Baidu: NVIDIA’s key partner in the AI space
In the previous part of the series, we saw that NVIDIA’s (NVDA) new Volta GPU (graphics processing unit) is the most advanced GPU in the data center space. Advanced Micro Devices (AMD) is unlikely to threaten NVIDIA’s position in the high-performance GPU market.
While focusing on developing advanced products, NVIDIA is also building partnerships to expand the reach of its GPUs in various AI (artificial intelligence) applications. Among its key partners is Chinese (MCHI) Internet giant Baidu (BIDU).
Unlike Google (GOOG), which has been developing its own AI hardware and software, Baidu is relying on NVIDIA for its AI hardware and software needs from cloud computing to in-home entertainment to the autonomous vehicle.
Baidu selected NVIDIA because its products are competitive and deliver performance in several markets. On the other hand, NVIDIA expanded its partnership with Baidu to penetrate the massive Chinese AI market.
- Baidu plans to use NVIDIA’s GPUs for cloud computing, Volta-based Tesla V100 GPUs for training, and Pascal-based Tesla P4 GPUs for inferencing.
- Baidu would use NVIDIA’s Volta GPU to optimize its open-source deep learning framework PaddlePaddle and make it available to academics and researchers. The Internet giant would use Volta’s Tensor core technology to increase the framework’s performance and make it the standard for China’s AI industry.
- Baidu would use NVIDIA’s DRIVE PX2 platform to power its Apollo autonomous vehicle initiative.
- Baidu would also bring its virtual assistant DuerOS to NVIDIA SHIELD TV. This would be an addition to the Google Assistant, which SHIELD TV would support through an update. NVIDIA also plans to launch expandable smart home mics to make Google Assistant accessible throughout the home. These mics would also support DuerOS in the future.
NVIDIA’s partnerships present future growth opportunities
NVIDIA-Baidu partnership is one of the most important partnerships in modern AI computing and could lead to many such broad partnerships in AI. This indicates that NVIDIA’s current partnerships could expand in the future, thereby increasing its revenue growth potential. These partnerships exclude existing and future opportunities from gaming and the traditional chip business.
NVIDIA is not only expanding through partnership but it is also broadening its research base to leverage AI across various applications. We’ll look at NVIDIA’s latest research projects in the next part of this series.