06 Dec 2023

Lenovo unveils new AI-based vehicle computing products in collaboration with NVIDIA

Lenovo has revealed three new automotive intelligent controllers based on the NVIDIA DRIVE Thor Artificial Intelligence (AI) vehicle computing platform

Hong Kong-based PC manufacturer Lenovo has announced what the company is calling its most comprehensive AI capabilities to date, revealing its vehicle computing roadmap that features three new products based on the NVIDIA DRIVE Thor AI vehicle computing platform.

Trio of vehicle computing products

The new automotive intelligent controllers are the Lenovo XH1: Central compute unit for ADAS and smart cockpit; the AH1: Level 2++ ADAS domain controller unit, and the AD1: Level 4 autonomous driving domain controller unit.

At a recent AI-centred company event, the brand demonstrated an intelligent virtual assistant based on the automotive foundation models inside a vehicle. These models will be applied to Lenovo’s central compute unit, the XH1, and according to the company will allow for the capacity to understand driver and passenger preferences, predict their needs, and respond as if it were a real human assistant. 

“The rapid development of AI, especially foundation models, will accelerate automotive intelligence. In the future, autonomous driving will be safer, and the smart cockpit will be smarter. The vehicle is not just a means of transportation but also an intelligent companion. With automotive AI built on NVIDIA DRIVE Thor, Lenovo can develop vehicle computing controller units to meet customers’ unique needs and requirements,” said Donny Tang, Vice President and head of Vehicle Computing.

Tang added that in addition to these new controllers, Lenovo is tapping its technical strengths and expertise in cloud, edge, AI, devices, and networking to augment its automotive operating systems and intelligent solutions.


Credit: Lenovo

Drive Thor for greater efficiency and safety

This is the latest development stemming from Lenovo’s news earlier this year that the company would be supplying automotive intelligent controllers based on NVIDIA DRIVE.

The next-generation NVIDIA DRIVE Thor system-on-a-chip harnesses advanced AI capabilities the company says were first deployed in NVIDIA Grace CPUs, as well as NVIDIA Hopper and Ada Lovelace architecture-based GPUs, but in a single architecture for greater efficiency and lower overall system cost. Lenovo expects DRIVE Thor to deliver up to 2,000 teraflops of high-performance compute for functionally safe and secure intelligent driving.

“The shift toward software-defined architectures requires a high-performance compute engine that can support the AI workloads necessary for safe vehicle deployment,” said Ali Kani, vice president of automotive, NVIDIA. “DRIVE Thor features the built-in headroom to continually support intelligent driving capabilities, including advanced driver assistance systems, autonomous driving, and smart cockpit.”

Since Lenovo launched its vehicle computing business last year, the firm has looked to establish partnerships with companies such as NVIDIA to strengthen its position within the automotive industry. “By coupling Lenovo’s core strengths in computing technology with NVIDIA’s AI hardware and software expertise, the company is poised to deliver leading-edge intelligent mobility solutions for the future,” the brand said in a release.

AI for All

In addition to the new vehicle computing offerings, Lenovo has also unveiled its vision of “AI for All.” Through this vision the company says it aims to take AI beyond the realm of research and bring real value to users. “From pocket to cloud, individual to enterprise, Lenovo is taking AI out of abstract headlines and putting its power in the hands of real people, everywhere,” the company claimed.

Earlier this year, Lenovo launched its AI at the Edge designed to enable mass deployment of remote computing capabilities and its TruScale for Edge and AI, which offers an infrastructure-as-a-service solution combining Lenovo hardware and software to enable the deployment of machine learning and AI models at the edge.

For more IFA news, click here.

By Bettina Badon