Overview
- The agreement announced on Sept. 8 makes OpenAI an anchor customer for dedicated compute from two Firmus sites in Malaysia and pushes Firmus’ total contracted capacity past 900 megawatts.
- Firmus will deploy NVIDIA’s next‑generation Vera Rubin processors and the DSX AI Factory platform integrated with its HyperCube liquid‑cooling system at the Malaysian sites.
- Firmus and OpenAI have not disclosed financial terms, how many megawatts OpenAI has reserved, exact site locations, or when the facilities will begin serving workloads.
- NVIDIA separately said it is working with eight Australian partners, including Firmus, on an effort that targets up to 2 gigawatts of AI capacity in Australia by 2027, highlighting a wider regional push for GPU‑heavy infrastructure.
- Malaysia’s fast buildout has prompted authorities to require secured power and water supplies and green compliance, which together with hardware delivery, grid connections and cooling needs are the main bottlenecks to bringing contracted capacity online.