BARCELONA, Spain, March 5, 2026 /PRNewswire/ — At the Huawei Product & Solution Launch during MWC Barcelona 2026, Yuan Yuan, President of Huawei Data Storage Product Line, officially launched Huawei’s AI Data Platform. The platform integrates technologies of knowledge generation and retrieval, KV cache for inference acceleration, memory extraction and recall, as well as Unified Cache Manager (UCM). It bridges the gap between models and business value.
Yuan Yuan, President of Huawei Data Storage Product Line, delivering a keynote speech
Despite rapid AI advancement, most AI models have not been thoroughly integrated into core services. This is largely due to more focus on training than inference, which is the key to model adoption. Improved data processing is needed to address challenges like AI hallucinations, slow response, and throughput constraints in inference.
Huawei Data Storage officially launched a "3+1" AI Data Platform to tackle these challenges:
- Knowledge generation and retrieval with high-accuracy multimodal knowledge for more accurate retrieval
It converts multimodal resources into high-accuracy knowledge through multimodal lossless parsing and token-level encoding, ensuring a retrieval accuracy of over 95%.
- KV cache for inference acceleration using historical memory data for faster and more efficient inference
It applies intelligent tiering and management for KV cache to greatly expand the context window and reduce repeated computing. This results in a 90% reduction in time to first token (TTFT), significantly improving the speed of AI responses.
- Memory extraction and recall with context memory management for making models smarter with use
It can accurately extract historical data and experience, and accumulate them into recallable memories. This memory mechanism makes models smarter with use.
- UCM for managing and scheduling memory data throughout the lifecycle
It manages memory data at three cache levels to intelligently schedule the knowledge base, KV cache and memory bank.
The Huawei AI Data Platform provides the appliance mode for greenfield deployment and independent mode for evolving deployment. The appliance mode uses the OceanStor A800 system. This ensures ultimate performance and flexible scalability. The independent mode uses the architecture of AI data engine nodes + OceanStor Dorado storage. Data engine nodes can be added to upgrade existing systems, protect prior investments, and enable smooth AI transformation for businesses.
"Huawei will continue to deepen technological innovation. With the AI Data Platform serving as a bridge, we will transform model capabilities into real business value," said Yuan Yuan.
Contact:
Zoe Yin, yinzijun@huawei.com
