NVIDIA and Firmus Partner to Build Large-Scale AI Infrastructure in Batam, Indonesia
June 29, 2026
NVIDIA and Firmus Partner to Build Large-Scale AI Infrastructure in Batam, Indonesia
Firmus Technologies has entered into a strategic partnership with NVIDIA to deploy up to 170,000 AI accelerators across the Grace Blackwell, Vera Rubin, and Vera platforms through 2027 and 2028. The deal is designed to power Firmus' AI Factory campus in Batam, Indonesia, a facility being developed in collaboration with Singapore-based digital infrastructure platform DayOne. Under the agreement, Firmus will procure NVIDIA infrastructure and resell NVIDIA-powered cloud services to AI-native companies, enterprises, and independent software vendors.
The financial scope of the partnership is substantial. Firmus has secured committed offtake agreements worth between US$25 billion and US$30 billion during the first six years, based on existing customer commitments. This scale reflects the broader industry trend identified by McKinsey, which estimates that meeting global AI-related data center demand alone will require US$5.2 trillion in investment by 2030, driven by projected AI capacity needs of 156 GW.
The partnership operates under a revenue-sharing and credit-support model. Under this structure, NVIDIA receives product revenue and a share of the cloud revenue generated from the supported capacity, while Firmus sells cloud services to end users. Tim Rosenfield, co-CEO of Firmus Technologies, said the model is designed to give AI-native companies, enterprises, and ISVs access to large-scale AI compute without requiring them to contract directly for capital-intensive infrastructure. "AI-native companies need access to scalable, energy- and cost-efficient compute infrastructure to compete globally," Rosenfield said, adding that the partnership offers customers greater certainty, scale, and flexibility.
The Batam campus will integrate NVIDIA's DSX full-stack AI factory platform with Firmus' HyperCube platform, a liquid-cooled AI Factory architecture developed in Australia and co-designed to NVIDIA DSX blueprints. Liquid cooling is critical for dense accelerator systems, which generate more heat than conventional server environments. NVIDIA notes that its liquid-cooled GB200 NVL72 rack design increases compute density, reduces floor space requirements, and supports high-bandwidth, low-latency GPU communication. The company also claims that the GB200 platform delivers 25 times more performance at the same power compared to H100 air-cooled infrastructure.
The combined platform aims to improve tokens per watt—a measure of AI model output per unit of energy—and reduce cost per token for customers using the Batam facility. The companies said the integration will allow the campus to operate under a common architecture, with the goal of improving deployment speed, energy efficiency, and resiliency. The Batam project follows other Firmus AI infrastructure deployments, including a long-term contract signed in March 2026 for dedicated AI infrastructure capacity at Project Southgate in Australia, which involves approximately 18,400 NVIDIA GB300 GPUs. This agreement expands Firmus' planned AI Factory capacity across the Asia-Pacific region.
Source: crnasia
