Huawei Cloud is accelerating its global footprint and doubling down on AI innovation. With its ‘Everything-as-a-Service’ strategy and AI-native infrastructure, the company aims to help enterprises build and deploy intelligent solutions at scale – securely, reliably and across every region. While attending Huawei Connect in Madrid, Alix Pressley, our Director of Strategic Content, sat down with William Dong, President of Huawei Cloud Marketing, who discussed this in more detail.
What is Huawei Cloud’s core strategy for its business globally?
The core strategy of Huawei Cloud is ‘Everything as a Service’, which means infrastructure as a service, technology as a service and expertise as a service. Our customers interested in understanding new technology or solutions can simply browse the Huawei Cloud website to find the latest services and solutions. Everything is accessible, as a service.
How is Huawei Cloud’s overseas business progressing and what sets Huawei Cloud apart from the competition?
In recent years, Huawei Cloud has been growing rapidly, especially in our overseas markets. Last year, our overseas public cloud services grew by more than 50%. This year, with AI booming, we’re maintaining steady growth.
Our core competence capability is built on AI. There are two main points here. First, we want to build AI-native infrastructure. Today’s hyperscalers, including Huawei Cloud, start with cloud-native technologies, such as virtual machines or containers, which operate like small computers controlled by CPUs. However, with AI trends, applications like AI agents require massive computing power as well as storage and networking. Therefore, the architecture must evolve from CPU-centric to parallel architecture – enabling resources like GPU and NPU computing, memory and storage to scale flexibly.
Based on this concept, we launched our Supernode solution this year at Huawei Connect in China. We believe the Supernode architecture is the best for AI. For example, if you take major Large Language Models (LLMs), they often use a Mixture of Experts (MoE) architecture. Our Supernode supports over 384 NPUs, allowing each expert to run on a separate NPU, maximising resource utilisation and performance while reducing costs.
Our second point of competitiveness is our focus on AI for industries. Many enterprises can access open-source LLMs, but the challenge lies in deploying them within their own environments and integrating their internal knowledge bases. Huawei aims to provide an AI platform that helps customers build their own models and AI agents. Having long served enterprise clients, we combine deep industry expertise with strong technological capabilities, which forms a core competitive advantage.
Are there are any regional nuances in terms of how cloud is used in the different countries or regions you operate in?
Regarding global competition, while regional differences exist, the more important factor today is what unites these regions. In the cloud era, we saw significant gaps between regions in terms of digitisation and cloud adoption. However, in the AI era, the gap is narrowing – many organisations are now adopting AI applications.
AI represents a huge opportunity globally. Whether in Africa, the Middle East, Asia Pacific, Europe or Latin America, our customers are increasingly leveraging AI technologies.
Currently, enterprises emphasise security and stability when moving to the cloud, especially in Europe. How does Huawei Cloud ensure these aspects for its customers?
Security is crucial. To use AI effectively, companies need unified AI computing resources and model platforms, which naturally depend on secure cloud infrastructure. Huawei Cloud has maintained a record of zero critical incidents over the past two years – a testament to both our technology and our culture of reliability and redundancy.
For example, in one data centre, a fire was caused by a malfunctioning cooling system. Huawei Cloud’s redundant resources (including dry ice systems) prevented any disruption. Similarly, during a power outage in Chile, our backup power engines kept services running for up to 144 hours. This culture of investing in redundancy – though costly – ensures reliability and customer trust.
Additionally, Huawei implements strict security awareness programmes and training for maintenance and R&D teams. We employ a seven-layer cybersecurity defence system to protect users at every level.
How does Huawei Cloud provide AI cloud services to promote the digital and intelligent transformation of enterprises?
We offer a full-stack AI solution built on our AI platform, which allows customers to select and customise LLMs as needed. Our ModelArts platform simplifies model training and fine-tuning, providing tools that make AI development accessible to enterprise users.
We’ve also launched Versatile, our AI agent platform, which allows customers to easily create AI agents by simply describing the type of agent they need – it then generates one automatically.
Are there any global customer success stories you’re able to share?
With our AI solutions, we have served over 1,800 customers all over the world. For instance, our client, Azerion, uses Huawei’s AI platform to build dedicated AI models for media and advertising, enabling its customers to generate ads independently.
A university in Hungary – and other universities today – are using Huawei’s AI platform to create AI assistants for students and teachers, integrating institutional knowledge bases with the latest LLMs to improve learning experiences and offer more choice.
Initiatives like Tech4All further showcase Huawei’s commitment to using technology for positive social impact – from education and wildlife protection to cultural preservation in Africa and Europe.
Looking ahead to 2026, what are the key development directions for Huawei Cloud?
Our ambition is to become the AI pioneer in industry. We’ve fully embraced AI and enhanced our investment in vertical AI models and AI agent platforms. We strive to provide our customers, partners and developers with the best AI infrastructure and the best AI agent platform.
