Chinas Data+AI sector About to Take off： Tencent Cloud Executive

　　TMTPOST — A Tencent Cloud excecutive said on Thursday that the Chinese tech giant is building comprehensive data-powered artificial intelligence capability and will launch a data agent product in the second half of this year.

　　Cheng Bin, the general manager of Tecent Clouds Big Data Infrastructure Product Center, made the annoucement at the Global Artificial Intelligence Development and Application Conference.

　　Cheng said that data serves as the crude oil for both traditional data platforms and AI large model platforms. Data+AI is not simply the addition of data and AI, but rather a deep integration and innovation, marking a leap from letting data speak to engaging in intelligent dialogue with data.

　　In 2024, China’s total domestic data production surpassed 40 ZB mark for the first time to 41.06 ZB, representing a year-on-year increase of 25%. Per capita data production reached approximately 31.31 terabytes , equivalent to more than 10,000 HD movies per person, up 25.17% year-on-year. Both total and per capita data production have surged.

　　With the growing data demands of AI large models, traditional data platforms are facing severe challenges in meeting the new data requirements brought by generative AI. At the same time, in generative AI application scenarios, data quality, compliance, and security have become core issues for enterprises.

　　Enterprises are grappling with a surge in data sources and severe data silos . At the same time, they need to build dynamic and traceable data governance systems to avoid hallucination outputs and privacy risks.

　　According to Gartner’s predictions, by 2027, 80% of data and analytics governance initiatives will become ineffective due to the absence of real or perceived crises. At the same time, IT spending focused on multi-structured data management will account for 40% of total IT expenditure on data management technologies and services.

　　Huang Shifei, the Vice President of Tencent Cloud and Head of Tencent Cloud’s Big Data Products, said that as AI capabilities become readily accessible, it is crucial for enterprises to awaken vast amounts of dormant unstructured data, bridge organizational silos, and ensure that the data fed to AI is high-quality and trustworthy, thereby avoiding hallucinations and risks.

　　Cheng pointed out that traditional data platforms are hitting bottlenecks in this new development cycle. The process from data mining to data training has become highly complex, driving up labor costs and lengthening the entire development cycle. From an objective perspective, the market focus has shifted from data engineering/science development to large models/Agents; in terms of usage, it has moved from people searching for data to Agents seeking data with specific questions; and in terms of platform positioning, it has evolved from a passive “data warehouse” to an active “knowledge substrate” that enables real-time querying and interaction by Agents.

　　Tencent Cloud Big Data has proposed the deep integration of “Data+AI,” aiming to build a natively unified Data+AI platform designed for the AI era—one that is data-driven and intelligence-oriented, to serve as the next-generation intelligent data platform and meet the evolving needs of the market and customers.

　　By integrating multimodal data processing, AI models, Data Agents, and more, Tencent Cloud is building a new-generation data intelligence platform, DIaaS. Through “Data+AI,” unstructured data can be effectively utilized, enabling all data to be understood by AI—so that data is not only stored, but also put to use. In Tencent Cloud’s view, DIaaS is not just a tool, but a capability that makes data intelligence as accessible as water, electricity, or gas.

　　 Cheng noted that using the multiplication sign suggests a desire for greater value creation. However, he believes that whether it’s addition or multiplication, the value of integrating Data and AI will only continue to grow in the future.

　　“Is the leap to the next generation of intelligent digital platforms driven by AI large models? Indeed, I believe there are two main reasons. First, the maturity of large language model technology has undoubtedly provided a better development path for data. Second, and most fundamentally, it comes down to demand. If users do not have such a need, or if the demand is minimal, then no matter how advanced the technology is, it cannot generate value,” Cheng explained.

　　Cheng emphasized that currently there are no AI big data startups in China that are comparable to Snowflake or Databric

　　He pointed out that after the advent of large models, many database companies have rapidly transformed into “Data+AI” companies. For example, both Snowflake and Databricks have successfully pivoted, with Databricks now valued at over $60 billion and Snowflake’s market cap reaching $74 billion. Their transformations appear to be quite successful. In contrast, there are no technically strong big data startups of this kind in China yet, but the future prospects are very promising.

　　“I think as soon as the second half of this year, we may see the emergence of new Data+AI companies in China that can benchmark against Snowflake and Databricks. We very much welcome this development—the more players in this track, the greater the possibilities. If we rely solely on ourselves, we may not be able to meet all the diverse customer needs,” Cheng said. He noted that building DIaaS is a long-term and systematic project that requires industry collaboration and joint exploration.

　　He predicts that

　　 Based on an annual growth rate of over 20%, it is estimated that by 2030, China’s data industry will reach a scale of 7.5 trillion yuan.