Our Story
In 2025, the race for AI dominance heated up. Some nations introduced tight legislation, while others
stepped back from controls. Global borders were tested in terms of tariffs and penalties.
All of this sparked a deep concern that Australia was now at far greater risk of being shut out of access
to foundational AI, or worse, forced to utilise technology that no longer matched our values and beliefs.
For some time, our co-founders, Troy and Simon, had discussed the need for greater sovereignty in AI for Australia.
However, the existing foundational models were so large and so readily available for Australians that building a model
here seemed unnecessary.
But the world shifted – and so did we.
Troy and Simon decided to step forward and change the narrative for Australia. So, they set about building a research model to prove the concept and continue to the next step. A huge step for Australia.
What sovereignty truly is
Sovereignty is the ability to determine your own future without outside interference or coercion.
In AI, sovereignty requires all four key layers.
It starts at the core infrastructure. The first two layers are the hardware that provides AI processing power (compute infrastructure) and the data centres that house and support it.
To protect our data and comply with Australian laws and standards, data centres must be Australian-owned and located here. Similarly, the acquisition, deployment and tuning of AI compute infrastructure (like AI chips or GPUs) must be delivered by Australian-owned companies.
Australia has limited GPU infrastructure and, although this is changing, it still means that when we type a prompt into an AI system, the data is often processed offshore.
Global cloud providers may process AI data on servers located anywhere in the world. Often, these countries don’t uphold the strict privacy standards that Australia depends on.
That’s why a crucial part of sovereignty is ensuring that the compute used to process (inference) your AI happens right here onshore.
The third layer is the foundational AI models. Australia needs a range of foundational capabilities, and although some work has been done, we do not have a sovereign large language model (LLM).
LLMs underpin almost all modern AI interfaces, from chatbots to medical transcription and educational tools.
This is the layer we will solve. We will ensure that Australia has a model that is built with transparency and explainability, and uses ethically sourced data.
The final layer is AI applications. Australia has a myriad of innovative applications that are both commercially developed and built in-house by large organisations – creating incredible value for business, government and the population at large. Today, most of these still rely on offshore foundational models. We want that to change.
Without sovereignty across all four layers, Australia could be cut off from the foundational models we increasingly depend on or continue to use models that offer no transparency in how they were built.
Bringing all four layers under Australian control will allow us to raise productivity and ensure ongoing security in our nation.
Making it happen
We will only work with data centres
that are Australian-owned and controlled.
We will set up strategic alliances with
the world’s leading AI chip providers.
This infrastructure will be sourced,
fine-tuned and managed by an
Australian-owned company. We will
ensure our compute capacity meets
global standards for size and speed.
We will provide complete transparency
on what data Australia’s foundational
model is trained on, with
appropriate opt-out mechanisms.
Our foundational models will
ensure Australia’s world-leading AI
applications stay sovereign and
secure.
