March, 2026 | Partners
Firmus to Build Grid-Integrated AI Factory Software with NVIDIA
Software-defined, energy-aware AI infrastructure built on NVIDIA DSX Blueprint
San Jose, CA / Sydney
Firmus today announced the development and deployment of a software layer for grid-integrated AI factories built with the new NVIDIA DSX Blueprint. The effort supports a new class of AI factory architecture that aligns compute performance with grid conditions.
The collaboration introduces an orchestration software layer that integrates with NVIDIA DSX Blueprint, to coordinate compute clusters with energy networks. This software layer forms part of Firmus’s Model-to-Grid platform, which links workload behaviour with grid dynamics in real time. The Model-to-Grid approach builds on research and development initiated in 2020 and includes patented technologies focused on grid-aware orchestration.
“The AI factory of the future must be as intelligent about energy as it is about compute,” said Dr. Daniel Kearney, Chief Technology Officer, Firmus. “As platforms evolve, power and thermal characteristics change fundamentally. Our architecture is designed to understand those shifts and apply that intelligence along with NVIDIA DSX Blueprint deployments to operate infrastructure at scale while working seamlessly with grid systems.”
“AI workloads are scaling toward gigawatts, and their relationship with power delivery systems is one of the defining engineering challenges of our era. Software-defined coordination between compute and energy networks is essential — and Firmus’s Model-to -Grid approach advances that directly.”
The architecture will be initially rolled out across Project Southgate in Australia, where complex grid dynamics provide a technically rigorous deployment environment. Operations will be coordinated with AEMO and the National Electricity Market. The deployment focuses on regional areas where concentrated investment can support renewable energy zones and long-term job creation.
This initiative aims to help establish an operational software standard for NVIDIA DSX Blueprint AI factories and enable deeper integration of AI infrastructure with energy grids worldwide.
About Firmus Model-to-Grid
Firmus’s Model-to-Grid engineering approach treats the AI factory as a continuously optimised system — from model workload to grid connection. The platform understands the electrical and thermal signatures of AI workloads and uses that intelligence to coordinate and orchestrate compute, cooling and power in real time. The result is AI factories that operate as managed grid participants that align energy policy and stability requirements – delivering lower energy per token, higher GPU utilisation and infrastructure that works with the grid, not against it.
About Firmus
Firmus Technologies builds world- class, highly-efficient AI factories, u tilizing its proprietary technology to drive down total cost of ownership and
accelerate AI adoption. With offices in Australia and Singapore, Firmus partners with government, enterprise, education and the innovators of tomorrow to create sustainable, sovereign AI solutions.
Follow @firmustechnologies on Linkedin.
