If Nvidia is looking to make a splash in the world of artificial intelligence it’s going about it in the right way, announcing a partnership with Baidu, one of the world's largest internet companies.

Jamie Davies

July 6, 2017

3 Min Read
Nvidia jumps in at the deep end with Baidu AI partnership

If Nvidia is looking to make a splash in the world of artificial intelligence it’s going about it in the right way, announcing a partnership with Baidu, one of the world’s largest internet companies.

Cracking the Chinese market is an ambition which almost every technology company around the world has harboured over the last couple of years, but this is perhaps one of the more significant steps forward made by a Silicon Valley business. According to the Alexa Internet rankings, Baidu is the most popular website in China, and the fourth most popular worldwide, it’s one of the country’s biggest cloud providers and has solid potential in the consumer electronics market. This is certainly a move which promises some riches for Nvidia.

“We believe AI is the most powerful technology force of our time, with the potential to revolutionize every industry,” said Ian Buck, GM of Nvidia’s Accelerated Computing business unit. “Our collaboration aligns our exceptional technical resources to create AI computing platforms for all developers – from academic research, start-ups creating breakthrough AI applications, and autonomous vehicles.”

The last few years has seen Nvidia focus part of its efforts exclusively on developing GPUs which are optimised for machine learning. It might have been viewed as a bit of a gamble at the beginning, but now it makes sense. The recent machine learning euphoria has been so monumental that companies such as Nvidia who focused on the niche in the early years are reaping the benefits now. Most people might forget that the machine learning craze is a very recent one, but this tie up could be seen as justification for the gamble.

The partnership itself will focus on three different areas; Baidu’s data centres, developing the self-driving car and finally the smart home.

On the data centre side of things, Baidu will deploy Nvidia’s HGX architecture with Tesla Volta V100 and Tesla P4 GPU accelerators for AI training and inference in its data centres. It will also combine its own PaddlePaddle deep learning framework, with Nvidia’s TensorRT deep learning inference software, with the intention of developing products for real-time understanding of images, speech, text and video.

In the self-driving car race, Nvidia’s Drive PX 2 AI supercomputer will form the foundation of Baidu’s self-driving platform. And while this is certainly a win for Nvidia, it is not the first time the pair have collaborated in this space.

Apollo, an open platform for self-driving cars, has seen the pair become chummy before as it includes various bits of Nvidia tech, including Drive PX 2, as well as software including CUDA and TensorRT. Apollo also counts a number of major Chinese automotive brands as members including Changan, Chery Automobile, FAW, and Greatwall Motor. Certainly a useful couple of handshakes to make in what is expected to be the worlds’ largest autonomous vehicles market.

The Apollo initiative was first announced back in April, with the ambition of creating fully autonomous vehicles for highways and regular roads by 2020. 2020 is a statement of intent, but so are Baidu’s recruitment strategies. It seems the internet giant is not happy at simply securing the Chinese market but has more global plans. Baidu has established a development team in Silicon Valley, while also securing autonomous vehicle testing permit from California’s Department of Motor Vehicles.

The final aspect of the partnership will focus on the smart home. Details are relatively thin here, but Baidu DuerOS, the company’s conversational AI system, will provide voice command capabilities to Nvidia’s Shield TV in the Chinese market.

Artificial intelligence promises fortunes for almost every segment in the technology world, but due to the speed and intelligence required to fuel the platform, chip geeks perhaps have more to benefit than most. All the major players are steering towards this revolution, but partnerships like this offer us a little bit of insight into who is winning the race. ARM, Intel and the rest might be edging the PR skirmishes, but Nvidia is quietly going about its work and might be winning the partnerships battle.

You May Also Like