Saudi Arabia's AI compute expansion gains operational foothold with Cisco and AMD partnership

A production system delivering GPU-intensive compute capacity has begun operation in Saudi Arabia through a collaboration between Cisco Systems, Advanced Micro Devices, and local partner HUMAIN. The infrastructure employs AMD's latest processor technology alongside Cisco's networking platform to enable businesses to access AI computation on a service basis, supporting both model development and deployment tasks. The three organizations have committed to scaling this capacity substantially, with plans to add up to 250 MW of resources beginning in 2027 and achieving 1 GW total by 2030, assuming sufficient customer interest materializes.

1. The operational deployment integrates AMD Instinct MI355X GPUs with AMD EPYC processors, connected through Cisco Silicon One networking infrastructure
2. Cisco's N9000 Series platform leverages 800G optics technology to create an AI-optimized network connecting the GPU resources with reduced latency across large installations
3. HUMAIN now offers GPU-as-a-service capabilities on this infrastructure, enabling customers to run AI model training and inferencing workloads
4. A joint venture among Cisco, AMD and HUMAIN plans to expand capacity to 250 MW, with initial rollout commencing in 2027 and initial supplies arriving in the second half of 2027
5. The partnership targets 1 GW of aggregate AI infrastructure capacity by 2030, with expansion dependent on securing sufficient customer commitments