AMD and HUMAIN Announced $10 Billion AI Partnership
The joint initiative targets the deployment of 500 MW of AI computing capacity over the next five years.
Updated on Oct. 1, 2026 in Data Centers
Live Poll
Do you support large-scale international partnerships for building national AI infrastructure?
AMD and Saudi-based HUMAIN have announced a $10 billion investment to build an open AI infrastructure ecosystem. The initiative focuses on deploying 500 MW of AMD-powered compute capacity across the United States and Saudi Arabia.
Why it matters
This partnership seeks to establish an open, scalable, and cost-efficient AI infrastructure by integrating AMD’s compute portfolio with dedicated hyperscale data center construction. The effort aims to accelerate regional AI development through specialized power systems and global fibre connectivity.
The project includes a 100 MW data center development at Oxagon, part of a larger 360 MW facility. This capacity will utilize the ROCm open software ecosystem alongside AMD’s AI compute hardware.
The players
AMD
A semiconductor company that produces high-performance processors and AI-optimized graphics cards for data center and consumer markets.
HUMAIN
A company backed by the Saudi Public Investment Fund focused on the development of hyperscale data center delivery and energy infrastructure.
DataVolt
A developer specializing in sustainable data center infrastructure and high-capacity site management.
The details
The infrastructure relies on the integration of AMD’s AI compute portfolio with the ROCm open software ecosystem, a set of tools designed to enable high-performance computing on AMD graphics processing units (GPUs). HUMAIN, a Saudi Public Investment Fund (PIF) entity, is responsible for building hyperscale data centers, integrating sustainable power systems, and managing global fibre interconnects to support the workload. A specific project at Oxagon serves as an early anchor for this infrastructure, where a 100 MW block of power is being dedicated to AI compute within a 360 MW total site capacity.
Timeline
500 MW of AI computing capacity will be deployed over the next five years.
The Tech Race
This initiative follows an aggressive trend among sovereign-backed entities to secure localized hyperscale compute capacity. It positions the region to compete with existing global data center hubs by building out indigenous power and connectivity networks.
This development primarily affects the availability of enterprise-grade AI compute resources for regional developers and global cloud providers. The physical build-out at Oxagon represents a significant expansion of regional high-performance computing infrastructure scheduled over the next five years.
The takeaway
This partnership highlights a shift toward sovereign-controlled AI compute hardware and infrastructure stacks. Watch for upcoming announcements regarding individual data center completion dates and specific power-utilization milestones for the Oxagon site.
Further reading
For more on the current expansion of global server capacity, visit our Data Centers section.
Live Poll
Do you support large-scale international partnerships for building national AI infrastructure?
