테슬라의 로보택시 야망: 도전 과제와 향후 전망

### 테슬라의 로봇택시 야망: 과대 광고, 도전, 그리고 경쟁
* 🚗 2015년 엘론 머스크는 **테슬라 차량이 2년 안에 자율주행 기능**을 달성할 것이라고 대담하게 주장했지만, 그 후로 기한을 지키지 못해 회의론에 부딪혔습니다.
* 📈 머스크는 최근 **올해 말까지 완전한 자율주행 기능**을 구현할 수 있다는 자신감과 함께 **내년에 100만 대의 로보택시를 도로에 투입**할 것이라는 전망을 내놓기도 했습니다.
* 🤖 곧 공개될 테슬라의 전용 로봇택시 공개에 대한 기대감에도 불구하고, 여전히 **카메라를 주로 사용하는 비전 기반 접근 방식**에 크게 의존하는 테슬라의 계획과 현재 기술에는 많은 불확실성이 남아 있습니다.
* ⚠ 소비자 보고서에 따르면 **Tesla의 완전 자율 주행(FSD) 소프트웨어는 기술적 진보**를 보여주지만, 운전 중 많은 운전자가 개입해야 하는 등 **진정한 자율 주행을 달성하기에는 아직 부족**하다는 점이 강조되었습니다.
* 💰 테슬라의 전기차 판매가 감소하는 가운데, 머스크는 **"테슬라의 가치는 압도적으로 자율성"**이라고 주장하며 회의론자들에게 **"이 비전을 의심한다면 주식을 팔라"**고 촉구하면서 회사의 미래를 위해 AI와 자율성의 중요성을 강조합니다.
* 🌍 테슬라가 대규모의 로봇택시를 만드는 것을 목표로 하고 있지만, 웨이모와 같은 다른 회사들은 이미 시장에서 상당한 진전을 이루었으며, **웨이모는 8월 현재 매주 10만 건 이상의 유료 차량 서비스를 제공**하고 있는 것으로 알려졌습니다.
* 🛠 테슬라의 로보택시 비전을 실현하는 길에는 **광범위한 규제 승인**과 **자율 차량 호출의 복잡성**을 처리할 수 있는 **강력한 운영 에코시스템 개발**의 필요성 등 여러 과제가 산적해 있습니다.
* 📅 분석가들은 **테슬라의 로봇택시가 2025년 중반에는 운행될 수 있겠지만 상당한 장애물이 남아** 있으며, 많은 사람들이 이러한 발표를 둘러싼 흥분이 즉각적인 현실로 이어지지는 않을 것이라고 추측합니다.

테슬라의 야심찬 로보택시 계획은 상당한 기술 및 규제 장애물에 직면해 있어 적시에 실현될 수 있을지 의문이 제기되고 있습니다.

In 2015, Elon Musk made a bold promise that Tesla vehicles will drive themselves in two years. While the concept of self-driving has investors and enthusiasts excited, Musk and Tesla have a history of missed promises. Next year, we'll probably be 90% capable of autopilot, Elon Musk out with a bold promise. Yesterday, he said there would be a million Tesla robocar, robo taxis, I should say, on the road next year. I'm very confident about full self-driving functionality being complete by the end of this year, because I'm literally driving it and I'm the boy who cried FSD. But I mean, I think we'll be better than human by the end of this year. Now, Tesla has announced plans to reveal a new vehicle which will operate as a dedicated robotaxi. Originally in August and now in October. You get one chance to make a first impression. They need to make sure they hit it out of the park.

As the rescheduled event approaches, there are still many mysteries about Tesla's plan for robotaxis. But during Tesla's 2024 Q2 earnings call, Musk reiterated his confidence in realizing its full Self-Driving efforts. Based on the current trend, it seems as though we should get miles between interventions to be high enough that it should be far enough in excess of humans that you could do unsupervised, possibly by the end of this year. I would be shocked if we cannot do it next year. Consumer reports have said Tesla's Full Self-Driving software, or FSD, represents technical achievements, but it has also been criticized and is still not capable of fully autonomous driving. See, I took over because that guy was taking a turn. Came too close. With Tesla's EV sales slowing and profit falling, much of the rhetoric from Musk in recent months emphasizes its AI ambitions, with its robotaxi and autonomy as critical to its future. The value of Tesla overwhelmingly is autonomy.

I recommend anyone who doesn't believe that Tesla will solve vehicle autonomy should not hold Tesla stock. They should sell their Tesla stock. Tesla's upcoming robotaxi event has generated investor interest. In the meantime, competitors have made progress. Tesla's falling behind. They were seen as the pioneer. And now? Now there are other companies that have commercial robotaxis. In one bullish estimate, Raymond James predicts $50 billion in robotaxi bookings by 2030. Try FSD. It is the best consumer product, I think, since the iPhone. Wall Street gets really excited about the idea of Tesla nailing it with both FSD and with the idea of a dedicated robotaxi. Musk has big plans for what Tesla's robotaxi offerings would look like. It's like Airbnb and Uber to some degree. So like there'll be some cars that Tesla owns itself. It's kind of like an Uber fashion. But then for the fleet that is owned by our customers,

It will be like an Airbnb thing. On a global scale. You know, the total potential for ride-hailing is probably well over a trillion, but actually being able to deliver on that is the big challenge and also doing it profitably. This is something that so far nobody has figured out. Tesla's ambitions to create autonomous vehicles started with advanced driver assistance features it began introducing in its vehicles in 2015. The first iteration, autopilot, which was sold as an add-on and comes standard today, combined adaptive cruise control and Autosteer, a lane-centering function. Enhanced autopilot became a paid upgrade to regular autopilot, adding navigate on autopilot, auto lane change, and a slate of other features such as Smart Summon. And then there's full Self-Driving, taking all of these features and incorporating traffic and stop sign control, which allowed Teslas to identify stop signs and traffic lights and automatically stop as needed. In a recent update, the brand name was updated to full

Self-Driving, supervised. For about a decade now, Elon Musk has been promising that the company was going to turn existing EVs made by Tesla into autonomous vehicles with a software update. Instead, their vehicles with advanced driver assistance systems. It's not the same thing as a so-called robotaxi. Recently, the company has expanded its scope for autonomy to also include a new vehicle that will operate as a dedicated robotaxi. Elon Musk recently referred to it as a cyber cab, so there's some excitement and expectation that it might have the aesthetic of the Cybertruck. But the company hasn't even taken the wraps off. To gauge whether Tesla is ready for robotaxis. One must first look at its software capabilities. Tesla says that they're what's known as a level two vehicle, and that's a technical term that's used in regulation and legislation. Level two vehicle means that the driver is there to supervise what's going on with the self-driving features. A level two vehicle is never, ever supposed

to be a robotaxi. Most companies that have developed autonomous vehicles, meaning level three plus hands-free systems, use a wider array of sensors in their vehicles, unlike competitors. Elon Musk has been adamant about using a vision-based approach to self-driving that relies mostly on cameras. When I started, they had eight cameras plus radar and ultrasonic sensors, and they've gradually eliminated the radar and the ultrasonic sensors, and now they just have the cameras. It's never going to be a sufficiently reliable solution, especially with the relatively low-resolution cameras that they use, which makes it even worse. I think the vision-based approach to autonomy over the long haul has enormous potential. At the current time, however, though, it's pretty clear that highly automated vehicle technologies are going to need more than vision-based sensing alone. While the company has managed to deploy functioning driver-assist features this way, it has shortcomings. Cameras are really good for doing object classification, but because you don't know where the

light source is. If it sees a billboard on the side of a bus with a picture of a person, it'll often identify that as a pedestrian because it doesn't have that depth perception. Whereas if you had a lidar on there, the lidar would show that it is actually a flat surface. So there's not actually a human being. So that's where having the combination of sensors is very valuable. In 2020, Tesla released a beta version of full Self-Driving, allowing customers who purchased the add-on to test the feature through its development. Fsd is supposed to be basically a robotaxi, except that the person behind the steering wheel is expected to act as a safety driver because it's a robotaxi and training its immature technology. CNBC recently took a ride with a few Tesla owners using FSD version 12.3.6. While the drivers acknowledged it doesn't turn their cars into robotaxis, they have seen it improve with recent software updates, but it still struggles with

certain things. This is where I'm going to be prepared to take over because I don't think it can make this turn. This is pretty sharp, even for me. And it decided not to take that turn. And actually, when I do take over, it gives me the opportunity to leave a voice comment so all I can tell why I'm taking over. And then I think they use that as feedback. Nitesh Mehta has had Tesla's Full Self-Driving for two years, but when he first started using it, the shortcomings were apparent. Honestly, it became a chore because it was. It required quite a lot of your attention. It was just harsh and bumpy and it would speed up. It would slow down, it would make unusual sounds. But that changed and that changed with the version 12. All right. So we are here in Mountain View. Bibhrajit Halder, another Tesla fan who works in the autonomous driving industry developing tech for

mining has been using Tesla's driver assist features since 2016. I've been doing autonomy since Autonomy Car started back in 2003, so I have been in the kitchen for over 20 years. I'm excited about what's coming, but I'm also very, very cautious about I know the limitations. Back in 2016, I used to let it drive in the highway and highway. It manages pretty well, but obviously that time I had to really, really monitor very closely and then over the time highway become much more comfortable. I think what the progression I'm seeing that in the city level. Now this is getting better. But city I still take over. There is a lot of area that it doesn't do well in city. Musk hyped the next FSD version as having a 5 to 10x improvement in miles per intervention, but the company also started calling this option FSD supervised and acknowledgment that the driver must stay in control of

their car's steering and braking at all times. According to the most recent data from the National Highway Traffic Safety Administration, Tesla has reported 1,399 collisions in which driver assistance systems were engaged within 30 seconds of the crash, with 31 of these resulting in alleged fatalities. NHTSA began gathering data from automakers on collisions involving advanced driver assistance systems. In June 2021, Tesla reported a higher number of collisions where ADAS was engaged compared to other automakers in the US. One factor is that Tesla started offering these systems as a standard option in all their cars before competitors. The company has made progress on advanced driver assistance systems, but they really haven't come out with anything that's safe to use, hands-free, and that could operate as a level four style robotaxi. Some Tesla owners have actually sued the company because they want money back after paying for FSD. What they expected would upgrade their cars to a true robotaxi experience, and they never got it in the

lifetime of the vehicle. With AI and automation at the forefront of Tesla's priorities, the company reportedly canceled plans for its affordable mass-market vehicle in favor of a dedicated driverless car built off of the same platform. But little is known about Tesla's latest vehicle and whether it can operate as a fully functioning robotaxi based off of its FSD software. I've been following Tesla for 16 years now since they unveiled the original Roadster, and Elon Musk has never set a deadline he couldn't miss. I was not at all surprised that they missed this one. They had to probably go back to the drawing board on something. I wanted to make some important changes that I think would improve the robotaxi. The thing that the main thing that we're going to show, and we're also going to show off a couple of other things. I think the prototype is going to be pretty eye-popping. I think there could be some announcements around partnerships, specifically new executives, you

know that that maybe lead this unit. That would be, I think, front and center for Wall Street. They need to actually now deliver can't delay it anymore. And a year from now in 2025, you actually need to see these things rolled out. As October 10th approaches, details have emerged that Tesla plans to host the robotaxi event at the Warner Brothers Studio backlot in Burbank, California. But despite whatever is showcased that day, the company issued a disclaimer in its 2024 Q2 earnings statement that its robotaxi deployment depends on technological advancement and regulatory approval. You just literally open the Tesla app and summon a car and send a car to pick you up and take you somewhere, and we'll have a fleet that's on the order of 7 million that are capable of autonomy soon. In the years to come, it will be over 10 million, then over 20 million. It's basically instant scale with a software update.

When you think about ridesharing, the likes of Uber and Lyft, I mean, this is Tesla's answer. It creates almost a whole nother mini company in terms of where this could all lead to. Elon Musk has talked about Tesla owners being able to generate income one day by adding their vehicles to its autonomous network. At its Autonomy Investor Day in 2019, Musk predicted one could generate $30,000 in gross income if their car operated as a robotaxi. He says the revenue that Tesla owners will get from making their vehicles part of the robotaxi fleet will far exceed the monthly payment for a Tesla. But even if existing Tesla owners receive the capability to operate their cars as robotaxis through an over-the-air software update, the vehicles are missing several other features necessary for driverless ride-hailing service. If you're going to operate an automated ride-hailing service like what Waymo does, you need a vehicle that has been expressly configured for that use case.

So, you need sensors and cameras on the inside of the cabin to monitor. Okay, has a passenger gotten sick? Have they left something behind? Did they smoke in the car? Tesla does have one camera mounted up by the mirror that is a fisheye camera, but it doesn't give you that much detail. With a human-driven cab or ride-hailing vehicle, somebody gets out of the car and they don't close the door properly. You know, the driver can always hop out and close the door. If you don't have a driver in the vehicle, you need to be able to do all of that stuff automatically. And obviously, Tesla vehicles don't have that capability. The amount of back-end capital involved in supporting robotaxis that exist today is immense. It is far more than developing the software to enable a robotaxi to operate. It is the entire ecosystem of what happens when things go wrong. How do vehicles deal with complex situations where the 5G network is no longer attainable for some

reason? Lots of unknowns that have to be factored in here that Waymo has been struggling with for years. Waymo recently received an additional $5 billion in funding to support its operations and expansion, adding to $4.75 billion in funding the robotaxi unit raised since 2020. You've got the expense of those vehicles, which is not going to be inexpensive. If you're going to do it right, you have to maintain those vehicles, you have to clean them, you have to charge them. So there's a lot of other operational expenses that you suddenly add in. There are also major regulatory hurdles. NBC News reported in April that Tesla hadn't contacted regulators in the states of Arizona, California, and Nevada to apply for the licenses and permits needed to test autonomous vehicles without a safety driver or to run a commercial robotaxi service. It took Waymo eight months to get its initial permit from the California Public Utilities Commission. With version 12.5 and maybe a 12.6, but pretty soon we

will ask regulatory approval of the Tesla supervised FSD in Europe, China, and other countries. I think we're likely to receive that before the end of the year. There's a very fragmented array of permits, licenses, and regulatory standards you have to meet. It is a significant challenge, and we haven't seen a lot of evidence that Tesla is knocking down all those hurdles. Tesla did not respond to CNBC's request for comment. In the race for robotaxis, some competitors have backed away from their driverless ambitions. Ford, Uber, and Apple abandoned plans to develop their own autonomous vehicles, and General Motors' cruise has run into setbacks. It recently said its purpose-built robotaxi, the cruise origin, is indefinitely delayed, with CEO Mary Barra citing regulatory uncertainty around the vehicle's lack of manual controls. But there are some companies pushing ahead. Amazon-owned Zoox is operating its driverless shuttles, giving employees rides, and gearing up for commercial robotaxi services in China.

Players including Pony.ai, Didi, and Baidu are launching driverless operations as well. Alphabet's Waymo, formerly the Google self-driving car project, has been undergoing testing since 2009 and launched an early rider program in the metro Phoenix area. In 2017, it began offering fully driverless rides to the public in October 2020. Since then, it has doubled its operating area in Phoenix. In 2023, launched service in San Francisco and Los Angeles, and has started initial rider-only testing in Austin, with plans to launch later this year. As of August, Waymo surpassed 100,000 paid rides per week. Among the companies operating in the US, they're far and away way ahead of everybody else in terms of getting this commercialized. I think Waymo has had a very successful strategy. Customers have loved it. Waymo tests and launches service in limited geographies, slowly expanding operations as regions are mapped and the vehicle's capabilities improve.

Musk has criticized this approach. Our solution is a generalized solution like what everybody else has. If you see Waymo and whatnot, they have a very localized solution. It's quite fragile. Their ability to expand rapidly is limited. Our solution is a general solution that works anywhere. It would even work on a different Earth. Waymo hasn't been without issues. However, the National Highway Traffic Safety Administration opened an investigation into Waymo Robotaxis, driving behavior, and potentially disobeying traffic safety laws to support its early driverless operations. Waymo relies on remote operators to assist when encountering scenarios that confuse its robotaxis. People say, well, these things won't drive drunk, and they won't drive distracted because they're not texting. But it turns out they have software defects. So the jury is still out on whether robotaxis will be safer than human drivers. In a recent no priors interview, Andrej Karpathy, Tesla's former head of AI, says he believes Tesla is

ahead of Waymo in its self-driving approach. Waymo looks like it's winning kind of right now, but I think when we look in ten years and who's actually at scale and where most of the revenue is coming from, I still think they're ahead in that sense. I don't think that Tesla is going to become a formative competitor to Waymo. I think there is a safety equation here that Waymo understands to a much greater degree. You can't beta test on public roads forever. You need to put a safety bar in place that is above and beyond what the average driver would accept. Even though Waymo is, you know, by far the most advanced right now, um, it's still only three cities in the United States. There's still lots of time for others to come in, you know, and grow the market. By no means is Waymo won. But until more details are revealed, analysts can only speculate about when Tesla's robotaxis will actually

be on the roads. Our expectation by mid-2025, you'll have rides out there. Now, it depends on the city in terms of the regulatory and how Tesla navigates that. But this is not years away now. I mean, we're talking months, and it speaks to my view. This is an autonomous AI story. They're actually going to be the ones that lead what I really almost view as like an FSD autonomous revolution. There will be a lot of hype. And then, you know, we won't see it again for probably at least another three years after that, maybe longer. I would not expect to see, you know, a functioning Tesla robotaxi on the road for at least several more years, regardless of what it looks like or what hardware it has on it. I don't expect the robotaxi day to come and all of a sudden every Tesla to become robotaxi. Not going to happen. These are very marketing events. It meant for creating excitement and the buzz.

I'll also sign up for the robotaxi thing if they, you know, I can use my car as a robotaxi just for the heck of it, right? I am always like the ahead of like early adopter of early adopter. So I will try it just for the experience of it.

Will Tesla Robotaxis Live Up To 10 Years Of Development And Hype?

Welcome to the official YouTube channel for CNBC. From 'Wall Street' to 'Main Street' to award winning original documentaries and Reality TV series, CNBC has you covered. Experience special sneak peeks of your favorite shows, exclusive video and more.

For real-time news updates from around the world, visit: http://www.cnbc.com

For CNBC's primetime TV lineup, visit http://tv.cnbc.com

For CNBC's International YouTube channel, visit: http://www.youtube.com/cnbcinternational

CNBC

테슬라의 야심찬 자율 주행 목표

엘론 머스크는 내년까지 100만 대의 로보택시를 도로에서 운행하겠다는 대담한 주장을 하며 Tesla 차량이 완전한 자율 주행 기능을 달성할 것이라고 여러 차례 약속했습니다.

테슬라가 약속을 지키지 못한 전력이 있음에도 불구하고 머스크는 테슬라의 미래를 위한 로봇택시의 중요성을 강조하며 테슬라의 자율주행 노력에 자신감을 보이고 있습니다.

그러나 경쟁업체들이 로보택시 시장에서 진전을 이루면서 Tesla의 지배력이 도전받고 있으며, Tesla의 야심찬 계획을 이행할 수 있는 능력에 대한 의문이 제기되고 있습니다.

엘론 머스크는 테슬라의 전기차를 자율 주행 차량으로 만들겠다고 약속했지만, 현재는 첨단 운전자 지원 시스템으로 작동합니다.

테슬라는 현재 사이버 택시라고 불리는 전용 로봇택시에 집중하고 있지만, 소프트웨어 기능은 아직 레벨 2에 머물러 있어 운전자의 감독이 필요합니다.

대부분 카메라에 의존하는 테슬라의 비전 기반 자율 주행 방식에는 단점이 있어 고도로 자동화된 차량 기술을 위해 더 다양한 센서가 필요합니다.

테슬라의 운전자 보조 기능 및 로보택시 요금제

Tesla는 2016년부터 운전자 보조 기능을 사용해 왔으며, 고속도로에서 시내 주행으로 자율 주행 기술을 발전시켜 왔습니다.

이러한 개선에도 불구하고 운전자 제어의 필요성과 보고된 충돌 사고 건수가 증가하면서 우려의 목소리가 높아졌습니다.

Tesla는 무인 자동차 및 로보택시 전용 차량에 대한 야심찬 계획을 진행 중이지만, 기술 발전과 규제 당국의 승인에 따라 배치가 달라질 수 있습니다.

Tesla는 자율 네트워크를 통해 미니 회사를 만들어 소유주가 자신의 차량을 로보택시 차량에 추가하여 수입을 창출할 수 있도록 하는 것을 목표로 하고 있습니다.

엘론 머스크는 테슬라 소유주가 자신의 차량을 로보택시로 운영하여 3만 달러의 총수입을 올릴 수 있을 것으로 예상하지만, 차량에는 무인 차량 호출에 필요한 필수 기능이 부족합니다.

로보택시를 운영하려면 상당한 자본과 운영 비용, 규제 당국의 승인이 필요하기 때문에 테슬라의 야심찬 계획에 어려움이 있습니다.

테슬라의 FSD 승인과 로봇택시 레이스

테슬라는 여러 국가에서 감독을 받는 FSD에 대한 규제 당국의 승인을 구하고 있지만, 허가 및 표준이 세분화되어 있어 어려움을 겪고 있습니다.

일부 경쟁업체는 무인 차량 계획을 포기한 반면, Zoox, Pony.

ai, Didi, Baidu와 같은 회사는 무인 차량 운영을 추진하고 있습니다.

미국을 선도하는 웨이모는 성공적으로 사업을 확장하고 상용 서비스를 시작했지만, 로보택시의 운전 행동에 대한 조사와 조사에 직면해 있습니다.

자율 주행 분야에서 웨이모와 테슬라의 경쟁

현재 자율 주행 기술을 선도하고 있는 것은 Waymo이지만, 2025년 중반에는 Tesla가 이를 따라잡고 상당한 영향력을 발휘할 것으로 예상됩니다.

Waymo는 안전에 초점을 맞추고 미국 내 3개 도시에서만 제한적으로 서비스를 제공하고 있기 때문에 다른 회사들이 시장에서 성장할 시간을 벌 수 있습니다.

Tesla의 로봇택시 이니셔티브는 잠재적인 게임 체인저로 여겨지지만, 제대로 작동하는 로봇택시의 광범위한 출시는 아직 몇 년이 더 남았습니다.