Tesla on Monday unveiled computer hardware for "full self-driving" capabilities as part of its strategy to bring autonomous cars to the mainstream.
At an investor presentation, the electric carmaker revealed technical details of a new chip and computer being installed in all its vehicles to advance its goals for autonomous driving.
Tesla already enables partial autonomy for its cars, but the road to full autonomy faces considerable legal and regulatory hurdles.
The chip announcement comes as Tesla races with Waymo, Uber and traditional automakers to bring autonomous vehicles to market.
Chief executive Elon Musk said the new custom-designed chip, which he called the best available, was a significant milestone in self-driving.
"At first it seems improbable — how could it be that Tesla, which has never designed a chip before, could design the best chip in the world?" he said.
"But that is objectively what has occurred."
Musk and other Tesla executives at the event in Palo Alto, California said the new chip was being installed in all its vehicles, clearing the way to improve its software and "neural networks" that will effectively drive the autonomous vehicles.
"All Tesla cars being produced have all the hardware necessary… for full self driving," Musk told the event.
"All you need to do is improve the software."
Despite Tesla's claims, its vehicles are not deploying the standard definition of "level 4" autonomy that would handle all functions with a human on standby or "level 5" autonomy that would need no human.
Musk said he expected that with Tesla technology, drivers "would not need to touch the wheel" sometime early next year and that he hoped for regulatory approvals in some areas later in 2020.
He maintained that Tesla's approach using data from its vehicles on the road was better than those of its rivals which rely on simulations.
"We have quite good simulation too, but it does not capture the long tail of things that happen in the real world," Musk said in response to a question.
"The real world is really weird and messy."
Musk also argued that the neural network technology, based on artificial intelligence, was better than the Lidar systems based on light pulses used by rivals.
Musk said his space exploration firm SpaceX uses Lidar for some purposes but that "in cars it's pretty stupid, it's expensive and unnecessary."
Tesla probes car fire in Shanghai
Shanghai (AFP) April 22, 2019 –
US electric auto maker Tesla said Monday it had launched an investigation after a video circulating in China showed one of its cars suddenly burst into flames in a garage in Shanghai.
The security camera footage has become on the hottest topics on China's Twitter-like Weibo platform since it was posted on Sunday, now viewed 22 million times with the hashtag "Tesla self-ignites".
The video shows white smoke emerging from under what appears to be a Tesla Model S car in a parking garage before flames erupted around it. Other footage shows that the fire destroyed an Audi next to it.
"We're actively contacting and cooperating with relevant departments to learn more about the incident. Based on current information, there were no casualties," Tesla wrote on its official Weibo account.
It is the latest incident to hit Tesla this year, just as it is stepping up its presence in China.
In January, Tesla recalled over 14,000 Model S cars in China as part of the global automotive sector's effort to replace potentially dangerous airbags made by Japan's Takata.
In March, a labelling hiccup made Shanghai customs temporarily suspend clearance for a batch of Tesla's cars.
The company is building its first overseas car factory in Shanghai to ramp up annual production, tapping into the world's biggest auto market and avoiding some of the impact of Sino-US trade tensions.
