Huawei Cloud Launches Hybrid Cloud Framework

The global launch of Huawei Cloud Stack (HCS) extends its on-premises cloud solution to international markets, building on a base of over 5,500 customers. This hybrid cloud offering is designed for industries like government, finance, and telecom where data sovereignty and security are paramount. The platform provides more than 120 on-premises cloud services, aiming for a consistent experience with its public cloud. For AI and machine learning workloads, HCS is tightly integrated with Huawei's ModelArts, a one-stop AI development platform comparable to AWS SageMaker and Google's Vertex AI. ModelArts supports the entire MLOps lifecycle, from data preprocessing and auto-labeling to distributed training and one-click model deployment to cloud or edge devices. This allows for the creation of full ML pipelines, moving beyond notebooks to production-level systems. A key component of Huawei's AI strategy is the Pangu large model series, which is available on HCS. These models come in various sizes, up to 100 billion parameters, and are designed for specific industries with foundation models for NLP, computer vision, and scientific computing. Developers can use ModelArts to fine-tune these pre-trained Pangu models with their own datasets and deploy them as real-time inference services via REST APIs. Underpinning its AI capabilities is Huawei's own Ascend AI hardware, which offers an alternative to Nvidia GPUs and is designed to provide significant performance for model training. For instance, the Ascend 910 chip is a core part of Huawei's large-scale distributed training systems. This vertical integration of hardware and software is a key part of Huawei's strategy to compete with other major cloud providers in the AI space. The platform also includes a suite of data engineering tools to support AI development. This includes FusionInsight, a lakehouse architecture for big data analytics, and ROMA, an integration tool for connecting new and legacy IT assets. For developers working with large datasets, Huawei offers a vector database and its Omni-Dataverse, a global file system to unify data across different locations. In the competitive landscape, Huawei Cloud positions itself as a leader in AI and machine learning, particularly in the Asia-Pacific market. While AWS offers a broader range of general cloud services, Huawei focuses on deep integration of its AI processors and industry-specific solutions. For companies with operations in regions where Huawei has a strong presence, or those looking for an all-in-one AI development ecosystem with specialized hardware, HCS presents a compelling alternative to AWS, Azure, and Google Cloud.