For years China has watched with envy as the West developed one frontier technology after another, while it could only count on low-cost labor to fuel growth.
Now a fundamental shift is underway in one of the hottest fields of technological innovation: Artificial Intelligence (AI). China is no longer simply catching up with the U.S., it is now taking the lead in some fields of AI, experts say.
The country's AI advances stem from significant state support coupled with an increasingly vibrant private sector. Beijing, interested in using the technology for wide-ranging purposes including surveillance, predicting crime, transforming city services and boosting national security, is directing massive capital and policy support to AI. It wants to be the world leader in this field by 2030, aiming to make the industry worth $150 billion by then.
Last week, a state-backed fund led a $460 million investment in Beijing-based facial recognition startup Megvii, eclipsing the previous record held by SenseTime Group, a Chinese facial recognition company that raised $410 million in July -- a sum it said then was the largest single round for an AI company to date.
More is on the way. While the Trump administration has proposed to slash the budget of the National Science Foundation -- a major hub of scientific research and innovation in the U.S. -- China's Ministry of Finance has plans to put some $1 billion into AI-related projects this year. The support includes funding for state-owned enterprises to develop automated oil production lines, design digitalized factories and make faster chips to crunch more data.
And not to be outdone, China's Ministry of Science is readying new funds totalling tens of billions of yuan for a new AI project, said Zha Hongbin, a professor of machine intelligence at China's Peking University. In the long run, the country's capital support for AI will be no less than major projects such as HeGaoJi, an initiative unveiled in 2009 that will receive more than 100 billion yuan ($15 billion) in funding by 2020 to develop technologies including better chips and computing software, said Wu Fei, a professor at Zhejiang University's College of Computer Science.
"China is devoting a lot of strength, a lot of determination and a lot of money to AI," Wu said. "Science budgets are growing by leaps and bounds in recent years, and the investment and support are indeed much better than abroad."
Although heavy state involvement is by no means certain to guarantee breakthroughs, China also has a dynamic private sector to back it up. Tech giants Alibaba, Baidu and Tencent now rival the likes of Google and Facebook in market influence, and they're racing to develop AI. The companies even setting up labs in California's Silicon Valley to write better algorithms, produce smarter personal assistants and testing self-driving cars.
Last month, Baidu opened its second Silicon Valley research lab, tapping local talent to achieve the goal of putting fully autonomous vehicles on Chinese roads by 2020. In the same month, Alibaba announced it would spend $15 billion over the next three years to recruit top talent for projects related to AI, cloud and quantum computing.
And in perhaps a direct challenge at Google's AlphaGo program, Tencent also developed a program on the complex board game Go. The company's Jueyi, which means fine art, won all 11 games at the Computer Go UEC Cup in Japan in March.
These Chinese companies have an advantage over rivals in the West -- unfettered by privacy regulations and debates, they are relatively free to scoop up gigantic amounts of data on consumption, payment, healthcare and transportation from China's estimated 750 million Internet users to perfect their algorithms and test new products. They can tap into China's national data trove of personal IDs, cross-referencing photos with their own data sets to develop facial recognition technologies, although some uses, for example naming and shaming jaywalkers, is sure to alarm some.
"In some cases, particularly in image recognition, China has done really well," said Paul Triolo, practice head of geo-technology at New York-based consultancy Eurasia Group. "These are now in the same realm as leading western companies."
In the meantime, however, the U.S. has looked at China's rapid advances with growing unease.
While much of China's AI research is peaceful and has genuinely done good -- including connecting a lost child with his parents 27 years after his abduction via facial recognition -- the Trump administration is increasingly concerned that Beijing might eventually use AI technology for military advances. For example, the People's Liberation Army has capitalized on AI to build unmanned weapon systems, including aircraft, drones and underwater vehicles, with the long-term goal of moving towards "unmanned, intangible,silent warfare," according to a February testimony in front of the U.S.-China Economic and Security Review Commission.
With these concerns in mind, the U.S. is weighing restrictions on Chinese investment in Silicon Valley, looking at in particular AI and machine learning, to shield cutting edge technologies with national security implications from Beijing. It comes as more and more Chinese capital is backing early stage U.S. AI startups, whose technologies are licensed by the military.
For example, China's Haiyin Capital, a government-linked fund, invested last year $1.2 million in Boston-based Neurala, an AI startup which has licensed its mapping and navigation software to NASA and the U.S. Airforce. Neurala has said that the company won't give core source codes to its Chinese investors.
Meanwhile at home, China's big AI push isn't all good news. State support has given rise to a potential bubble in big data analytics, as funding is more likely to be given to projects yielding immediate results -- hence the rush to write algorithms to parse various data sets, Peking University's Zha said.
That means the U.S. still leads when it comes to the core of AI, and that is teaching machines to behave like sentient beings. The West for one thing has more talent related not just to AI but other cognitive subjects critical to building intelligent machines.
"I don't think we can be overly optimistic," Zha said. "Still, China is catching up fast. With the government's help, we are seeing some very good results now."