The artificial intelligence infrastructure market continues to gain momentum. Nvidia, together with Australian Firmus Technologies, has announced a large-scale deployment of up to 170,000 AI accelerators at the Firmus campus in Batam. Deliveries are scheduled for 2027–2028 and will cover advanced platforms Grace Blackwell, Vera Rubin, and Vera.
Financial Parameters and Scale
Firmus expects to receive between $25 and $30 billion under capacity guarantee agreements during the first six years of the deal. The Batam campus, being developed with the participation of DayOne, is positioned as one of the largest AI infrastructure projects in the Asia-Pacific region. This is not just an expansion — it is a strategic move that lays the foundation for the next generation of computing.
Technology Foundation
The Grace Blackwell, Vera Rubin, and Vera platforms represent Nvidia's flagship solutions for high-performance computing and neural network training. Accelerators based on these architectures are capable of processing trillions of parameters, which is critically important for modern AI models. Deploying such a volume of equipment in a single location is a rare event even for the global market.
My Expert Analysis
The deal between Nvidia and Firmus is a clear indicator that the race for computing power is reaching a new level. $25–30 billion over six years is not just an investment, but a bet that demand for AI infrastructure will grow exponentially. However, the key question is whether Firmus can ensure stable power supply and cooling for 170,000 accelerators, as this will require gigawatt-level capacity. If the project is implemented on schedule, Batam will become a new AI gravity center in Asia, changing the balance of power in the cloud computing market.