Tencent AI execs clash over pace of AI development at cloud conference
At the 2026 Tencent Cloud AI Industry Conference, Tang Daosheng questioned if Tencent AI is moving too slowly. Chief AI Scientist Yao Shunyu deflected, then offered his own answer.
The tension was laid bare on stage. At the 2026 Tencent Cloud AI Industry Conference, Tang Daosheng, a senior Tencent AI executive, turned to Chief AI Scientist Yao Shunyu and asked a pointed question: is Tencent AI moving too slowly? The room fell quiet. Yao deflected at first, offering a non-answer about industry trends and the importance of careful iteration. Then he paused, and gave his own answer. “We are not moving slowly,” Yao said. “We are moving deliberately.” The distinction mattered.
He argued that Tencent’s strength lies in integration — weaving AI into its sprawling ecosystem of social, gaming, and cloud services — rather than racing to launch flashy models. But the deflection itself was revealing. If there was no issue, why dodge the question? The exchange was not scripted. Attendees described it as unusually candid for a corporate event where harmony is typically enforced. Tang’s question reflected a growing unease within Tencent’s AI ranks.
Rivals like ByteDance and Baidu have pushed large language models into consumer products at breakneck speed. ByteDance’s Doubao chatbot, for instance, now rivals ChatGPT in monthly active users across China. Tencent’s Hunyuan model, while technically strong, has been slower to reach users beyond its own apps. What casual observers miss is that Tang Daosheng is not a rogue critic. He leads Tencent’s cloud AI business, which depends on selling AI tools to enterprise clients. Those clients want speed.
They see competitors offering faster updates, cheaper inference, and more aggressive pricing. Tang’s question was a plea from the revenue side of the house to the research side: give us something to sell now, not next year. Yao’s response, while polished, did not fully address that commercial pressure. He emphasized that Tencent’s AI research prioritizes reliability and safety — a nod to regulatory concerns that have tripped up rivals.
But in a market where first-mover advantage is real, caution can look like paralysis. The audience, a mix of developers and corporate buyers, seemed unconvinced. Internal sources suggest the rift is not new. Tencent’s AI teams have long debated whether to prioritize foundational research or product deployment. The conference stage merely made the disagreement public. Yao represents the academic tradition of deep research; Tang embodies the commercial imperative to monetize.
Both are necessary, but their priorities are diverging. The company’s leadership is now watching closely. Pony Ma, Tencent’s co-founder and CEO, has publicly backed a “slow and steady” AI strategy. But that stance is being tested by quarterly earnings calls where investors ask why Tencent’s AI revenue growth lags behind peers. The conference exchange suggests that internal patience is wearing thin. What comes next will depend on whether Tencent can reconcile these two speeds.
The company has the data, the talent, and the distribution. What it lacks is a unified clock. Tang wants to run. Yao wants to walk. The market, meanwhile, is sprinting.