Huawei is expanding its semiconductor business with new AI chips and services. Its new AI strategy, as well as its full-stack AI driven portfolio was also unveiled by chairman Eric Xu at the Huawei Connect event in Shanghai.
Xu said Huawei’s AI strategy is "to invest in basic research and talent development, build a full-stack, all-scenario AI portfolio, and foster an open global ecosystem". Huawei’s full-stack portfolio includes the new Ascend AI chips, chip enablement, a training and inference framework, and application enablement. Along with the HiAI engine launched in April for smart device processors, they will run on the Cloud EI service platform, introduced at the end of 2017 to help businesses develop cloud-based, AI-driven services.
The bottom layer of the stack is the new Ascend 910 and Ascend 310 chips, for respectively data centres and smart mobile devices. The Ascend AI IP and chip series "delivers excellent performance per watt in every scenario, whether it’s minimum energy consumption or maximum computing power in data centres", Huawei said. The unified architecture also makes it easy to deploy, migrate, and interconnect AI applications across different scenarios, the company added.
Huawei said it plans to expand the portfolio to "all scenarios, both standalone and cooperative scenarios between cloud, edge, and device". The portfolio already includes the SoftCOM AI for telecom network O&M, HiAI to bring intelligence to its consumer devices, and the Huawei EI public cloud services and FusionMind private cloud solutions to help businesses and the public sector adopt AI. The portfolio will also include an AI acceleration card, AI server, AI appliance, and many other products, Xu said.
The company also showed an upgrade of its Atlas cloud hardware platform running the Ascend AI chips. Designed to build an AI infrastructure covering the end, edge and cloud, the new Atlas includes the terminal-facing Atlas 200 AI accelerator module, the DC-facing Atlas 300 AI accelerator card, the edge-oriented Atlas 500 AI edge station, and the Atlas 800 AI appliance for enterprises.
