In February 2024, Fudan University’s campus newspaper featured an article on Sun Tianxiang, titled 'Sun Tianxiang: The Idealist with a Bold Vision and Practical Approach'.

Sun considers his primary strength to be his forward-thinking vision, fueled by a boundless sense of adventure. He prefers not to be confined by established boundaries; to conduct groundbreaking research, one must transcend conventional thinking, achieve what others deem impossible, break free from current limitations, and approach problems from novel angles.

On July 1, 2026, the 29-year-old Sun Tianxiang officially joined Baidu's organizational chart as the Head of the Fundamental Model R&D Department and a member of Baidu's Model Committee.

While Alibaba has Lin Junyang (born in 1993), Tencent brought in Yao Shunyu (born in 1998), and Xiaomi has Luo Fuli (born in 1995), it becomes evident when comparing the AI strategies of major tech firms that Baidu previously lacked a young, unifying figure to lead its AI initiatives.

Over the past year, major tech companies have embarked on a collective youth movement, aiming to leave behind the vestiges of the internet era and demonstrate their commitment to the AI age. This has become a prevailing trend.

As of yesterday, Baidu has finally addressed this notable gap.

Sun Tianxiang's credentials are impressive. He earned a Ph.D. in Computer Science from Fudan University, led the development of China's first ChatGPT-like open-source model, MOSS, during his doctoral studies, was an early advocate of the MaaS (Model as a Service) concept, and has published over 40 papers in top-tier conferences. Upon graduation, he became a professor while also launching a startup, securing angel investment from Sequoia Capital and Gaorong Capital. His investors describe him as 'a person with fire in his heart'—a fire that Baidu desperately needs. In recent years, the frequently seen Robin Li has led Baidu, the earliest adopter of AI in China, through a transformative phase. However, when it comes to whether Sun can help Baidu AI make a remarkable comeback, the answer hinges on numerous complex, long-term factors. For Baidu today, Sun at least serves as a short-term catalyst, determining whether the company can restore market confidence. Surprisingly, the buzz around Sun's arrival seems to be overshadowed by the IPO of Baidu's 'own child,' Kunlunxin.

I. Financial Resources?

The extent to which market confidence in Baidu will recover may hinge less on Sun Tianxiang's performance and more on how Baidu nurtures this young talent moving forward.

Baidu has never been short of funds—or the resolve to invest in AI.

According to Baidu's official disclosures, the company's cumulative R&D investment over the past decade has surpassed 180 billion yuan, with over 100 billion yuan invested in AI from the release of Wenxin Yiyan in March 2023 to the end of 2025—all within a span of less than three years.

Baidu's Q1 2026 financial report, released last month, indicates R&D expenses of 5.2 billion yuan, accounting for 16.2% of total revenue, with 279.3 billion yuan in cash and equivalents on hand. Li Yanhong also stated clearly at the earnings call that Baidu will sustain its level of investment in foundational models.

Following Sun Tianxiang's high-profile appointment, a direct question he faces is: How much autonomy will he have in resource allocation?

While Baidu's overall AI investment is substantial, the flexibility to reallocate existing resources is limited.

A significant portion of Baidu's past AI R&D investment has been directed towards Kunlunxin. From its inception in 2011 to its independence in 2021, Kunlunxin has been a continuous drain on resources. After independence, Baidu no longer directly funds its R&D costs but still maintains a controlling stake of over 57%, providing indirect support.

Now, Kunlunxin is nearing its IPO. For Baidu, Kunlunxin represents a core asset capable of directly generating revenue and telling a compelling capital story, making resource allocation to it inevitable.

Next, Baidu must continue to allocate significant resources to Apollo Go.

The Q1 2026 financial report reveals that Apollo Go completed 3.2 million fully driverless orders, up 120% year-on-year, covering 27 cities, achieving break-even in its largest operating city, and expanding into Europe and the Middle East. Apollo Go is Baidu's most mature AI application, a narrative that the capital markets are willing to value, necessitating ongoing investment.

Objectively, the resources Sun Tianxiang can command may already be constrained, not to mention whether this young leader has the conditions to truly secure his due resources within a structure dominated by seasoned veterans. This is where Baidu's sincerity will truly be put to the test.

II. Authority?

In November 2025, Baidu established the Fundamental Model R&D Unit (BMU), led by Wu Tian, and the Applied Model R&D Unit (AMU), led by Jia Lei. Both departments report directly to CEO Li Yanhong.

In May 2026, Baidu announced the formation of the Baidu Model Committee (BMC), comprised of young researchers with a deep understanding of large models. BMU and AMU now report to BMC.

Sun Tianxiang holds the positions of Head of the Fundamental Model R&D Department and member of the Baidu Model Committee, meaning he can theoretically report directly to Li Yanhong.

Under the new structure, Baidu CTO Wang Haifeng's responsibilities have been significantly diminished. His purview has been reduced to supportive sectors like technical platforms, PaddlePaddle, and research institutes, no longer directly overseeing core large model R&D.

This arrangement is intriguing.

Wang Haifeng is no ordinary figure. He was the first Chinese president of ACL, a three-time candidate for the Chinese Academy of Engineering, and led the development of Baidu's knowledge graph, PaddlePaddle platform, and Baidu Brain. More critically, he has been with Baidu for sixteen years, a seasoned technical veteran.

Yet, the gray-haired Wang Haifeng may no longer fit the market's current image of a tech giant's top model leader. Whatever the reason, the result is that Wang has failed to propel Baidu to prominence in the large model era, even giving the impression of complacency—Baidu has all the necessary AI components, but each seems slightly lacking.

In foundational models, Li Yanhong has repeatedly stated that the speed of technological iteration is the only competitive advantage. Wang's R&D system retains a strong research institute ethos, emphasizing papers, parameters, and awards over product implementation and market feedback.

Perhaps for this reason, in applications, Wenxin Yiyan continues to be criticized for starting early but arriving late, struggling to gain traction among consumers (C-end), and while it showcases a steady stream of bidding wins in the enterprise sector (B-end), concrete commercial success remains elusive.

In the large model era, technological iteration occurs weekly, and the efficiency losses caused by Baidu's longstanding internal silos may have motivated Li Yanhong's restructuring. Consolidating core model departments under his direct control aims to shorten decision-making chains. Theoretically, nothing is more efficient than direct CEO approval. However, historical experience suggests that direct reporting to Robin at Baidu may be a double-edged sword.

Many have held this authority before, but few have ended well. Lu Qi is the most representative example—a key figure in Baidu's All in AI strategy, with Li Yanhong once stating, 'Lu Qi knows more about AI than I do.' To some extent, Lu Qi's joining Baidu as COO in 2017 marked Li Yanhong's first true delegation of power, with all business lines reporting to Lu, making him the de facto second-in-command. At the time, Lu disbanded the medical business unit, integrated autonomous driving, and went All in AI, boosting Baidu's market value by 60% in just 486 days.

But ultimately, Lu Qi clashed with the interests of longtime Baidu employees. Media reports suggest Lu advocated eliminating certain vertical industry's bid-ranking ads, facing joint resistance from four Baidu executives. Lu's departure also indirectly revealed Li Yanhong's principles in handling conflicts between old and new forces. After Lu left, Baidu's atmosphere of prosperity seemed to revert overnight.

At the time of Lu Qi's departure, media statistics showed that from 2007 to 2017, Baidu had seen at least ten vice presidents and over twenty executives leave. After Lu, Baidu's senior management remained turbulent, with Li Jing, Xiang Hailong, Yu Zhengjun, Ma Jie, Li Zhenyu, Jing Kun, and others departing in succession.

Li Yanhong's alternating delegation and reclamation of power have created this long list of departures and personnel fluctuations.

Li Yanhong once responded to questions about delegation at a media briefing, saying, 'My management style is generally quite decentralized. If my subordinate and I disagree, we proceed with their approach first. If it works out, great; if not, we try my way.' What Sun Tianxiang faces in the future may still be a delicate power dynamic. Can he lead the technical direction of foundational models? Can he effectively manage existing team members? Can he secure unwavering support from Li Yanhong when conflicts arise with other departments? All remain unknown.

More critically, time is of the essence. After more than three years of large model waves, what Baidu has gained and missed may not allow Li Yanhong to give Sun Tianxiang much time for gradual integration.

III. Corporate Culture?

The topic of why Baidu cannot retain talent has been debated repeatedly for nearly a decade.

Many attribute it to the erosion of engineer culture. Baidu initially rose to prominence through search technology, with engineer culture once being its proudest core—simple, direct, and technology-driven. However, as the company grew, PPT culture and factionalism emerged alongside engineer culture.

These derivative cultures have unconsciously become a bucket of water to extinguish flames, also presenting an irreconcilable tension with Sun Tianxiang's personal aura.

In February 2024, Fudan University’s campus newspaper featured an article on Sun Tianxiang, titled 'Sun Tianxiang: The Idealist with a Bold Vision and Practical Approach'.

Sun considers his primary strength to be his forward-thinking vision, fueled by a boundless sense of adventure. He prefers not to be confined by established boundaries; to conduct groundbreaking research, one must transcend conventional thinking, achieve what others deem impossible, break free from current limitations, and approach problems from novel angles.

From 2019 to 2023, during his Ph.D. at Fudan, his greatest takeaway was deep gratitude for the resources provided by the lab and his advisor's full trust, allowing him to pursue his interests.

He used a metaphor: 'From undergraduate to Ph.D., it was like moving from a main road into a wilderness—I could sprint in any direction. For research, nothing mattered more to me than that freedom.' Today, this bold idealist has officially joined Baidu, attempting to reignite the company's former flame.