On September 20, 2023, Chinese tech giant Huawei held the Huawei Connect 2023 conference in Shanghai, bringing together business leaders, tech experts, partners, developers, and industry stakeholders from around the world to “explore new opportunities for an intelligent future.” Sabrina Meng, deputy chairwoman, rotating chairwoman, and CFO of Huawei, delivered the keynote address to kick off the conference.
In her keynote address, Meng shared the all-intelligence strategy put forward by Huawei with the goal of accelerating the intelligent transformation of various industries. The objective is to make all things interconnected, all applications modelable, and all decisions computable. She highlighted three key points:
- Huawei is committed to building a solid computing power foundation in China, creating an alternative option for the world. In the future, it will continue to enhance the integration capability of “hardware, software, core, edge, and cloud” to meet the diverse artificial intelligence computing needs of various industries.
- Huawei will support the proliferation of large models in the era of intelligence through the openness of its computing power foundation, AI platform, and development tools. “We support every organization in training their own large models with their own data, allowing each industry to develop its own industry-specific large models,” Meng said.
- Huawei has, for several decades, built its technological advantage by delving into the core of communication and computing theories, continuously investing in fields such as mathematics and algorithms, chemistry and materials science, physics and engineering technology, standards, and patents. Looking toward the future, Huawei will continue to uphold hardware and software openness and collaborate with industry partners to develop the industry and build a thriving ecosystem.
David Wang, Huawei’s executive director, chairman of its ICT infrastructure managing board, and president of its enterprise business group, unveiled Huawei’s new Atlas 900 SuperCluster. This new AI computing cluster, the latest offering in Huawei’s Ascend series of computing products, utilizes new architecture optimized for training massive AI foundation models with over one trillion parameters. The Atlas 900 SuperCluster is equipped with Huawei’s Xinghe Network CloudEngine XH16800 switch. With AI entering the era of large models, high computing power is becoming the core engine for AI development.
The introduction of the cluster, equipped with “innovative super node architecture” is part of a move to transform the traditional server stack model. It is expected to boost overall computing power and take the speed and efficiency of foundation model training to new levels while improving system reliability at the component, node, cluster, and service levels.
Huawei has also upgraded the Ascend Computing Language (AscendCL), a unified programming interface to meet developers’ requirements for custom frameworks and operators. Using a more efficient programming approach significantly reduces development cycles, enabling rapid development of AI models and applications.
In collaboration with customers and partners, Huawei has released a whitepaper titled “Architecture for Intelligent Transformation of Public Services”, shedding light on intelligent transformation trends and implementation directions, as well as unique ways to make public services more inclusive and well-targeted. The whitepaper features nine major industries, including finance, government affairs, manufacturing, power, and railways.