Huawei launches AI Data Platform

Huawei unveiled an Artificial Intelligence (AI) Data Platform at the Huawei AI Dual Connectivity (DC) Innovation Forum during MWC Barcelona 2026 to address barriers to AI agent adoption and to reinforce the data foundation for enterprise digital and intelligent transformation.

Enterprises still struggled to deploy AI agents at scale because of delayed knowledge acquisition, low retrieval accuracy, inefficient inference in long-sequence and multi-turn interaction scenarios, and the lack of task memory and experience accumulation, which kept most AI agents confined to the demonstration stage and not ready for production-level enterprise applications.

Xie Liming, the President of the Flash Storage Domain of the Huawei Data Storage Product Line, introduced the AI Data Platform, which integrated a knowledge base, a KV cache, and a memory bank coordinated by an Update Compression Module (UCM). The platform continuously detected source data changes and converted raw inputs into knowledge in near real time, used multimodal lossless parsing and token-level encoding to create multimodal knowledge, and reported retrieval accuracy of over 95%.

The KV cache applied intelligent tiering and management to reduce repeated computing during inference, lowering inference latency and improving inference throughput for long-sequence and complex agent inference. The memory bank accumulated working memory and experiential memory during agent interaction, supported memory backtracking and multi-agent collaborative learning, and continuously optimized inference accuracy and efficiency.

Huawei said it would strengthen its investment in AI data infrastructure and work with global customers and partners to drive broader AI adoption across more fields, unlocking the full potential of data.