What happened
Meta signed a lease for AI capacity from Firmus' new Southeast Asia data centers.
On September 29, 2026, Data Center Dynamics reported that Meta has signed a lease to obtain AI capacity from Firmus' upcoming data center deployments in Southeast Asia. The agreement builds on an existing partnership in which Meta already leases Nvidia GB300 NVL72 compute from Firmus' Melbourne data center in Australia.
Firmus is constructing "AI factories" with its partner DayOne, including a facility in Batam, Indonesia that will house 170,000 GPUs, and additional sites in Malaysia where OpenAI is also a customer. The Southeast Asian capacity destined for Meta will be based on Nvidia's full‑stack DSX platform and will use Firmus' HyperCube modular data center solution.
Meta’s vice‑president of engineering and infrastructure, Gaya Nagarajan, said the partnership is strategic for building the next generation of AI across the Asia‑Pacific. Firmus co‑founder Tim Rosenfield highlighted the need for high‑performance, energy‑efficient to support demanding AI workloads.
Source details: datacenterdynamics.com ↗
Why it matters
The deal expands Meta’s AI infrastructure in the Asia‑Pacific region and signals growing demand for dedicated AI‑focused data center capacity.
The agreement underscores the accelerating shift toward purpose‑built AI infrastructure, as major tech firms seek to secure dedicated rather than rely on generic cloud services. By securing capacity in Firmus' modular "AI factories," Meta can potentially reduce and improve energy efficiency for its large‑scale model training.
The deal also illustrates the growing role of specialized data center providers like Firmus in the AI supply chain, especially in emerging markets such as Indonesia and Malaysia. This diversification may influence competitive dynamics among cloud providers and impact regional data‑center investment trends.
From a broader industry perspective, the partnership highlights the importance of Nvidia's DSX platform as a de‑facto standard for high‑density AI workloads, reinforcing Nvidia’s position in the AI hardware ecosystem.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
In AI, what are a model's "parameters"?
What to watch next
Future capacity expansions, pricing terms, and the impact on Meta’s AI model training roadmap.
Whether Meta will expand the lease to additional Firmus sites or other regional providers as its AI models scale.
The pricing and contract terms, which have not been disclosed, could set benchmarks for future AI‑ leasing agreements.
Potential regulatory scrutiny or geopolitical considerations surrounding large‑scale AI deployments in Southeast Asia.