Shares of Baidu (BIDU -1.78%) tumbled 10% on May 18, after the company announced the upcoming departure of COO Qi Lu in July. Prior to joining Baidu in early 2017, Lu served as Microsoft's executive VP of its Applications and Services Group. As one of the industry's leading AI experts, Lu played a key role in Baidu's transition from an online search company to a cloud and AI services provider.

When Lu initially joined Baidu, CEO Robin Li stated: "With Dr. Lu on board, we are confident that our strategy will be executed smoothly and Baidu will become a world-class technology company and global leader in AI." Lu stated that he was "excited to help realize Baidu's visionary AI strategy."

A female robot's head shattering.

Image source: Getty Images.

Lu will remain vice chairman of Baidu's board but stated that he was "no longer able to work in China on a full-time basis" for "personal and family reasons." His departure will likely hurt Baidu, which is battling Tencent (TCEHY -0.30%) and Alibaba (BABA 2.12%) in the cloud and AI markets.

This also wouldn't be the first time Baidu lost a major leader in its AI efforts. Let's take a look back at Baidu's AI initiatives and analyze the unit's brain drain issues.

Understanding Baidu's AI business

Baidu's "Baidu Brain" unit improves the company's machine-learning algorithms, core search technologies, analytics, and big data applications.

The division created Baidu's virtual assistant Duer and its platform DuerOS, opened labs in Silicon Valley to work with American tech companies, and launched an open-source platform for autonomous cars called Apollo. Robin Li believes that Baidu's self-driving cars, which are powered by its AI and mapping technologies, will hit public roads in China "within three to five years." The unit is also building a state-backed engineering laboratory for deep-learning technologies in China.

Baidu Brain also partners with other companies, like Huawei and Xiaomi, to expand its reach by deploying its APIs into mobile and Internet of Things devices. Those moves widen Baidu's moat against Alibaba -- which leverages its e-commerce and cloud infrastructure platform to expand its AI efforts; and Tencent, which leverages its WeChat messaging app, video games, and streaming services to do the same.

If Baidu holds these rivals at bay, it can gather more data from its users via DuerOS devices, process its search data into better-targeted ads, and expand its AI-powered ecosystem into adjacent markets like streaming media, fintech, and smart retail.

But it's been a bumpy ride

Baidu hired Qi Lu to sharpen those efforts, but Lu's arrival preceded the departures of three other key AI executives: Andrew Ng, Zhang Tong, and Adam Coates.

A brain melting like ice cream on a tablet.

Image source: Getty Images.

Andrew Ng, who became Baidu's chief scientist in 2014, resigned in March 2017 to pursue other opportunities. Prior to joining Baidu, Ng was an associate professor at Stanford University and a founding leader of Alphabet's Google Brain Team.

Just two days after Ng's resignation, Tencent poached Zhang Tong -- who previously led Baidu's Big Data Lab -- to lead its own AI Lab. At the time, Tencent's AI team of 50 scientists and 200 engineers was much smaller than Baidu's team of 1,300 AI employees.

In September, Adam Coates, the director of Baidu's Silicon Valley AI Lab, also resigned. Coates was responsible for promoting Baidu's machine-learning efforts in the U.S. and applying deep learning to natural language processing tasks like text-to-speech and speech recognition. Coates notably did his post-doctoral research at Stanford, under Ng.

Is Baidu losing its edge in the AI market?

Baidu's loss of four key AI executives in just over a year is troubling, and it indicates that the company could be struggling with management issues as it faces tougher competition from Tencent and Alibaba.

Lu's successor as AI chief, Haifeng Wang, has been with Baidu since 2010. However, it's unknown if Wang will fare better than his predecessors, who all seemed to jump ship just as the market was getting more interesting.