Tesla aimed to surprise investors Thursday night with its long-awaited “robotaxis announcement,” marking a milestone in a decade of Elon Musk's unfulfilled promises to offer self-driving cars. There is a possibility that
The automaker has unveiled a prototype called a “CyberCab” rather than a road-ready driverless taxi. Musk said the vehicle would be capable of “supervised, fully autonomous driving.”
“There's no steering wheel, there's no pedals, so we're hoping this will work,” the founder said as 20 cars drove unmanned inside at an event at Warner Brother Studios. “Let us find out!
He added that he expects robotaxis to be introduced “by 2027” at a cost of less than $30,000.
Convincing regulators and passengers that the vehicles are safe can be much more difficult and time-consuming. Meanwhile, major competitors such as Alphabet Inc.'s Waymo are expanding their robotaxi fleets, which are already operating in some cities.
Tesla has historically pursued a different technological path than its major self-driving competitors, according to Reuters interviews with more than a dozen executives, consultants and academics. That path could potentially bring higher rewards, but it also comes with higher risks for both the business and passengers, he said. Three former Tesla self-driving car engineers specializing in self-driving technology.
Tesla's strategy relies on “computer vision,” which aims to use cameras like the human eye, and an artificial intelligence technology called end-to-end machine learning that instantly transforms images into driving decisions. It depends only on the combination of
The technology already powers “fully autonomous” driver-assistance features, which, despite its name, cannot operate safely without a human driver. Musk said Tesla is using the same approach to develop fully autonomous robotaxis.
Tesla's competitors, including Waymo, Amazon's Zoox, General Motors' Cruises and a number of Chinese companies, are using the same technology, but to ensure safety and gain regulatory approval for driverless cars. , typically layered with redundant systems and sensors such as radar, lidar, and advanced mapping. vehicle.
Tesla's strategy is simpler and much cheaper, but it has two significant weaknesses, an industry executive, self-driving car expert and a Tesla engineer told Reuters. Without the layered technology used by its peers, Tesla's system would struggle even more with so-called “edge cases,” rare driving scenarios that self-driving systems and human engineers have a hard time predicting.
Another major challenge is that end-to-end AI technology is a “black box,” making it “nearly impossible” to “know what went wrong when it malfunctions and causes an accident.” said a Tesla engineer. The inability to accurately identify such failures makes it difficult to prevent them, he said.
Tesla did not respond to requests for comment about its technology.
Nvidia founder and CEO Jensen Huang used a similar “black box” explanation in an interview, highlighting weaknesses in the end-to-end technology without specifically mentioning Tesla's system. I explained. End-to-end artificial intelligence involves training computers to make decisions directly from raw data, without intermediate steps that require additional engineering or programming.
Nvidia, the world's leading manufacturer of AI computing chips, also uses end-to-end technology in the self-driving systems it develops and plans to sell to automakers. But Nvidia is combining that approach with more traditional computing systems and additional sensors such as radar and lidar, Huang told Reuters.
End-to-end technology usually, but not always, makes the best driving decisions, which is why Nvidia takes a more conservative approach, Huang said. “We have to build the future one step at a time,” he said. “We can't go directly into the future. It's too dangerous.”
robotaxi pivot
Tesla's ability to offer robotaxi has become more important this year as sales and profits have declined amid slowing global demand for electric vehicles and fierce competition from rising Chinese EV makers.
If Tesla can overcome the technical challenges of its self-driving strategy, the rewards could be huge. Competitors like Waymo already have robotaxis on the road, but they operate much more expensive vehicles in smaller, more comprehensively mapped zones.
Tesla aims to sell affordable robotaxis that can drive themselves anywhere.
Mr. Musk has a long history of making bold promises about self-driving cars. In 2016, he predicted that drivers would be able to hail their cars from across the country within two years. Musk predicted in 2019 that Tesla would produce a working robotaxis by 2020.
This week's robotaxi rollout announcement came on April 5, the same day that Reuters exclusively reported that Tesla abandoned plans to build a $25,000 electric car for the masses, informally known as the Model 2. Following the news, Tesla stock initially fell. Mr. Musk replied: Post Later that day, he posted on his social media platform X that “robotaxis will be announced on August 8th,” sparking intense speculation from investors. Tesla later postponed the event until this week.
That April day marked a fundamental shift in Musk's stated priorities. He previously promised to build Tesla into an electric car giant the size of Toyota, a promise that has helped fuel Tesla's soaring stock price and make it the world's most valuable automaker. He has now vowed to monopolize self-driving technology.
Sudden cost-cutting measures, including mass layoffs, followed as Mr. Musk diverted investment from EV manufacturing priorities such as battery development, gigacasting and expanding the automaker's Supercharger network.
The retreat from mass-market EVs has only increased investor pressure on Tesla's self-driving car development. In April, Mr. Musk took a tougher stance on scrutiny, saying anyone who doubts Tesla's ability to “solve autonomy” should not invest in the company.
Nicholas Marsh, a portfolio manager at Tesla investor Purpose Investments, said Musk “has a lot of convincing to do.”
Still, Mersch said Musk's autonomous strategy is a “really bold bet” with potentially immense payoffs, even if it takes Tesla significantly longer to crack the code. At Tesla, “we have to keep in mind the big picture of how much iterative innovation is happening,” he said. “I don't disrespect them.”
data driven
For now, unlike its robotaxi competitors, Tesla only offers semi-autonomous solutions in its “Autopilot” and “Full Self-Driving” capabilities. The naming and marketing of these systems has sparked investigations and lawsuits over whether Tesla overstated its vehicles' self-driving capabilities and put drivers at risk.
According to a study by the National Highway Traffic Safety Administration (NHTSA) released in April, from January 2018 to August 2023, there were 542 accidents, including 14 fatalities, involving Tesla vehicles with Autopilot or FSD engaged. A collision occurred.
But incorporating Autopilot and FSD into mass-produced models gives Tesla a distinct competitive advantage. That means vast amounts of data collected by cameras in millions of vehicles that can be analyzed and used to develop self-driving technology.
Two former Tesla engineers said the company's technology is capable of large-scale data collection because of its relatively low cost compared to smaller competitors like Waymo. One of the engineers said Tesla's high-resolution cameras cost much less than lidar and could eventually make fully self-driving cars affordable for customers.
Lidar uses lasers to create a three-dimensional image of the vehicle's surroundings as it navigates around obstacles.
Speaking to analysts and investors this summer, Musk touted the “exponential” improvements and predicted that Tesla could achieve unsupervised driving “by the end of this year.” I would be shocked if I couldn't do it,” he added.
Sasha Ostojic, a former self-driving car engineer and head of software development at Nvidia, Cruise, and Zoox, believes Tesla will need at least “three years more” to even match the level of self-driving that Waymo is achieving today. He said he believes it will take some time. Ostojic currently advises Palo Alto venture capital firm Playground Global on technology investments.
“I don't believe Tesla will converge toward true 'eyes off, brain off' self-driving on the timeline that Elon Musk has promised,” he said.
with post wires
