- NVIDIA and Baidu announced a partnership to create a cloud-to-car autonomous car platform for self-driving car makers.
- The platform is based on high performance mapping systems in the cloud (Baidu) and in-car components for machine vision, AI and ML (NVIDIA).
- NVIDIA is consolidating its leadership in advanced AI/ML hardware components for autonomous driving.
In August Amigobulls reported that NVIDIA (NSDQ:NVDA) stock had soared 146 percent in the previous year, outperforming all other tech stocks - no other tech stock gained more than 50 percent in the same period. The spectacular performance of NVIDIA's stock is due not only to the company's leading Graphic Processing Unit (GPU) product line, but also due to the fact that NVIDIA is becoming a leading supplier of hardware for Artificial Intelligence (AI), Machine Learning (ML), cloud computing, and autonomous driving.
In November last year Amigobulls covered the launch of the first embedded computer designed for AI applications, a credit-card sized supercomputer for embedded autonomous devices. In January, at CES 2016, NVIDIA unveiled the world's most powerful engine for in-vehicle Artificial Intelligence (AI) for autonomous cars. The NVIDIA DRIVE PX 2, with ten times the processing power of the previous-generation unit announced at CES 2015, and able to process 24 trillion operations per second, could enable cars to sense the world around them and drive themselves. With these product launches, NVIDIA wants to consolidate its position as the leading supplier of embedded AI technology for self-driving cars - a market expected to explode in the next decade. The recently announced partnership with Baidu Baidu (NSDQ:BIDU), the leading Chinese internet operator and a loose equivalent of Alphabet (NSDQ:GOOGL) in China, represents a milestone in NVIDIA's roadmap.
At the recent Baidu World Conference in Beijing, Baidu CEO Robin Li and NVIDIA CEO Jen-Hsun Huang announced a partnership to use artificial intelligence for the creation of a cloud-to-car autonomous car platform for self-driving car makers. The partnership will combine Baidu's cloud platform and mapping technology with NVIDIA's self-driving computing platform to develop solutions for HD maps, Level 3 autonomous vehicle control and automated parking.
"Level 3" is defined, in the classification of the Society of Automotive Engineers as "conditional automation: the driving mode-specific performance by an automated driving system of all aspects of the dynamic driving task, even if a human driver does not respond appropriately to a request to intervene." In other words, a driver should be present in a "Level 3" vehicle but the car can handle most functions on its own, including autonomous parking. It's worth noting that other self-driving car makers such as Ford (Ford (NYSE:F) are already planning future "Level 4" (full automation) vehicles.
"We're going to bring together the technical capabilities and the expertise in AI and the scale of two world-class AI companies to build the self-driving car architecture from end-to-end, from top-to-bottom, from the cloud to the car," said Huang. "We can start applying these capabilities to solve the grand challenges of AI, one of which is intelligent machines. One of the intelligent machines we would like to build in the future is the self-driving car."
Baidu announced that Baidu USA has been issued an autonomous vehicle testing permit by the California Department of Motor Vehicles. Baidu, which started its autonomous driving research program in 2014 and performed successful road tests in China, will now be able to test its self-driving cars on roads in California and demonstrate compliance with the stricter US regulations.
Baidu is planning to use its self-driving technology for a public, shared shuttle service in 2018, said Baidu VP Jing Wang, who leads Baidu's autonomous cars team, as reported by Business Insider. Wang noted that Baidu's autonomous car platform uses a high-performance image recognition subsystem developed by Chinese company Inspur in partnership with NVIDIA. Baidu and NVIDIA will develop a cloud platform for Level 3 autonomous vehicle control, and the platform will be made available to other car makers.
A fully autonomous car is a complex system with many separate subsystems both in the car, such as the in-car supercomputer running machine vision and AI algorithms, and in the cloud, such as 3D maps continuously updated. While Baidu's contribution to the partnership will be more focused on cloud computing and infrastructures, NVIDIA will focus on the in-car components. It's worth noting that the similarities between GPUs and AI accelerators, both of which rely on intensive parallelization - running many elementary operations at the same time - indicate that in-car specialized hardware for autonomous driving is a natural evolution of NVIDIA's traditional product lines.
"No company in the world today has demonstrated the ability to integrate this system and deploy it at scale for millions of cars to drive by themselves," said Huang, as reported by The Wall Street Journal. "We are going to bring together our technical capabilities and expertise in AI” to make this happen."
It's very likely that Baidu's technological and financial strength will permit the Chinese company to conquer a profitable niche in the self-driving car market of the 2020s, at least in its native Chinese market but perhaps, if it can compete on price, also worldwide. This is good news for Baidu investors. But the self-driving car sector is becoming crowded, which is why Baidu is targeting platforms and infrastructures that can be sold to other autonomous car makers. NVIDIA doesn't make cars and owns the sub-niche of specialized advanced hardware components for autonomous driving, and therefore the partnership with Baidu is very good news for NVIDIA investors.