Tencent Cloud at Davos: Unlocking AI for All, Making Value Accessible

23 January 2026

DAVOS, Switzerland: At the Annual Meeting of the World Economic Forum (WEF) in Davos, Dowson Tong, Senior Executive Vice-President of Tencent and CEO of Tencent Cloud and Smart Industries Group, emphasised that the real-world value of artificial intelligence is further realised when open ecosystems empower consumers and businesses to choose from a variety of large models tailored to specific, practical applications.

“When people talk about AI, we tend to think of one large super system and refer to it as Artificial General Intelligence (AGI). However, there are many different types of models serving different purposes,” said Tong. “For Tencent Cloud, we recognise that customers want choices. That’s also our AI strategy: to provide tools and products that are model-agnostic. I believe this restores the power of choice to our customers, allowing them to select the most suitable model for their needs,” he added.

This approach forms the cornerstone of Tencent’s AI strategy, which adopts a two-pronged approach. The company is simultaneously enhancing its own proprietary model capabilities while remaining open and collaborative towards global innovations in large models. The aim is to enable every user and enterprise to select the most effective model for their specific requirements, empowering them to address real-world challenges efficiently.

Tencent’s internal operations serve as a testing ground for its AI vision. Its AI coding assistant, CodeBuddy, is now utilised by more than 12,000 engineers and assists in generating over half of all new code. The tool has reduced average coding time by over 40%, and notably, 90% of the code for Tencent’s recently upgraded AI CLI tool, CodeBuddy Code, was generated by the assistant itself.

Furthermore, Tencent is committed to investing in its full-stack Hunyuan large model series. Over the past year, the company has launched more than 30 new models across various domains, including advanced hybrid reasoning, image generation, video synthesis, and 3D content creation. The high-performance Hunyuan 2.0, which employs a Mixture-of-Experts (MoE) architecture with 406 billion total parameters (and 32 billion active parameters), positions itself at the forefront of the industry in both inference capability and operational efficiency. Meanwhile, Tencent’s open-source Hunyuan 3D model has been downloaded over 3 million times on the open-source platform Hugging Face, becoming one of the most popular 3D open-source models for creators, developers, and communities worldwide.

To date, Hunyuan powers more than 900 internal Tencent scenarios, from Tencent Meeting and Weixin/WeChat to gaming and advertising. The tangible impact is reflected in Tencent’s 2025 third-quarter financial report, which highlights strong AI-driven growth: AI contributed to a 21% year-on-year increase in marketing services revenue and a 22.8% rise in gaming revenue.

Tencent extends this expertise to its customers through a straightforward yet powerful framework: agent-based tools, cloud-native platform support, and plug-and-play solutions. This approach has already delivered results across more than 30 sectors. Retailers leverage Tencent’s AI to accelerate 3D product design; medical researchers utilise it for new medicine discovery; and marketers employ it to boost campaign conversions.

A notable example is Bambu Lab, a global leader in 3D printing, which reengineered its design workflow using the Hunyuan 3D model, making professional-grade modelling accessible to a broader audience and sparking a new wave of consumer creativity.

Tong also emphasised the importance of providing AI tools to young learners. The new generation can learn to use AI through educational institutions and via applications like Tencent Yuanbao.

He stated, “Encouraging young people to be curious and to use these freely accessible AI tools is the best way to cultivate the habit of integrating AI into their learning and daily lives.”