Huawei is rolling out its Intelligent Computing Platform Service Solution internationally. The end-to-end package allows companies to set up AI infrastructure in 4 to 6 months and roll out new AI models in 5 days.
The company announced this at MWC Barcelona 2026. The package includes facility and infrastructure design, PUE, liquid cooling, and cabling. One of the package’s pillars is the Intelligent Computing DC Facility & Computing Cluster Integration Service. By using factory prefabrication and parallel construction, Huawei reduces the time required to install infrastructure. A project that normally takes 7 to 9 months now takes only 4 to 6 months.
In addition, the company deploys clusters of 1,024 nodes in 15 days, including automated deployment and stress testing. To date, Huawei has delivered more than 130 AI data center projects worldwide.
At MWC, Huawei also presented the Atlas 950 AI SuperPoD, a system with up to 8,192 Ascend NPUs that competes with Nvidia and AMD.
AI models quickly in production
The second pillar is the AI Computing Enablement and Optimization Service. Huawei claims to have adapted more than 150 industry-standard models, covering 90 percent of the most important customer scenarios. With a knowledge base of more than 10,000 expert cases, adapting and rolling out new models now takes five days. Thanks to seven-layer performance optimization, the platform delivers 30 percent higher throughput for typical AI workloads.
With the introduction of this platform at MWC 2026, Huawei is targeting customers who want to rapidly scale up their AI infrastructure. The company has already laid the foundation for AI transformation at European companies and is now continuing along this path with a globally focused service package.