Summary Bullets:

• Tencent continues to advance its Hy AI model, focusing on cost-efficiency, performance, and specialization around productivity tasks.

• With Tencent WorkBuddy, Tencent is helping enterprises leverage AI to raise workforce productivity to benefit from the AI technology.

Tencent Cloud has been advancing its AI solutions despite not grabbing headlines internationally. The company is preparing the release of Tencent Hy4, the fourth generation of the Tencent Hy (Hunyuan) AI model. Tencent Hy4 is a 770 billion parameter, open-source, mixture-of-expert (MoE) model with 49 billion active parameters per token (i.e., optimizing for speed and compute resources by activating only a subset of the full model parameters). It is designed to improve reasoning efficiency and to support productivity tasks and visual understanding. The company has also designed Tencent Hy4 to support its AI products and applications by bringing expertise across Tencent including software engineering, gaming, finance, and security. As cost is one of the major concerns for enterprises, Tencent Cloud will also be adding Tencent Hy4 Flash, a more cost-effective model that carries Tencent Hy4’s visual understanding and agentic capabilities while reducing computational overhead. This is also responding to the broader industry trend toward a two-tier strategy, with a deep reasoning and a highly optimized, low-latency version (e.g., Gemini 3.8 Flash, GPT-4o mini, Claude 3.5 Haiku, DeepSeek-V4.1-Flash, and Qwen3.8-Flash).

Tencent Cloud’s approach to AI is also to drive adoption by working with enterprises to develop agentic solutions that are practical, meet the needs of business, and allow for faster return on investment (ROI). It has launched Tencent WorkBuddy a desktop autonomous AI agent that can execute multi-step office workflows to deliver finished tasks. This removes the need for users to switch between business applications to complete a deliverable, for example, by connecting with Tencent’s WeCom (WeChat enterprise version), Tencent Docs, Tencent Meeting, and ima (knowledge management) for productivity use cases. Key Tencent WorkBuddy use cases include market research, content production, web/app development, and data analysis. This product has gained rapid adoption in China, and Tencent Cloud has extended it to international customers.

Tencent WorkBuddy allows users to turn a single instruction into a completed output through multi-agent orchestration, assigning tasks to various specialized agents (e.g., data collection, trend charting, presentation of data). The product offers over 100 AI skills (through Tencent WorkBuddy Skills Marketplace) that are pre-built for roles such as market research, financial analysis, UI/UX design, and more. Tencent WorkBuddy is an open system that works across different business applications (e.g., Microsoft Office), and it also connects with third-party systems to complete cross-platform workflows. This is either through pre-built connectors (e.g., for GitHub, Jira, Confluence, Google Drive, Slacks, Telegram, and more) based on the Model Context Protocol (MCP) approach; or custom connectors through APIs.

Through Tencent TokenHub (a model-as-a-service platform supporting multiple LLMs), Tencent WorkBuddy allows users to leverage Tencent Hy4, connect to third-party models (e.g., DeepSeek, Kimi, and GLM), or route sensitive tasks to open-source models. Users can also connect to closed models such as OpenAI’s Claude and Google’s Gemini through APIs. Tencent WorkBuddy also has a governance and control layer to help customers impose roles-based access controls, analyze usage, and track spending. Besides offering Tencent WorkbBuddy as an SaaS-based solution, Tencent also has a dedicated virtual private cloud (PVC) option that allows data isolation and security.

Tencent WorkBuddy is not the only AI agent that Tencent is offering today. Its ‘Buddy’ family of agents also includes CodeBuddy for AI coding, DataBuddy for data engineering and governance, Miora an AI-native creative studio, and Ardot for UX/UI design. Expect Tencent to continue expanding its agentic AI capabilities to deliver faster business outcomes for enterprise customers. However, the majority of Tencent Cloud’s enterprise customers are in mainland China and the ecosystem built around Tencent WorkBuddy is predominantly Chinese business applications and systems. Customers overseas will need to urge Tencent Cloud to enable easier integration with their existing technologies such as ERP, CRM, and productivity tools.

Tencent Cloud is making good progress in winning customers outside of China, doubling its overseas customers over the past year. While there are other competitors offering various cloud and AI solutions, Tencent Cloud enjoys some competitive advantages. Leveraging Tencent, the parent company, Tencent Cloud has been winning clients in media, entertainment, fintech, and other digital and B2C segments. In emerging markets across Asia and the Middle East, enterprises are looking to diversify their technology ecosystem, manage geopolitical risks, and include more cost-effective offerings to combat rising costs. Tencent Cloud has a team of technical specialists that work with clients to implement the solutions (also known as forward deployed engineers or FDEs) and provide support to ensure client success. While major technology vendors often pursue large enterprise customers (e.g., Forbes Global 2000), there are many mid-to-large enterprises across Asia that are often underserved. There are significant opportunities for Tencent Cloud to engage these companies to transform their business with AI, either directly or potentially more effectively through service provider partners.