Towards the Thousand-T Computing Power Era, the Most Powerful Intelligent Driving Chip Thor is on the eve of mass production

Tang Xinyue estimates that in the next five years, the deployment scale of Robotaxi will reach 100,000 units.

In addition to the AD1 product line for Robotaxi/unmanned driving, Lenovo Automotive Computing is also developing AH1, a domain controller based on other versions of Thor. AH1 is aimed at the mass production demand of L2+ for passenger cars, and it is expected to be gradually fixed in the second half of this year and the first half of next year.

On another front, Desay SV helped Xpeng mass-produce the industry's earliest domain controller based on NVIDIA DRIVE Xavier through a tripartite collaboration with Xpeng and NVIDIA starting from 2017-2018.

After accumulating experience with the Xavier generation of products, Desay has almost monopolized the designated DRIVE Orin domain controllers of most domestic OEMs in the past two years.

At the Desay Technology Festival not long ago, Desay executives disclosed that it has obtained orders for Thor from at least two OEMs and had completed the development of the A sample based on Thor in June.

The upstream arms dealers are gearing up for battle, and the downstream intelligent driving competition is about to start.

According to HiEV, Ideal Auto may launch models equipped with Thor as early as the first half of next year;

Zeekr Motors has officially announced that it will release a mid-to-large flagship SUV equipped with Thor in the third quarter of next year;

Other authorized manufacturers, such as BYD, have also obtained Thor samples, and the IDC Intelligent Driving R&D Center under the New Technology Institute is leading the development of domain controllers.

At the same time, unlike the OrinX generation of products, the OEMs are more deeply involved in self-research. Not only have Ideal and Xiaopeng begun to deeply develop domain controllers, but manufacturers such as BYD, Zeekr, and Leapmotor are also accelerating their self-research of controllers.

In addition to providing purely open computing hardware, NVIDIA's internal team, led by Wu Xinzhou, is also developing a full-stack intelligent driving software solution based on the DRIVE platform. HiEV learned that Desay has also formed a team to develop intelligent driving software.

Whether it is open collaboration or the integration of hardware and software, the technical competition in intelligent driving is far from over before the most powerful intelligent driving chip is mass-produced.