CAIO Connect.
Design · Build · Power · Scale
For enterprise and government buyers of dedicated AI compute: Australia's AI ambitions run through GPU supply chains that mostly sit offshore.
Sovereign ambition, offshore compute.
A closed-door discussion for infrastructure and government technology leaders on what it takes to build AI capacity locally, at the scale actually required.
Australia's AI ambitions depend on compute capacity, but most AI infrastructure investment and GPU supply chains run through offshore hyperscalers, leaving local enterprises and government dependent on overseas providers for sovereign workloads.
Firmus has outlined more than $70 billion in planned Australian capacity, having raised over $11 billion in combined equity and debt including a round with NVIDIA as an investor. Its modular AI Factory design is codesigned across compute, network, power and cooling to scale across GPU generations without retrofit.
What we'll be exploring.
Talking points on the table include, but are not limited to:
The offshore dependency
What sovereign workloads actually require, and where the current supply chain breaks that.
The scale of the build
$70 billion planned, $11 billion raised, and NVIDIA on the cap table.
Modular by design
Built to scale across GPU generations without rebuilding the facility each time.
Power as the real constraint
A 600 MW energy agreement in South Australia, and why energy decides the timeline.
APAC Series. Six cities.
Brought to you by Firmus and Innovatus Media.
Firmus designs and operates AI Factories, modular, energy-efficient, GPU-dense infrastructure engineered from the ground up for AI training and inference, anchoring Australia sovereign AI infrastructure buildout.
Your session host
Each city session is chaired by an Innovatus Media host alongside a Firmus speaker local to the region.
Host name to be confirmed per city.
Save your seat for CAIO Connect.
This is an invite-only executive roundtable. Register your interest and we'll confirm your city, venue, and date as they're locked in.