The most striking change at the 2026 Yunqi Conference is the absence of Zhou Jingren from the limelight. For nearly three years, he was a pivotal guest speaker at the main forum, making major AI business announcements. This year, however, his name is notably missing from the main forum's speaker lineup, the list of high-profile guests, and even the entire official guest list.

Taking his place is Liu Dayiheng, the newly appointed head of the Tongyi Qianwen project, announced just a day before the conference. The technical announcement segment, once a comprehensive showcase handled by a single individual, has now been divided into multiple sections, including models, agents, and infrastructure, reflecting a modular approach with clear divisions of labor.

For a company renowned for its technological prowess, this arrangement signals a strategic pivot in Alibaba AI, transitioning from a phase of technological exploration to one of commercial implementation. The focus has shifted from who builds the best model to who can derive the greatest commercial value from it.

A Year's Transformation: From Prominent Figure to Missing Name

At the 2025 Yunqi Conference, Zhou Jingren was the shining star of Alibaba AI. As CTO of Alibaba Cloud and head of Tongyi Lab, he unveiled seven successive iterations of the Tongyi large model, spanning intelligence levels, Agent tool invocation, coding capabilities, deep reasoning, and multimodality.

That year, the number of derivative models from Tongyi Qianwen exceeded 170,000, surpassing Meta's Llama series to become the world's largest open-source model family.

Zhou was also elected as an Alibaba Partner by year-end, joining the company's highest decision-making body, which had been trimmed to just 17 members.

At that time, Zhou Jingren embodied Alibaba AI's technological vision. In a 2025 Yunqi Conference interview, he articulated his belief that the rapid release of new products in the large model field was in line with the global AI industry's competitive landscape. This philosophy of rapid iteration aligned perfectly with Alibaba's strategic needs at the time to catch up.

But just a year later, the landscape had changed. Lin Junyang, once the technical soul of Tongyi Qianwen, resigned in March 2026 and founded the AI startup Pragmatik Labs several months later. Zhou Jingren was appointed as Alibaba Cloud's CTO on an emergency basis to temporarily oversee the Qwen model, stabilizing the R&D front.

This marked his first role adjustment, more akin to a firefighting effort.

In the following months, Alibaba AI underwent intensive organizational restructuring. The ATH Business Group was established in March, the Group Technology Committee was set up in April with an upgrade to the Tongyi Large Model Business Unit, and the Token Foundry Business Unit was formed through a merger in June, with Zhou Jingren transitioning to Chief Scientist.

Three adjustments in three months saw Zhou's every move appear to be a promotion, from CTO to Chief AI Architect of the Group Technology Committee, and then to Alibaba's first-ever Chief Scientist. While his titles grew more prestigious, his actual business execution power was continuously diluted.

The title of Chief Scientist is the highest academic honor in Alibaba's technology system, but in the context of Chinese tech companies, it often signifies a shift from the frontlines of business to the backlines of research.

The core model R&D team that Zhou previously oversaw, the Tongyi Large Model Business Unit, was entirely transferred under the Token Foundry Business Unit directly led by CEO Wu Yongming. Instead, Zhou was tasked with leading the establishment of the AI Future Research Institute, focusing on the exploration and breakthrough of cutting-edge AI technologies.

The transition from overseeing core business to exploring cutting-edge technology involves a transformation and gap that is difficult to encapsulate in a few words. Regarding rumors of his departure that had circulated previously, Alibaba officially clarified and denied them. At the Shanghai Bund Conference on September 10, Zhou still attended as the Group's Chief Scientist for a roundtable discussion on agent systems and AI's understanding of individual contexts, alongside Ant Group CEO Han Xinyi.

He remains active at the industry's forefront, just no longer on the main stage of product launches. This reflects how, in Alibaba AI's strategic shift from technological breakthroughs to commercial monetization, the role of technological leaders is being redefined.

Beyond Fragmented Speeches: The Logic Behind Alibaba AI's Power Restructuring

The decision to split Zhou Jingren's 50-minute solo speech at the 2025 Yunqi Conference into four separate speeches this year reflects a comprehensive logic of organizational restructuring. The core direction of Alibaba AI's 2026 organizational restructuring can be summarized by two words: centralization and alignment.

At the beginning of the year, Alibaba maintained a decentralized AI architecture, with Tongyi Lab, MaaS business line, Qianwen Business Unit, and AI Innovation Business Unit scattered under different reporting lines. However, by August, the Cloud Intelligence Group merged with T-Head to form AI Cloud and Computing Services, while the AI Model Lab, Qianwen Consumer Business, and Qianwen Office were integrated into the AI Lab and Applications.

The entire AI business was consolidated into two major segments, reporting directly to CEO Wu Yongming. The naming of the Token Foundry Business Unit itself hints at the new strategic direction: Alibaba aims to become a Token factory, standardizing and scaling AI capabilities for billing in the form of Tokens. The MaaS platform is replacing traditional cloud resource leasing as the core of Alibaba AI's new business model.

This transformation means Alibaba AI's KPIs have shifted from model benchmark scores to Token sales. When evaluation metrics switch from technical to commercial indicators, the types of talent the organization needs also change. Model R&D requires top scientists and algorithm engineers, while a Token factory needs product managers, delivery engineers, and sales teams.

Correspondingly, on the Yunqi Conference stage, there is no longer a need for a single executive to outline grand technological visions. Instead, specific business leaders are needed to discuss model capabilities, product solutions, industry implementations, and developer ecosystems.

Thus, we see that this year's Yunqi theme is "Intelligence for Practical Use," with a core focus on Agentic AI. Liu Dayiheng's speech theme is "Qwen: Towards Real-World Agents," emphasizing real-world applications rather than technical parameter comparisons. This itself is a direct reflection of the strategic shift.

Microsoft and Amazon made similar adjustments in 2026. According to public reports, Microsoft established the 6,000-person Frontier Company with a $2.5 billion investment, aiming to help clients implement AI solutions. Amazon Web Services invested $1 billion to set up a dedicated AI delivery department, deploying thousands of frontline engineers.

Global tech giants are converging in logic: the first half of the AI competition was about model capabilities, while the second half is about who can faster turn model capabilities into client purchase orders.

Wu Yongming's speech at the main forum of the Yunqi Conference confirmed this point. He judged that machines are becoming the main force in thinking and proposed that AI models, AI chips, and AI cloud are the three cornerstones of the machine intelligence era. This is a typical infrastructure narrative.

Wu is telling the market that Alibaba's AI is not just an AI model company but a full-stack infrastructure from chips to cloud to models. In this narrative, Zhou Jingren's position becomes nuanced. He built the most core model layer of this infrastructure, but when the infrastructure enters the operational phase, the operator does not need to be its original builder.

The Significance of Absence: Technological Idealism and the Second Half of China's Large Model Competition

The reason Zhou Jingren's absence is worth discussing lies not in him personally but in the broader context of the type of role he represents within China's AI industry. He was the architect of the Tongyi Large Model from scratch, a key builder of Alibaba AI's technological system, and a technical representative elected as a Partner in 2025.

His disappearance marks that China's large model competition has entered a new phase, where the window for technological idealism is closing, and the pressures of engineering and commercialization are reshaping everything.

One telling detail is worth noting. When Zhou was elected as a Partner in December 2025, a report by LatePost mentioned that one of the key factors for his selection was Tongyi Lab's efforts over the past year to maintain Qwen's leading model position, which management evaluated as highly challenging. However, just half a year later, when Alibaba AI underwent organizational restructuring, Zhou's role was redefined.

This is not a denial of his personal abilities. Alibaba explicitly stated when denying rumors of Zhou's departure that the appointment as Chief Scientist was both a high recognition of Zhou's contributions and a strategic deployment for the future. However, the phrase "strategic deployment for the future" precisely illustrates the issue: in Alibaba's future strategy, Zhou's position is no longer at the frontlines.

This recalls Jack Ma's bold announcement at the 2017 Yunqi Conference when he founded DAMO Academy, pledging over 100 billion yuan in technological R&D investment within three years and introducing the legendary "Thirteen Sweeping Monks" of DAMO Academy.

Nearly nine years have passed, and most of the thirteen have left or faded from the core spotlight, with Zhou Jingren being the exception throughout. He has been present from the era of cloud computing's foundations to e-commerce's data intelligence era and now to the large model era, hitting all the right technical shifts. But when he stands at the 2026 node, the rules of the game have changed.

Alibaba no longer needs a technological leader who can rapidly iterate models with his team. Qwen's model capabilities are already strong enough, with Qwen-3.7 ranking second globally and first domestically in the Code Arena evaluation. What Alibaba needs is to transform large models from laboratory projects into group-level revenue businesses.

Of course, this is not a pessimistic narrative. Zhou's transition to Chief Scientist and leading the AI Future Research Institute may precisely be Alibaba's play for a bigger game, allowing top technical talent to break free from the constraints of daily business management and focus on cutting-edge exploration.

The AI Future Research Institute is positioned to focus on the exploration and breakthrough of cutting-edge AI technologies. Zhou may be planning something even further beyond Qwen. However, this path has a longer return cycle, higher uncertainty, and keeps him further from the spotlight. In an AI industry universally pursuing quarterly commercial progress, such a choice requires immense patience.

Another layer of speculation is that Alibaba is actively reducing the large model business's reliance on a single core figure. The top-level strategy and resource coordination are personally handled by Group CEO Wu Yongming; long-term cutting-edge technology exploration is led by Chief Scientist Zhou Jingren; and daily R&D and product delivery of foundational models are handled by the grown technical backbone Liu Dayiheng.

Three individuals each manage a segment, non-overlapping and non-bound, ensuring that even with subsequent personnel changes, the large model business narrative will not fracture. For large model competition, which has now entered the deep waters, organizational stability is far more critical than the endorsement of a single star executive.

Zhou Jingren's absence from the main forum of the Yunqi Conference may seem like a minor personnel arrangement detail on the surface. However, against the backdrop of Alibaba AI's organizational restructuring, it serves as a footnote to the era that China's large model competition is transitioning from the age of technological heroes to the age of systems engineering.

When heroes exit the stage, it does not represent failure; but the spotlight has indeed shifted to another group of people.