AMD, Cisco, and HUMAIN launch live AI infrastructure in Saudi Arabia

AMD, Cisco, and HUMAIN Launch Live AI Infrastructure in Saudi Arabia In a significant milestone for Saudi Arabia’s digital economy, AMD, Cisco, and HUMAIN announced that their joint venture’s artificial intelligence infrastructure is now live in the Kingdom. Unveiled on August 31, 2026, during the LEAP and DeepFest 2026 events in Riyadh, this production deployment marks a concrete step toward providing scalable, sovereign AI compute capabilities. The infrastructure, powered by AMD Instinct MI355X GPUs, AMD EPYC CPUs, and Cisco Silicon One-based AI networking, is designed to serve HUMAIN customers in Saudi Arabia and the broader region. This operational deployment distinguishes itself from generic executive previews by offering immediate, tangible resources for AI model training and inferencing. The newly live AI infrastructure relies on a robust combination of hardware technologies to deliver high-performance compute capabilities. At its core are the AMD Instinct MI355X GPUs, which provide the processing power necessary for demanding AI workloads. These graphics processing units are specifically engineered to handle the complex calculations required for machine learning and deep learning applications. They are complemented by AMD EPYC CPUs, ensuring efficient data handling and system management. The integration of these central processing units with the GPUs creates a balanced and powerful compute environment capable of tackling a wide range of artificial intelligence tasks. On the networking side, the deployment utilizes Cisco Silicon One-based AI networking, integrated with Cisco 800G optics. This combination aims to deliver the high bandwidth and low latency required for distributed AI training and large-scale inferencing tasks. The Cisco Silicon One architecture provides a unified, programmable networking solution that can scale to meet the demands of modern data centers. By incorporating 800G optics, the infrastructure ensures that data can move quickly and efficiently between compute nodes, minimizing bottlenecks and maximizing overall system performance. By integrating these advanced components, the joint venture provides a production-ready environment for HUMAIN’s customers. The focus on sovereign AI compute is particularly relevant, as it allows organizations within the Kingdom to develop and deploy AI models while keeping their data localized. This approach aligns with broader national strategies to build self-sufficient technological ecosystems and reduce reliance on external infrastructure providers. The availability of such high-performance resources within the country is expected to accelerate the adoption of artificial intelligence across various sectors, from healthcare and finance to energy and manufacturing. While the current deployment represents a significant achievement, the partnership has outlined ambitious plans for future expansion. Beginning in 2027, the joint venture intends to deploy up to 250 megawatts (MW) of additional AI infrastructure. This next phase will introduce the AMD Instinct MI400 Series GPUs, further enhancing the system’s computational capabilities. The introduction of the MI400 Series is expected to bring improvements in performance, efficiency, and scalability, allowing the infrastructure to handle even more demanding workloads. Additionally, the expansion will leverage AMD ROCm open software, providing developers with the tools needed to optimize their applications for the new hardware. The ROCm platform offers a comprehensive suite of libraries, compilers, and tools designed to facilitate the development of high-performance computing and artificial intelligence applications. This planned 250 MW expansion is part of a longer-term strategy to scale the infrastructure significantly. The partnership remains on track to reach a total capacity of up to 1 gigawatt (GW) of AI infrastructure by 2030. Such a massive scale underscores the commitment of AMD, Cisco, and HUMAIN to supporting the growing demand for AI compute resources in the region. A 1 GW capacity would position the infrastructure as one of the largest and most powerful artificial intelligence computing facilities globally. The phased approach allows for the integration of next-generation technologies as they become available, ensuring the infrastructure remains active of AI development. It also provides a clear roadmap for customers, allowing them to plan their artificial intelligence strategies with confidence in the availability of future compute resources. The launch of this live AI infrastructure has practical implications for various stakeholders in Saudi Arabia and the Middle East. For customers, it offers access to localized, high-performance compute resources, which are critical for training complex AI models and running inference workloads efficiently. This localized approach not only improves performance through reduced latency but also addresses data sovereignty concerns, a growing priority for