Huawei Cloud unveiled its full-chain AI-ready data infrastructure capabilities at the 2026 China International Big Data Industry Expo in Guiyang.

On August 28, during the 2026 China International Big Data Industry Expo, the Data Elements Conference was held in Guiyang under the theme “From Data to Agent: Reconstruction and Innovation of the Data Value Chain.” Government agencies, leading big data groups, industry enterprises, and academic experts gathered to discuss the market-oriented allocation of data elements, the construction of high-quality datasets, trusted data spaces, and innovations in data infrastructure for the AI era.

As large language models and agents accelerate their penetration across industries, a fundamental shift is underway: models and agents have become the new consumers of data. This change in consumption patterns is driving big data platforms to evolve from traditional lakehouses to multimodal lakehouses and data agents, imposing entirely new requirements on data infrastructure in the AI era.

“Data is the core fuel for AI, and AI drives the release of data value,” stated Gu Xuejun, President of Huawei Hybrid Cloud, at the conference. As the AI industry enters the Agentic era, data has become the core fuel driving AI model evolution and agent growth. Huawei Cloud Stack delivers a comprehensive data capability system covering the entire data lifecycle to customers in critical sectors, enabling enterprises to effectively utilize their historically fragmented and accumulated data, transform data from a resource into a productive force, and unlock its true value.

Historically, enterprise data has been widely dispersed across clouds, data centers, production lines, and office systems. Without data integration, agents struggle to penetrate core business operations. Making an enterprise’s complete data portfolio truly accessible for AI has thus emerged as a new imperative for digital and intelligent transformation.

In response to this shift, Huawei Cloud has leveraged its “Hybrid Cloud + Data + AI” strategy and the Huawei Cloud Stack hybrid cloud foundation to build a full-chain capability spanning reliable data management, efficient data supply, intelligent data usage, and trusted data circulation. This architecture supports the construction of AI-Ready data platforms.

Reliable data management establishes a solid foundation for AI. The big data engine supports cross-AZ, cross-region, and cross-cloud disaster recovery across all scenarios, while metadata management unifies control over hundreds of millions of records. Through architectural decoupling of storage, compute, and management layers, the system ensures data reliability, manageability, and cost efficiency.

Efficient data supply transforms data into readily available “fuel” for AI. Huawei Cloud AI DataLake, a multi-modal intelligent data lake, relies on the synergy of three core engines to eliminate bottlenecks across the AI data supply chain. Tailored for model training and agent inference, it comprehensively upgrades data supply efficiency, addressing longstanding industry pain points such as slow delivery, limited volume, low precision, and weak compatibility.

Intelligent data usage elevates data from raw information to logic, and ultimately to action. Huawei Cloud’s intelligent data decision platform, DataArts, is designed for complex enterprise data scenarios. It unifies the ingestion of IT system data, OT production data, ET engineering data, and KT enterprise knowledge. Featuring over 40 built-in data processing operators, support for self-evolving ontology construction, and high-performance computing, DataArts enables efficient data access and utilization without requiring data migration.

Trusted data circulation unlocks full value realization. To address challenges such as uncontrollable data sovereignty, unverified participant identities, and non-auditable processes, Huawei Cloud employs privacy computing, data APIs, data sandboxes, and blockchain technologies to create a trusted data circulation environment. This ensures that data flows remain fully traceable, controllable, and verifiable. These capabilities have already been deployed in projects including the Shanghai City Data Space, the Nantong Home Textile Industry Data Space, and Yunnan Communications Investment.

The transition from “managing data well” to “enabling data to serve AI” marks a paradigm shift: data evolves from a static asset into a productive force that continuously drives model and agent evolution, thereby redefining the data value chain.

Technical paradigms must ultimately be validated through industrial practice. At the conference, XCMG Group, Guangdong Power Grid, Yangtze River Delta Data Chain, and Guizhou Shujubao presented frontline implementations showcasing how data better serves AI across data governance, high-quality dataset creation, trusted circulation, and data value operations.

As a leading enterprise in the construction machinery sector, XCMG Group initiated its “Smart Transformation, Digitalization, and Networking” strategic shift in 2023. Guided by this initiative and structured around a “1+4+2” framework and four major engineering projects, XCMG deeply integrated data elements into its product development, manufacturing, and operational chains. Data governance has thus moved beyond merely “organizing data” to laying a robust data foundation for enterprise intelligence.

As the data foundation solidifies, the next challenge becomes forming high-quality data that AI can “consume and learn from.” Guangdong Power Grid developed the “Yi Wan Wan” multi-dimensional tagging system, implementing intelligent labeling through a “pre-labeling + manual fine-tuning + multi-dimensional tagging” workflow. In customer service scenarios, smart recommendations of standardized scripts exceed 20,000 daily calls, achieving a 98% accuracy rate.

When data moves beyond single organizations, secure and trusted circulation becomes key to unlocking value. Yangtze River Delta Data Chain has built a “1+1+1+X” architecture centered on “blockchain + privacy computing.” It currently hosts 232 participating institutions, covers 140 application scenarios, has processed 4.3 billion blockchain transactions, and has interconnected logistics nodes across nine provinces and 40 cities.

As data begins to flow, new forms of value emerge. Zhou Lin, CEO of Guizhou Shujubao, noted: “The endpoint of data is not reports, but tokens.” Leveraging data sources from over 50 ministries and bureaus and 1,800+ data products, Shujubao has built an integrated “Data + Compute + Token” platform. Its subsidiary, TopenRouter, operates an independent token trading platform handling over 100 billion daily calls, and has established a “Base × Operations” collaborative model with Huawei.

From data governance and high-quality datasets to trusted circulation and value operations, a new pathway is taking shape: data no longer stops at governance, reporting, and analysis. Instead, it continuously feeds into models and agents, forging new value loops within real-world business operations.

During the event, the joint innovation launch for Guizhou Province’s public data resource “One Lake” project was announced, aiming to accelerate the efficient aggregation and intelligent provisioning of public data. Simultaneously, the “Data Element Scenario Joint Innovation Action” was unveiled, bringing together industry partners including Guotai Xindian, Sifang Weiye, Guizhou Shujubao, Hua’ao Data, Stardust Era, Shuxin Technology, Yonghong Technology, Shuyu Technology, and Primeton Information to drive the release of data element value in practical scenarios.

Behind these initiatives lies Huawei Cloud’s sustained accumulation in the data sector. In recent years, Huawei Cloud has partnered with the Guizhou Big Data Group to establish a provincial public data “One Lake” and jointly release outcomes for an AI-trusted data space. It has facilitated the effective integration of public, industry, and enterprise data in Shanghai, pioneered a manufacturing trusted data space pilot with XCMG, and continues to explore city-level data spaces and high-quality datasets empowering AI in Beijing, Shenzhen, Jiangsu, and other regions.

Institutions and enterprises face a dual mandate: protecting locally accumulated core data, business processes, and industry know-how, while continuously accessing rapidly evolving AI capabilities. Huawei Cloud Stack’s hybrid cloud architecture bridges data distributed across clouds, data centers, and production sites, enabling efficient data flow within secure, controlled boundaries while continuously supporting model training, agent execution, and intelligent decision-making.

Crucially, enterprise business data, interaction data, and intelligent experience will continue to accumulate within their own AI data lakes, establishing a virtuous cycle where “data drives model evolution, and models entering business operations generate new data.” This feedback loop ensures that enterprise AI grows progressively smarter with use.

Looking ahead to the intelligent era, Huawei Cloud will continue to increase investments in cloud infrastructure, data, and AI technologies. By leveraging its “Hybrid Cloud + Data + AI” approach to connect the full chain of reliable management, efficient supply, intelligent usage, and trusted circulation, Huawei Cloud will work alongside clients and partners to continuously advance the evolution of data infrastructure.