Nvidia-backed Firmus will build a 360MW AI data centre in Indonesia and expects $30 billion in offtake deals

The business move

Firmus Technologies, an Australian AI infrastructure startup valued at $5.5 billion, is breaking ground on a 360-megawatt data centre campus in Batam, Indonesia. This facility will be built through an eight-year partnership with Nvidia featuring Nvidia’s DSX AI Factory design. The project is also backed by Singapore-based DayOne, situating the data centre just off Singapore’s coast. Firmus expects the Batam campus to anchor a pipeline of AI computing contracts valued at $30 billion in offtake agreements over time.

Why it matters

Building a large-scale AI data centre outside the usual US, Chinese, and European hubs marks a significant strategic expansion for infrastructure capacity in Southeast Asia. Firms constructing generative AI models and running large compute workloads will find a new regional option closer to other Asian markets. The involvement of Nvidia adds technological credibility and specialized AI hardware as part of the deployment. With such a large power footprint dedicated to AI, the project signals accelerated AI adoption pressure on power grids and facility operators in emerging markets.

This kind of multi-year offtake deal also means Firmus and Nvidia are locking in long-term enterprise commitments, which lowers risk for investors and operators but raises the stakes for the facility’s uptime and efficiency. The scale and scope position Batam as a target cluster for AI compute demand moving forward.

Who gains and who gets squeezed

Regional AI developers, cloud operators, and AI-first enterprises benefit from lower latency and more proximate infrastructure options. Indonesia and Singapore’s economies may secure jobs and technical talent development around AI infrastructure. Nvidia gains by embedding its AI hardware portfolio within next-gen facilities, locking in demand and expanding its geographic footprint.

Traditional data centre operators and cloud providers outside Southeast Asia may face pressure as companies look for more specialized and regional AI infrastructure deals. Power generation and network providers on Batam will have to scale rapidly and maintain reliability under unprecedented load growth.

What to watch next

Monitor the actual build progress and the timeline for going live, as delays or cost overruns would test the partnership’s claims. Watch for public announcements of enterprise customers and AI model workloads that firm up the $30 billion offtake target. Regulatory and grid readiness in Indonesia and Singapore could become bottlenecks or points of friction.

Follow how Nvidia integrates its DSX AI Factory hardware into operations and whether this model shapes new AI data centre designs elsewhere. Finally, check how this influences competitors—both cloud giants and emerging AI infrastructure players—in Southeast Asia and beyond.

AI Quick Briefs Editorial Desk