Jan 25, 2019 10:51 PM

Star Researchers’ Departures Marks Shift in China’s AI Sector

Andrew Ng, Baidu Inc.’s then-chief scientist and head of research, poses for a photo in his office in Sunnyvale, California, in March 2016.
Andrew Ng, Baidu Inc.’s then-chief scientist and head of research, poses for a photo in his office in Sunnyvale, California, in March 2016.

When Andrew Ng, an associate professor at Stanford University and an academic star in the world of artificial intelligence (AI), chose to join Baidu Inc. in May 2014, Silicon Valley took notice.

The Chinese company ran China’s dominant search engine, but didn’t have much of a profile in the cutting-edge field of AI.

Ng, who co-created Google’s machine learning system, Google Brain, joined Baidu as its chief scientist and head of research to lead its push into AI. Ng was not the first Chinese AI scientist who made the shift from Silicon Valley to China, but he was perhaps the most well-known. Due to his Google fame, more AI experts with academic and business experience followed him.

“At the time, Andrew Ng’s joining Baidu was a symbol of AI researchers shifting to big tech in China,” a senior executive at an AI company told Caixin.

Later, in June 2016, Baidu founder and CEO Robin Li predicted that AI will become a growth area in China in the following years as the mobile internet waned. “Baidu will make AI its strategic focus over the next 10 years,” he said.

AI scientists were once attracted to these tech giants for their financial resources and business infrastructure, as the success of their research hinges on both, according to multiple researchers who used to work at Baidu. However, the past few years have shown that it remains difficult to turn cutting-edge AI technology into new products. Meanwhile, many of the once-sought-after AI stars have found themselves out in the cold.

Less than a year after Li made his pledge, Ng resigned from Baidu to set up his own venture. By 2018, most of the AI scientists that Baidu recruited had departed, some for their own startups, some for different places to continue trying to apply AI to business.

The transition from academia to the business world hasn’t always gone smoothly. Monetizing academic research remains a challenge, and investors have been losing patience as China’s economic growth has slowed. After Ng’s departure in March 2017, many of his colleagues left for rival Chinese tech giant Tencent Holdings Ltd. One of them, Zhang Tong, who had been a professor at Rutgers University in the U.S., joined Tencent as the new chief of its AI Lab.

A person familiar with the matter told Caixin that Tencent later put Zhang in charge of AI research for its most popular and lucrative mobile game, “Honour of Kings,” but he failed to live up to the company’s expectations.

This month, after 22 months in the position, Zhang stepped down to return to academia. Some saw it as a sign of the times. “The AI industry just doesn’t have the money anymore to wait around for academic research to be developed into a product,” said an AI technician at Tencent who wished to remain anonymous. “Setting up labs allows companies to tout their tech power, and it helps them market their products. But when they run into budget constraints, the first place they cut is R&D (research and development) because research is never a profit center.”

Overall, China’s tech sector has been struggling of late, especially in 2018. Many companies, big and small, have reportedly laid off workers or restructured their operations. And startups have found funding much harder to come by.

Big tech’s new moves after the loss of the scientists show that a more business-oriented strategy is taking over.

After Ng’s departure, Baidu appointed Wang Haifeng, its senior vice president, as the new head of its AI division. Wang believes the industry will live or die based on how fast and how well it can commercialize AI technology.

Tencent’s AI team used to focus more on researching the fundamentals of machine intelligence, but Zhang’s departure may have signaled its shift in focus from scholarship to commercial application. Founded in 2016, the lab had touted its publication of more than 100 papers in top-notch AI conferences and journals in 2017.

“AI research takes a lot of data. We don’t have data, and we don’t have the business know-how either,” said Pan Tianyou, deputy chief of Microsoft Research Asia. “The AI industry has moved fast over the past couple of years, but we must find more business applications for the technology. They are not something that research institutions can find by themselves.”

Microsoft Corp.’s own research facility used to be criticized for disregarding practical needs. Then at the end of 2017, Microsoft Research Asia announced a new project — a member-based partnership with external companies, mostly from traditional business sectors, to collaborate on exploring the digital transformation of enterprises.

Pan said the project charges its clients a high price for its services, but the ultimate goal isn’t making a big profit, but rather to come up with more useful products and services that could benefit more people in the future.

This is the fourth in a series of reports on China’s tech industry this week by Caixin. For previous stories, click:

In-Depth: A Winter Storm Looms for China’s Tech Sector

As Economy Cools, Tech Startups Feel the Pinch

Private Funds Find It Harder to Raise Money as Economy Cools

Contact reporter Isabelle Li (

You've accessed an article available only to subscribers
Share this article
Open WeChat and scan the QR code