Huawei Cloud Stack Provides the Industry’s First Hybrid Cloud for Large AI Models, Driving AI Industry Momentum in Asia Pacific

BANGKOK, April 30, 2024 /PRNewswire/ — After more than half a century of development, AI is now flourishing. Generative AI and large AI models are revolutionizing applications in almost every industry.

Over the past decade, cloud computing has witnessed significant growth in China. Many companies have been integrating AI into practical use cases to accelerate digital transformation. Industry intelligence is key to future success.

Many large government organizations and enterprises have been spearheading digital transformation and pioneering cloud transformation. They are seeking to spur innovation on the cloud to accelerate intelligence.

Many government and enterprise customers in Asia Pacific are eager to deploy their own AI capabilities. Huawei Cloud Stack has created the industry’s first hybrid cloud for large AI models by incorporating Pangu models to Huawei Cloud Stack. This solution helps enterprises build their own large AI models. It provides the industry’s most comprehensive AI pipeline including computing platforms, cloud services, development suites, and professional services, enabling numerous models and applications and empowering a myriad of industries.

Enterprises in Asia Pacific are actively promoting data-AI convergence and the integration of innovative digital technologies into R&D, production, and sales. They are increasingly interested in diving into cloud. In this context, Hu Yuhai, Vice President of Huawei Hybrid Cloud, announced the release of Huawei Cloud Stack 8.3 in the Asia Pacific market.

Johnny Lyu, CTO of Huawei Hybrid Cloud for International Business, said: “Huawei Cloud is striving to become a preferred partner for enterprises to build large models and accelerate intelligent upgrade. With the extensive cloud services and powerful Pangu models provided by Huawei Cloud Stack, AI construction, development, and operations have never been easier. We are committed to building custom large AI models for every enterprise.”

An Intelligent World Is Approaching Fast

The past two decades have seen numerous emerging technologies such as virtualization and containers. The cloud technology has also continued to evolve and expand fast. It has been regarded as “fertile soil” for digital transformation. More and more pioneers have been putting tremendous effort into cloud transformation. And the cloud adoption is accelerating. In this new era, there are many more new opportunities than ever. Government organizations and enterprises are diving into cloud for continuous innovation in technologies, services, and business models. They are also accelerating intelligent transformation to unlock more possibilities. Many industries are rapidly changing now:

• Government: In the past, to get something done, residents or businesses needed to make multiple visits to a local government office, but now, with the Government Big Data Platform, many services can be handled using mobile apps. Government services are now much more efficient. What’s more, the Government Service Model can handle thousands of government services. With accurate semantic understanding, the model can accurately identify user intent and generate useful answers. Services can be handled faster and with fewer interactions needed. The service experience is much better.
• Finance: As financial systems continue to expand and become more complex, a vast amount of complex professional knowledge is spread across various systems. As a result, to query information, customer service personnel had to switch back and forth between different systems. This was time-consuming and would affect customer satisfaction. Now, an intelligent customer service system, powered by a large AI model pre-trained on data and knowledge about banking processes, policies, and case studies, can automatically generate service workflows and guides based on customer requests. The average number of steps needed to complete a task has been reduced from 5 to 1. The average handling time is reduced by 5 minutes, and the call handling time has been reduced by 15%.
• Manufacturing: A typical digital factory has more than 50,000 devices and more than 10,000 data collection points. Such huge amount of data needs to be collected and used to empower manufacturing. There are more and more smart factories powered by AI. Industrial IoT platforms and big data technologies are used to streamline the IT and OT systems and create comprehensive data assets. And AI-powered analysis enables smarter decision-marking, optimization, and execution for the factories.

AI is reshaping almost every industry and is radically changing the way we live and work. An intelligent world is approaching fast.

Government and Enterprise Customers Prefer Hybrid Clouds

A hybrid cloud is a great option for large government organizations and enterprises to establish a technical foundation for intelligent upgrade. Due to regulatory and compliance reasons, government organizations, financial institutions, and state-owned enterprises have to prioritize data security, so they often opt for on-premises cloud data centers. Driven by intelligent transformation across all industries, these customers are seeking to spur innovation on cloud. They want to embed innovative technologies like AI, cloud native, and big data into all processes, including production, consumption, and services, and quickly move all their core applications to the cloud. They are increasingly interested in diving into cloud.

A hybrid cloud suits all their needs. It can be deployed on-premises and keeps data on-premises. The innovation strengths of cloud vendors enable its continuous evolution.

Huawei Cloud Stack 8.3 is such a hybrid cloud. It is the industry’s first hybrid cloud for government and enterprise customers to build custom large AI models in one stop. It has three highlights:

• A solid computing backbone: Huawei Cloud Stack provides the industry’s most comprehensive AI pipeline, covering computing power, storage, networks, operators, development frameworks, development suites, and professional services, which lowers the barriers to building large AI models. A vast range of capabilities, such as software-hardware synergy, long-term uninterrupted training, cloud-edge synergy, and security and compliance services make it easier for government and enterprises to address core challenges in using large AI models.
• Numerous models and applications: Huawei Cloud Stack supports Pangu Models with a decoupled “5+N+X” three-layer architecture. It also supports more than 30 mainstream open-source large AI models in the industry and provides efficient, easy-to-use tools to help customers smoothly migrate models. Plus, Huawei Cloud Stack provides an intelligent orchestration tool to standardize and visualize model and application development, accelerating rollout from months to days.
• Empowering a myriad of industries: Huawei Cloud Stack provides more than 10 large AI models for diverse sectors like government, mining, and finance, along with professional services for planning, construction, development, training, and operations. Customers can conduct training, fine-tuning, and inference deployment on the models based on their own data. They can also share their own algorithms and applications. This helps consolidate the experience building large AI models and makes the models easy to share and monetize.

To help customers in Asia Pacific dive into cloud, Huawei Cloud released a new hybrid cloud version, Huawei Cloud Stack 8.3, with six highlights: one solid cloud foundation and five innovations covering databases, mainframe modernization, data lakes, large AI models, and industrial Internet.

Huawei Cloud Stack provides high availability, robust security, multi-architecture compute, and an open, trusted, more resilient cloud native infrastructure.

Huawei Cloud Stack has been stepping up efforts to boost innovation in the following key areas and making new technologies available on-premises as fast as possible:

• GaussDB provides a distributed database with high performance and reliability: GaussDB, in terms of high-concurrency handling, delivers excellent reliability and 30% better performance than many rival products. It also provides flexible disaster recovery and one-stop migration capabilities, which are critical to core systems of government and enterprise customers.
• Mainframe modernization accelerates transformation of core systems: Customers in many sectors, like finance, energy, and manufacturing are eager to move their core systems from siloed legacy mainframes to open platforms. To address these requirements, Huawei launched the mainframe modernization solution using innovation in software, hardware, and services. The solution provides three cloud migration paths: re-hosting, re-platforming, and rearchitecting of core systems. It provides cloud-hardware synergy, more dependable processes, and scenario-specific solutions.
• A converged data lake improves data sharing and circulation: Huawei Cloud Stack provides three innovative services: MRS (a big data service), DWS (a data warehouse service), and DataArts Studio (one-stop data governance). It also provides an enterprises-grade real-time data lake that supports OLAP and HTAP workloads and unifies data standards. With Huawei Cloud Stack, customers in diverse sectors like government and finance are now able to achieve efficient cross-source, cross-domain data collaboration and real-time data analysis and sharing.
• Industrial Internet is integrated with AI to enable new industrialization: Huawei Cloud Stack provides a reference architecture for a new industrial Internet platform. It enables industrial all-domain connections, data intelligence, and application innovation. It helps resolve pain points in different stages of industrial production and drives integration with AI.
• A hybrid cloud for large AI models helps enterprises in every industry build custom large AI models in one stop.

Unleashing Cloud Value Together

Some pioneers diving into cloud early on have already reaped a myriad of benefits. However, there are still many challenges for customers diving into cloud. Attitudes are slow to change; the way forward is often unclear; the technology stack maybe inadequate; the ecosystem collaboration can be poor; and there are not enough clear standards.

Based on a wealth of experience serving enterprises and insights into future-oriented technologies, Huawei Cloud released the Visions of Diving into Cloud white paper that predicts 10 most powerful technologies or models to be put into place in the government and enterprise markets. Huawei Cloud also proposed an initiative to dive into cloud, which calls for intensive efforts to tap into cloud native, innovate on cloud, and draw on profound expertise and experience. This initiative aims to promote best practices, establish new standards and practices, set up a robust system for talent cultivation, and ultimately, to help customers accelerate intelligent upgrade and thrive with digital and intelligent technologies.

View original content to download multimedia: Read More