The Application Of Ai To Nvidia Is Only A Difficult Challenge

Artificial intelligence (AI) has begun to show its application potential in various vertical applications, and the trend of moving towards the edge nodes is becoming more and more obvious. For the past more than 10 years, the GPU operation has been vigorously pushed forward, and it is imperative to push the edge operations in the supercomputer, efficient computing, AI and other areas of excellence, but the company will be very strategically focused on some applications.

Jen-hsun Huang, executive director of NVIDIA, thinks that a simple topic is not worth doing, and only a highly difficult topic is worth the investment of the company. For technology companies, pursuing profits is important, but more important is to contribute to the advancement of science and technology.

More than 10 years ago, NVIDIA decided to transform from a graphic chip company into a computing company and began to build up the infrastructure and ecosystem needed to promote the popularization of GPU operations. More than 10 years later, the achievements of the transformation of the company are obvious to all. From the most sophisticated physical and medical research, to the current hottest artificial intelligence and self driving development, the NVIDIA platform continues to provide higher operational efficiency for developers, and has also become an engine for the advancement of technology.

Looking forward to the future, the era of ubiquitous AI will soon come and bring considerable opportunities for the technology industry. With NVIDIA itself as an example, the data center related products have been the second major source of the company, and the Department's latest quarterly revenue is still paying 71% of its annual growth. This is evident from the strong growth momentum of AI demand. The AI revolution has just started. In the future, the training and inference of AI will not only be carried out in the data center, but the mobile phone, the sound box and even the fridge will also support the machine learning (ML) inference.

Jen-hsun Huang pointed out that edge computing will be a very large market, and NVIDIA will not be absent. But the company will carefully choose to cut into the path of the market, and will not do anything. For example, mobile phone, smart sound box or household appliance such as applications, although the future will have machine learning and inference demand, the market size is not small, but the demand for such products is very simple, not worth the NVIDIA investment. Autonomous Machine, such as self driving, intelligent robot and intelligent manufacturing, is a topic of high complexity that NVIDIA wants to challenge.

In fact, all kinds of AI, which seem complex, do the same thing in essence: let the machine have the ability to learn, and then let the software automatically write new software to solve the problem and achieve a higher degree of automation. Therefore, NVIDIA is very interested in all kinds of complex automated applications.

Therefore, Jen-hsun Huang believes that although NVIDIA is giving priority to resources in transportation, medical and other fields, the AI application of manufacturing is also a potential market in the future, and the company will be laid out in the future. Now the automation of manufacturing is mostly rigid, equipment can only be fixed, repetitive work, but if the production equipment into the AI, manufacturing will become more flexible and more efficient.