Chipmaker Nvidia is teaming up with Firmus Technologies, an Australian startup, to build a massive artificial-intelligence data center in Indonesia, according to reporting from Yahoo Finance, Quartz, Reuters and others.
The scale is the headline number: the facility will house 170,000 Nvidia graphics processing units, or GPUs — the specialized chips that train and run modern AI systems. Firmus will deploy those chips and operate them as a cloud service.
According to The Times of India and reporting carried by MSN, the deal centers on making AI computing more affordable. The aim, per those reports, is to lower the cost of access to AI infrastructure for emerging and smaller companies that otherwise struggle to afford the hardware. Firmus, described by Quartz as an Australian startup and by The Times of India as Australia's Firmus Technologies, will run the Indonesian site and rent out its capacity.
GuruFocus framed the arrangement as Nvidia backing a Firmus AI cloud deal, signaling the chipmaker's continued strategy of seeding demand for its processors by supporting the cloud providers that buy them in bulk.
The sources reviewed here agree on the core facts — a 170,000-GPU build in Indonesia, a Nvidia–Firmus partnership, and a focus on affordable cloud access — but do not include a disclosed price tag, timeline, or specific GPU model.
Why it matters: large AI data centers have mostly clustered in the United States, Europe and a handful of wealthy hubs, so a project of this size in Indonesia signals that the race to build AI computing capacity — and the demand for Nvidia's chips — is spreading into fast-growing Southeast Asian markets.