Transwarp Launches GPU-Native Cognitive Database at WAIC 2026, Targeting AI Agent Era

Transwarp (星环科技), a leading Chinese AI infrastructure software company, unveiled a preview version of its GPU-native cognitive database at the World Artificial Intelligence Conference (WAIC) 2026 in Shanghai on Friday. The company, participating in WAIC for the ninth consecutive year, showcased the new database under the theme "Cognitive Emergence · Data Rebirth" at Booth H1-B124.

The GPU-native cognitive database was the centerpiece of Transwarp's exhibition. According to the company, the GPU-native architecture delivers order-of-magnitude performance improvements over traditional approaches in TPC-DS benchmark testing, and significantly accelerates AI inference pipelines in scenarios including intelligent investment research and full-chain credit analysis. Interactive demonstrations at the booth allowed visitors to experience the performance advantages in real-time data processing and intelligent decision-making.

Beyond the cognitive database, Transwarp demonstrated its complete AI infrastructure product stack. The AI-Ready Data Platform unifies multi-modal enterprise data management with natural language-driven data processing. Sophon LLMOps covers the full lifecycle of large language model operations from corpus development and prompt engineering to model training and deployment. AgentBox enables flexible orchestration of models, knowledge bases, and plugins into personalized agents, with integrations to major Chinese messaging platforms including Feishu, WeCom, and DingTalk. XClaw leverages an Agentic Loop mechanism for data analysis and presentation generation, while Astro enables intelligent data governance through simple natural language commands.

Transwarp presented industry practice cases alongside clients including Shanghai Pudong Development Bank and Tianjin Energy Investment Group, demonstrating real-world deployments in the finance and energy sectors. The showcase highlighted how Transwarp's AI infrastructure supports the complete pipeline from data preparation and model training to agent deployment in enterprise production environments, driving knowledge asset accumulation and operational efficiency.

As AI applications across Chinese industries accelerate the transition from proof-of-concept to production deployment, Transwarp's GPU-native strategy represents a significant bet on hardware-accelerated data infrastructure as the foundation for enterprise AI adoption in the Agent era.