China is one the most important markets in the world of technology. As many auto manufacturers are beginning their autonomous driving projects in the country owing to the huge population base, China also needs to make its mark in other areas of technology such as semiconductors. Experts and analysts have opined that the world’s second largest economy is lagging in the race of development of artificial intelligence (AI). The progress of the country in AI is slower as compared to the U.S. In the recent era of technology, development of AI will be crucial for any country to establish itself as a frontrunner in the world. Computer chips are vital part of any AI system. As China is lagging in the semiconductor industry, it has a lot of potential to become one of the established force in the world by providing hardware for AI systems. The development of chips that can be mounted in AI systems will help the country to progress.
Tsinghua University in Beijing developed a computer chip, known as Thinker, which can perform facial recognition tasks and other tasks by understanding voice commands in Chinese. This chip supports neural networks and uses only eight AA batteries to power itself for a year, thereby saving a lot of energy. Thinker adapts memory and computing requirement according to the needs of software on which it is being run. In AI applications, this chip will be so helpful as it can understand human voice and determine objects in images. The development of Thinker is a major step as existing chips do not use small batteries and are expensive. This small chip can be embedder in watches, smartphones, robots, or equipment employed in remote areas. The Chinese research sector can flaunt Thinker to show the world its achievement.
Chinese Academy of Sciences’ Institute of Computing Technology (ICT) will collaborate with a local semiconductor manufacturer to manufacture small chips for use in robots. The chip, Dadu, will have two cores, one for control motion and other to power neural networks. The core for neural network sets algorithms to allow motion to find an optimal route to reach the destination or determine the optimal motion to lift an object.
China’s AI startups and companies have been developing their own hardware. Fengxiang Ma, director of ASIC design at Horizon Robotics, a Beijing-based startup which researches and develops AI techniques in driving and cameras, said, “In the future, companies that only make chips may be fewer and fewer.” This presents opportunities for companies in China to invest in hardware industry and experience growth in future as number chip manufacturing companies reduce.
However, there needs to be more innovations based on hardware to establish itself in the AI industry. Major U.S. tech giants such as Google and Intel along with few startups have been investing enormously in research and development of hardware for AI. Chinese government has taken major steps to progress in AI sector. China’s Ministry of Industry and Information Technology revealed a three-year action plan for development of AI in December 2017. It set a goal to manufacture neural-network processing chips massively by 2020. The central government also announced a manufacturing policy in 2015, called Made in China 2025. In this policy, chip design and fabrication was crucial sector in which the government sought breakthroughs.
There are challenges for Chinese manufacturers to develop chips such as commercialization, scaling up, and adapt according to transformation in AI sector in future. However, they need to capitalize on opportunities in hardware development for AI.