Firmus and Meta have announced strategic AI infrastructure agreements that will provide GPU compute capacity at Firmus AI factories in Southeast Asia. The agreements, announced on September 29, 2026, include contracted capacity and additional expansion options built on NVIDIA’s full-stack DSX platform. The collaboration is intended to support Meta’s AI research, model development, and training initiatives, and it builds on an existing relationship between the two companies in Australia.
Strategic Partnership Details
Meta is already an existing Firmus customer in Australia and currently uses NVIDIA GB300 NVL72 systems at Firmus’ Melbourne AI factory. This facility hosts the largest deployment of NVIDIA Blackwell Ultra systems in the Southern Hemisphere, which highlights the scale of the existing infrastructure. The new agreements extend the relationship into Southeast Asia and provide the infrastructure scale needed for demanding AI workloads.
Regional AI Factory Expansion
The agreements cover GPU compute capacity at Firmus AI factories in Southeast Asia and include both contracted capacity and additional expansion options. This builds on Meta’s existing presence in Australia and extends the relationship into a region with growing demand for AI infrastructure. The capacity will support AI research, model development, and training at significant scale while providing room for future growth.
HyperCube Platform and Technology
Firmus designs, manufactures, and operates AI factories that support large-scale AI workloads. Its HyperCube platform, co-designed with NVIDIA, combines liquid cooling, power, and mechanical systems into factory-built modules that can be produced at scale and deployed across the region. The latest generation supports NVIDIA GB300 and Vera Rubin NVL72 systems and Spectrum-X networking, giving customers a path to next-generation AI performance.
Energy Efficiency and Integrated Design
Firmus brings together computing, cooling, and power systems with the goal of delivering more compute from each watt of energy. By designing computing, networking, cooling, and power systems together, the company aims to deploy AI factories faster and use energy more efficiently. This integrated approach is central to supporting large-scale AI workloads across the Asia-Pacific region and meeting demanding operational requirements.
Strategic Importance for Meta
The partnership gives Meta access to infrastructure that can support some of the world’s most demanding AI workloads. Meta views the collaboration as central to building the next generation of AI across Asia-Pacific and expanding the foundation for AI research and model development. The agreements also provide flexibility through expansion options that can accommodate future growth in model training and AI capabilities.
Executive Perspectives
Meta Vice President of Engineering, Infrastructure, Gaya Nagarajan said, "We see Firmus as a long-term strategic infrastructure partner as we build for the next generation of AI across Asia-Pacific and expand the infrastructure supporting our AI research and model development efforts, the foundation of personal superintelligence." She emphasized the importance of infrastructure for these efforts. Firmus Co-Founder and Co-CEO Tim Rosenfield said the company is building for partners whose models demand exceptional performance and efficient energy use.
The announcement reflects a deepening alignment between Meta and Firmus around scalable, energy-efficient AI infrastructure. Firmus’ AI factory build-out across Asia-Pacific is positioned to support Meta’s current and future AI research and model development requirements. The partnership underscores a shared focus on delivering higher compute performance per watt while supporting next-generation AI capabilities across the region.