Australian AI infrastructure company Firmus Technologies has struck a strategic partnership with Nvidia, the chipmaker at the center of the global artificial-intelligence boom.
According to InnovationAus, Firmus said on Monday it had signed the deal to help emerging AI firms get more cost-effective access to Nvidia's sought-after chips. Securing affordable compute power has become one of the biggest hurdles for AI startups, so a pipeline to Nvidia hardware is a meaningful advantage.
A central piece of the partnership is a major new data center. The Straits Times reports that Firmus and Singapore-based DayOne will develop the facility in Batam, Indonesia, as part of an eight-year agreement with Nvidia. The project has been billed by Futu (富途牛牛) as the basis for a Southeast Asian "AI factory" — industry shorthand for large-scale facilities built specifically to train and run AI models.
The ambitions are costly. The Australian Financial Review reports that Nvidia's enthusiasm for Firmus has helped spur a spending plan worth more than $20 billion. Per the AFR, Firmus maintains the new development will not slow its existing Australian rollout, though the publication notes the company will need to raise a great deal of money to deliver on both.
The deal also reflects how AI infrastructure is spreading beyond traditional U.S. and Chinese hubs, with Southeast Asia emerging as a contested region for data-center investment. TradingView and other outlets framed the agreement primarily as an "AI access" arrangement, underscoring that the value lies as much in who gets to use the chips as in the buildings themselves.
Why it matters: control over AI computing capacity is shaping which companies and regions can compete in the technology's next phase, and a multibillion-dollar Nvidia-backed buildout signals that the race for that capacity is now firmly global.