2026 China Database Technology Conference Launches: Data Fusion and AI Leadership

Conference Overview

From August 20 to 22, 2026, IT168 together with ITPUB and ChinaUnix hosted the 17th China Database Technology Conference (DTCC 2026) in Beijing. The event’s theme, “Data Fusion and AI Leadership,” featured two main venues, 16+ technical sessions, and dozens of industry experts covering database kernel innovation, cloud‑native and distributed practices, AI‑enabled data, vector databases, real‑time warehouses, and industry pain points.

ITPUB’s Full‑Chain Service and AI Strategy

ITPUB General Manager Tang Chuan introduced the community’s full‑chain service capabilities and its AI‑landing strategy, highlighting 25 years of technical accumulation, 45+ past DTCC events, and over 3,000 expert participants.

Dameng PAI Database Appliance

Dameng’s product line manager Zhu Honglei presented the latest research results of the Dameng PAI database appliance, focusing on two dimensions: performance and intelligent computing.

Performance: Three core breakthroughs – ultra‑fast storage with tiered architecture and smart caching, self‑developed compute offload embedding a compute engine in storage, and full‑stack acceleration using RoCE v2 to build a 200 Gb/s network, aiming to turn “usable” systems into “truly replaceable” ones.

Intelligent Computing: A three‑layer foundation – multi‑model DM9 kernel, an open platform (MCP Server and Open API) supporting vector retrieval and mixed queries, and 24/7 top‑DBA expert team for autonomous operation.

Tencent Cloud’s TDSQL Nexa for the Agent Era

Vice President Wang Yicheng explained the fragmentation problem of data across stacks when building AI Agent applications. Tencent’s TDSQL Nexa unifies transaction processing, full‑text and vector search, AI computation, and large‑scale analytics on a single platform, supporting direct connection to existing data sources. Benchmarks show over 10× performance improvement and more than 60% cost reduction compared with open‑source combos, and a 50% average performance boost for medium‑scale workloads with AI‑generated SQL.

TDSQL‑C Serverless provides per‑instance isolation with instant branch creation, while TDSQL Nexa and Boundless deliver unified metadata, multi‑model tables, and unified storage, retrieval, and analysis for Agents. Additional components (DatabaseClaw, DBbrain, DTS) handle memory management, diagnostics, and migration.

Alibaba Cloud’s Agentic DB Capabilities

Chen Zongzhi described the “Agentic DB” concept, outlining four core capabilities: multi‑model data handling (structured, vector, JSON, graph, full‑text), serverless elasticity with second‑level cold start, Git‑like branching for zero‑copy forks and rollbacks, and natural‑language interfaces via MCP/Tool/Skill standards.

Alibaba also launched the AI‑Native Database Service (AIDBS), a one‑stop platform offering control‑plane Agent for instance management, DevOps Agent for change review and SQL optimization, and application‑development Agent for metadata editing, analysis, and one‑click web app generation.

OceanBase’s AI Lakehouse Evolution

OceanBase AI Data Platform R&D Director Yang Zhifeng traced the product’s evolution from a distributed database to an AI lakehouse. The lakehouse unifies structured, semi‑structured, and unstructured data with real‑time processing and intelligent retrieval, using a storage‑compute separation architecture with hot/auto/cold tiers, multi‑modal hybrid search (vector, full‑text, scalar, graph), and built‑in high‑performance vector indexes (HNSW, IVF). It also supports Spark and Ray for open AI computing.

Key features include multi‑model tables, forkable databases for versioned data sandboxing, and a unified data foundation for AI Agents.

GoldenDB’s Core‑Base and Intelligent Fusion

GoldenDB Vice President Tu Yaofeng presented two themes: “Core Foundation” and “Intelligent Fusion.” Under the core foundation, GoldenDB leverages 20 years of technology and over 1,000 patents to support critical industry workloads, handling billions of daily transactions and serving thousands of customers.

In the intelligent fusion layer, GoldenDB envisions databases serving massive Agents, requiring support for MCP, external memory, and inference engines. It implements a unified architecture for JSON, vector, and file data, hybrid search and analysis (HSAP), and AI‑driven query cost estimation, index recommendation, and performance tuning, achieving self‑awareness, self‑diagnosis, and self‑evolution.

Conclusion

The three‑day DTCC 2026 highlighted the rapid transformation of databases from mere storage engines to foundational platforms for AI Agents, with industry leaders showcasing performance breakthroughs, multi‑model capabilities, and cloud‑native designs that will shape the future of data‑centric intelligent systems.