Canada. Waymo. Is eyeing a Canadian expansion, not to one market, but to two markets that OMEGA uncovered lobbying records. She has great access to all this publicly available data. The markets that Waymo’s looking at deploying in based on lobbying records, Toronto and Vancouver. We know that the Canadian regulatory environment is, uh, not the easiest to get through. I’m not saying it’s as strict as the EU It’s complex. What do we know about these two potential deployments in Vancouver, in Toronto? That Waymo was ironing Rob.

Rob Grant: Yeah, I think, I think it’s a smart move by Waymo to look at the further expansion in North American market. Right? It, I think what people who aren’t familiar with kind of the Canadian way is that it’s a very provincial driven system, right? They do have a federal government, but the, the federal government powers aren’t as expansive as they are in the United States, and so. What we have here is kind of a tale of two provinces. Uh, you have the Ontario province, uh, where Toronto is located. Uh, they have kind of, I would say, given a soft embrace to AV technology. They’ve been in this position for a while. I, I would give them credit. They’ve not been kind of, uh, a lake comer to this. They’ve done a lot of research, a lot of talking with companies. Uh, you know, Toronto, the city of Toronto has done some pilot programs and things like that. So I think for Waymo, given what we’ve seen of its technology, right, and how well it’s worked in multiple cities in the United States, I think Toronto is a natural place to start expansion in Canada. I think, um, you have. Uh, a relatively open market. Uh, they are, uh, going to kind of wanna put the technology through its paces. They’re gonna want, um, reporting and they’re gonna want robust interaction with, with local law enforcement and things of that nature. But I think they have a real opportunity to, um, kind be the initial, um. Uh, a dominant player in the Toronto market, even if it starts at a small scale. Um, and so that’s a credit to Ontario and the folks there and the city of Toronto and, and, and some of the kind of forward thinking they’ve had on a lot of technologies in, in Toronto. Um, now Vancouver. Uh, they have a much different approach, uh, right. So, uh, in the, in the British Columbia province, uh, they’ve been a little bit hesitant, uh, to adopt new technologies and that, that’s putting it kindly. Um, we saw a similar divide happen when Uber and Lyft tried to get there in the, the, you know, middle part of the teens, 2000 teens, uh, where Toronto. It was a robust debate. I remember being in city Hall and at some of those meetings where a lot of the taxi drivers from Mississauga and other places were really opposed to Uber and Lyft coming there, but we eventually got it done, uh, whereas it took many years after getting into Toronto to get into Vancouver. Uh, and some of that, um, is being played out again in the AV space. So British Columbia has already taken a position. Bars, anything above level three technology from being on its roads. So there’s your initial impediment, right? It’s not like you’re trying to, uh, convince them to do something for the first time. They’ve already taken a stance, which is, we don’t want this technology here. Um, now if you look at, um, what the Vancouver City of Vancouver and other folks, uh, in the government there have said it. Span doesn’t prevent potentially a pilot project with somebody behind the wheel or something that might be akin to, you know, uh, manual testing or things like that. Um, but they are going to need to take an affirmative step to allow this technology on the road. And one other thing to keep in mind, um. You not only have to go through the government, uh, in terms of the city of Vancouver and the British Columbia province to get approval to operate, but you have a separate entity there called the ICBC, which is, um, the Insurance Commission of British Columbia. Uh, now they have kind of dominant, unique, singular control. It’s like the singular insurance body for automobiles in British Columbia. Um, and if they don’t. Provide you insurance, it provides you an ability to insure your vehicles out there. Your vehicles can’t go on the road. So you not only if your Waymo have to get the government in terms of operationally allowing you to move forward, then you have to get the ICBC to ensure your vehicles moving forward. Uh, and so there’s a whole nother challenge as well. And so. I think it’s important to start there, uh, and start that conversation, uh, which I know, um, OMEGA has uncovered that they’ve started this conversation a couple of months ago. Uh, and they’ve sent some high level, uh, folks including the CEO up there to talk to some, some of the government folks. Um, but. Actually seeing a meaningful expansion in Vancouver is going to be significantly after you see meaningful expansion in Toronto. And that has, and this is the interesting, right, that has nothing to do with the advancement of way most technology, right? It, it’s not as if. It is a completely different driving environment in Vancouver than it is in Toronto. Yes, I know a lot of people who are from Vancouver will say, we’re completely different. And look at our highways. We’re surrounded by the mountains. There’s more snow. And, but I’m like, I’ve driven in both. It’s not that different. Um, you know, the cities are very different, um, and each person has their favorite city, um, in, in, in the different provinces, but, um. Materially, the driving is not that different. What we’re talking about here is a difference in regulatory approach, and that is going to either slow or accelerate adoption.

Grayson Brulte: You’re right about the driving environment. I’ve done both and so very similar, but the regulatory is a lot different and I wanna hone in here in Toronto before we, we get back to Vancouver because in January, Waymo Co-CEO, Tekedra she met with council staff members. And this is an exact quote from the lobbying record that I’ll make and cover I wanna read to discuss the potential development of bylaws and regulations related to autonomous driving technologies, including as it relates to. Operational authorizations, commercial authorizations, goods, delivery operations, and ride hail operations. Okay. We understand the, the ride hail, the commercialization goods delivery. That was that, that stood out to me. Is that a potential opening that Waymo could explore that, or is that, can we get everything under one regulatory environment? So if we, if we decide to expand once again into goods, it’s already in place.

Rob Grant: I mean, my, my guess and based upon, you know, having had those discussions up there, you wanna go as broad as possible, uh, to cover as much because you don’t want to have to keep coming back. Uh, every time you come back, that opens up further discussion from those that are opposed to what you’re already doing. Um, and so look, they may very well be trying anyway to get a foothold in there. It could start with goods delivery, right? Um, it could start with, uh, ride hailing. It could start with, uh, you know, ride hailing, um, in areas that are. Outside of the financial district right up by where, you know, uh, York and, uh, the University of Toronto are, which is a beautiful campus by the way. People should go visit it. It’s got a great view downtown. Um, but you know, I think if I am Waymo, I go in and I’m saying, Hey look, this is this expansive. Of an ask as I can get, and we’ll see what gains traction. Uh, and it gives you an opportunity to say, well, maybe we’ll start here and then move to there. Uh, because you want to provide as much flexibility to the city of Toronto as possible, uh, to allow you to go in, in a way that works for them. Um, because ultimately my experience up in, in Toronto, uh, with Lyft was. If you’re willing to work with them to find a way in, you’ll eventually get in. Um, and so it’s a little bit different of a, of a place. You can’t, I don’t find it a place to be in and, and just say, we’re coming deal with it. Um, if they’re willing to work with you. So they should be willing to work with, with, uh, the city as well. And so I think that’s what I read into that ask is like, Hey, we can work with you along many different fronts, but we really like to get started on just any of them right now.

Grayson Brulte: That’s a positive. And from a regulatory standpoint in Canada, how do the airport negotiations work to do drop off and pick up at the airport?

Rob Grant: That’s, that’s, that’s a, a whole different ball of wax. Right? The airports are regulated differently in Canada than they are here. Um. Plus the, uh, you know, Pearson airport is, is quite a distance away from downtown Toronto. It’s 20 plus miles, uh, outside of it. Um, and then they do a little airport there called Billy Bishop, uh, which I find is the coolest airport in the world to fly into. ’cause you can take a tunnel under the Lake Ontario to get up into downtown, uh, Toronto. Um, I think, I think airports is a stretch, uh, right now to even look at airports there. Um, you know, it, like I said, you’ve got a, a different kind of regulatory body for the airports. It brings in federal oversight as well from from Ottawa. Um, and things of that nature. And so I think that’s a complexity that you get to after you’ve shown that you can deliver a service that is safe and reliable in, in downtown, uh, Toronto or, or greater Toronto, Hamilton area, depending on, on where they might be able to operate.

Grayson Brulte: as you clearly laid out, there’s a lot of complexities and a lot of cooks in the kitchen. Let, let’s go back to Vancouver because there was a statement on Wednesday, March 25th from the ministry to the Vancouver Sun. The stated at this time, the ministry is not pursuing. An autonomous vehicle pilot project, and Waymo has not approached the ministry to request one That’s very blunt language. And it, and it’s a, it’s, it’s a clear contrast between the, the filing that we had in Toronto versus Vancouver. So is it your assumption or your guess that if you look at Canada as a whole, Toronto was more likely to come online before Vancouver potentially.

Rob Grant: Yes, I think well in advance, I, I, I would put it in, in the years in advance, you know, at least 12 to 24 months in advance of a Vancouver launch.

Grayson Brulte: That’s really good insight and that insight brings us to OMEGA’s take on this and OMEGA says. The regulatory angle most miss is a jurisdictional gridlock between provincial and federal authorities. British Columbia is deferring to transport Canada’s overarching safety rules, meaning Waymo must secure federal exemptions before provincial lobbying can yield commercial results. Combined with local pushback over congestion and extreme weather highway safety, the capital deployment timeline for Canadian robax operations is structurally delayed. Did you program OMEGA, because it sounds very similar to what you just elegantly described to the audience.

Rob Grant: I think OMEGA and I are having a mind meld at times, but, uh, that’s not surprising since, uh, I’ve tried to dump most of what is good from this brain into OMEGA. And now that may not amount to much, but at least it’s powering OMEGA for now.

Grayson Brulte: It’s Powering OMEGA with, with our proprietary research, our, our brains and publicly available data. This is our new show, autonomy Signals each and every Thursday, Rob Will and I will be here breaking down the signals. Autonomy economy. If you’re interested in learning more about a proprietary, uh, AI powered by OMEGA, please send an email to alpha (at) autnmy.ai That’s alpha (at) autnmy.ai The Future is Bright, The Future is Autonomous, The Future is Autonomy Signals. Rob, great first show sir.

Rob Grant: I loved it. It was a lot of fun. Uh, I think, uh, I can’t wait for the next one and we’ll, we’ll keep on moving. The future is bright.

Grayson Brulte: The future is bright and we’ll keep uncovering signals.