저자: DanIves및영상진행자 · 외국인
요약 보기
테슬라 로보택시와 Cybercab은 오스틴 폭우 때도 운행을 계속한 반면 Waymo는 전면 중단됨. 영상은 순수 비전 방식이 라이다보다 우천·분진·공사구간 등에서 확장성이 높다고 주장. Dan Ives는 2027년을 테슬라의 황금기로 보며 Optimus·로보택시 확대를 주요 촉매로 제시. 진행자는 Waymo 대비 우위를 강조하지만 자막상 출처·데이터 검증은 부족함.
다듬은 글
Hello everyone, a few days ago it poured with heavy rain. Other robotaxi companies, including Waymo, completely stopped all autonomous commercial operations, while Tesla did not. That’s right. Not only did Tesla’s Robotaxi not stop, but the Cybercab, the autonomous taxi with no steering wheel or pedals, also continued to operate on the roads.
UBS analyst Dan Ives says that 2027 will be Tesla's golden year, hmm. Let’s take a look later at the famous Tesla test driver Chris, who recently shared why Tesla’s autonomous driving does not use LiDAR. At night, when it comes to avoiding small animals or pets, it is not inferior to other autonomous driving companies; at the very least, it is just as good.
Although Elon Musk recently said that Tesla’s operating hours were extended from 10 PM to 11 PM , it is currently not yet operating 24/7. Some articles have attacked Tesla , claiming that this is precisely when LiDAR is needed. So why are their arguments incorrect? I [weekend] and guess what happened, Waymo shut down completely as they do when there's rain , but Tesla taxis they kept driving.
One says light scatters in rain, snow, and dust. Tesla's AI chief said the weather doesn't bother the cabs, and finally, it goes very naturally. Tesla’s Cybercab continues normal commercial operations. Actually, Elon Musk mentioned before that when LiDAR encounters rain, it will scatter— these raindrops, and also dust—LiDAR will have this scattering phenomenon.
But if you use an autonomous driving pure-vision solution, with the power of Tesla’s AI computing behind it, you can solve these kinds of problems. We have verified this point again with this reality. Previously, it happened that during heavy rain, there was flooding, and Waymo cars drove into the flood.
Waymo stopped all of its highway commercial driving services and taxi services; everyone can look back, these records exist. Now, let’s look at Dan Ives again. Why did he say 2025? Or was it 2027? Sorry, it’s GL. Tesla. All right, Tesla is moving high on delivery beat. One part of this beat that I thought was interesting: the model, the Model Y, 98%of deliveries—is that an issue for this company?
I mean, that's going to continue to really be the growth story for Tesla going forward, but it's really about the European bounce back. I mean, this is a relief for investors. You go...dire what you think Tesla obviously vehicle truck, that they saw strength in CyberTruck over the last year, but the reality is, given the high-end part of the market, it's not going to significantly move the needle from a demand perspective.
I think going forward, as you go into 2027, the question is: do you actually have a new model released? And then for the stock, it all comes down to robotaxis, autonomous, Optimus. I mean, this is essentially a company transitioning from an EV player into what is really a true AI player. And right now, this is sort of that, that sort of time in between where you're seeing the stock investors are waiting.
Optimus out here, rolled out, how it's going to be out—do you have a very clear one? And obviously, we know that Elon, he often sets a date and then pushes it back, sometimes once, sometimes several times. Yeah, I think 2027, I mean, call it early to middle part of the year. The big thing is going to be about the robotaxi launches because, so far, it's been slow regulatory.
As you have the autonomous story take hold, I think Optimus is the second half of the year going into next year, and that's why I think 2027 will be a golden year for Tesla. And we continue to believe, you know, by the end of the year next year, there's over an 80%chance that SpaceX and Tesla ultimately walk down the aisle, a little marriage there.
There are some other stories, but very quickly, when you're...I mean, this is about scale, because as you go to more and more cities, it's that. Then when you actually get to the Cybercab, when does this actually take the story to the next? I think that's. Ambassadors are waiting for I think right now this is sort of the appetizer to which really going to be a key year for in 2027 but in terms He believes the first half of the year will mainly see more and more Robotaxis entering urban roads for operations, while in the second half, he expects there will be a lot of mass production of vehicles.
And he used an adjective to say that, crucially, the most important thing is for Robotaxis to scale up. Right now, this is just an appetizer, a side dish—like in Western dining where you are served appetizers or small bites before the main course. It is that kind of analogy for an opening act. This reporter, and in fact NBC has never been friendly to Tesla, didn't mention that Tesla's deliveries far exceeded market expectations; instead, they mentioned Tesla's deliveries were 99.
Almost all of which were Model 3 and Model Y. The Cybertruck was originally expected by the market to have very high potential, but now that's gone. They mentioned that this is because it was never meant to be a high-volume vehicle; it's a high-end product, so it wouldn't account for a large share, which is quite normal.
However, I think there is one point he mentioned that I need to correct. He said that the reason Tesla hasn't scaled up is due to regulation. I think the main reason is what Musk explained earlier this year, and what his VP of AI and the SOC also mentioned . The main reason is that their focus was previously on safety, but after solving the safety issues, Tesla's Robotaxi team suddenly realized that operational problems like parking and navigation were surprisingly troublesome.
Musk and his team expressed their realization earlier this year that they didn't expect an operational issue to take so much time. Musk initially said Tesla's FSD V 12.3 was the final piece of the puzzle, but later he said we have to wait for V13. Compared to V12, FSD V13 has 10 times the parameters.
We have recently seen that navigation and parking have gradually improved, and many car owners have verified that the effect is much better when using their own FSD. Regarding parking, I see there are still some issues with navigation, and some netizens are reporting these problems, but Tesla has responded that they are in the process of solving them .
Well, I think this is what is hindering Tesla from rapidly scaling up their Robotaxis as Musk expected by the end of 2025, but this day is getting closer and closer, just with a slight delay as a minor interlude. Next, regarding the delay of Tesla’s Full Self-Driving event from 10:00 to 11:00, while this was originally very good news, some media outlets that are consistently unfriendly to Tesla attacked the company, claiming that LiDAR solutions can handle it, but Tesla cannot.
Here, Chris, a well-known Tesla driver, offered an excellent explanation on the matter. Generally speaking, Google’s Waymo also had an incident where it hit a small animal just last year, even though it was quite a common occurrence. However, in his own driving experience, he uses FSD for 99%of the time.
Near his home, there are squirrels, cats, and dogs, and the car is able to avoid them all. As Musk suggested, it is becoming safer . As for whether you use vision-based technology or LiDAR, you will always encounter small animals that require more careful avoidance, especially since the smaller the animal, the higher the risk.
LiDAR will provide feedback when it encounters an obstacle , but the nature of LiDAR technology is that it only knows something is there, not what it is; deciding what avoidance measures to take still depends on intelligence. So, Tesla’s pure vision approach can actually solve this problem better through greater data accumulation.
In fact, it can already dodge squirrels; it’s just a question of whether we are talking about 99.5%or 99.995%reliability. This also includes Tesla Navigation, which many people find works well, and parking is generally handled very well, but Tesla just wants to solve these issues as effectively as possible.
And as it improves, the results will be even better. This will better reduce the need for remote operations. For instance, Waymo often uses remote operators; when a vehicle encounters a situation and gets stuck, it hits the help button, and a remote operator comes in to help maneuver the car out of a jam.
This kind of situation is actually a significant waste of human and physical resources. Each Tesla autonomous vehicle has a full AI chip inside. Tesla hopes to minimize human and remote operational costs and use a technical solution to achieve this, ultimately requiring less human intervention, which is what will make them more sustainable.
Even now, in Austin, Tesla saw its Robotaxi density surpass Waymo's just a couple of days ago. To put it another way, for every 1,000 vehicles, there were over two Tesla Cybercabs, while Waymo's numbers were lower during that same period. So, what exactly happened to lead to this? Let’s look at what Chris shared.
He discussed Tesla's bad weather capabilities, which explains why we saw more Cybercabs in Austin compared to Waymo during that brief rainstorm. A post showed a Tesla driving 65 mph in heavy rain, and the response highlighted their major focus on traction control. I’ve personally never had issues with my Tesla self-driving in these conditions; the performance in rain is phenomenal.
If you ever pull the camera feeds directly from the car and compare them to what you see out the window, you’ll find the cameras have a much easier time seeing through rain than you do. If you're driving at night in the rain and can't see the lane lines, save your footage and check it later—you’ll be amazed at how clear the lines are compared to your own vision.
Chris mentioned that the dynamic range of the camera sensors is underappreciated, especially when dealing with blinding reflections. He noted that even when it’s too bright for human eyes, FSD handles it without a problem. He also explained that FSD locks into certain modes during inclement weather to prevent hydroplaning, proving Tesla's deep focus on road friction.
This is where Tesla has a huge advantage because its autonomy software is deeply integrated with the vehicle's hardware controls. Tesla makes every part of the vehicle, which other companies can't claim. This is vital because a Cybercab must control torque, braking, and steering in every condition. Dallas is a great testbed for this due to its severe rainstorms, unlike California.
Chris explains that Tesla's cameras possess a level of recognition and resolution for bad weather that far exceeds human capability. In other words, Chris often reviews Tesla’s driving footage and finds that even in extremely strong sunlight that blinds human eyes, Tesla's cameras can still accurately identify the road and lane lines.
It’s the same with rain. Because Tesla seamlessly integrates its intelligent camera hardware, onboard chips, and neural network software, it achieves results that other companies simply cannot match. This is exactly why, when heavy rain hit Austin, Waymo was forced to stop while Tesla continued to operate.
This is also why Tesla's density surpassed Waymo's at that time, because Waymo’s entire commercial operation was forced to shut down under those heavy rain conditions. When we look toward a future where we rely heavily on autonomous vehicles, we can't have a system that becomes unusable just because it starts to rain , right?
Of course, we know that when it rains heavily, many human drivers might choose not to go out, but there will always be people who need to drive or have errands to run. In those situations, we must have vehicles available. Especially since we are thinking about a future where most people may not drive at all, Tesla is truly the one solving this problem, not Waymo.
There are many limitations to vehicles using LiDAR technology. We say there are issues when it rains, and the problems are just as significant when it snows. In dusty areas, the light scatters, and furthermore, if the vehicle hasn't been on a road before— wherever there is no high-definition map, it cannot go.
If a road is under construction or flooded, its systems won't be able to recognize it. Therefore, this is why the LiDAR approach is not a solution that can scale up, whereas Tesla’s is a scalable solution—an approach that can be expanded. Alright, that is all for today. Thank you for your valuable time.
If you enjoy my content, please like, comment, and follow me to encourage me to make even better shows. Thank you.
Comments 0