SCX.ai and DDN have formed a partnership to expand Australia’s sovereign AI inference cloud, integrating SCX.ai’s domestic infrastructure with DDN’s data platform to keep AI workloads and data onshore for Australian enterprises, government agencies, and research institutions.

SCX.ai already runs its first sovereign AI node at Equinix SY5 in Sydney, with a second node planned to be operational by year’s end and additional nodes slated for nationwide expansion.

The partnership positions onshore inferencing as a core value driver to protect data sovereignty, support real-world applications, and deliver cost-effective, energy-efficient AI within Australia.

No external information is introduced; all points derive from the provided article content.

Industry leader warns against regulatory overreach and advocates balancing innovation with prudent governance, aiming for end-to-end sovereign control and onshore value capture across the AI stack.

The deal is framed as an economics play focused on ASIC-style performance per watt and a dedicated data layer to optimize utilization and cost per token onshore.

The company argues that sovereign control of data and value creation is essential for Australia’s economy and exchange rate, warning against over-reliance on foreign providers and offshore leakage.

CEO David Keane emphasizes reducing data sent overseas, advocating sovereign Australian alternatives that minimize energy and water use while maintaining data control.

Keane envisions Australian MSPs and integrators owning and orchestrating the AI stack, with hyperscalers as just one component of a broader, locally managed solution.

Both SCX.ai and DDN intend to expand sovereign capabilities with a focus on domestic control and compliance for government and enterprise workloads.

Australia’s sovereignty appeal centers on data residency and concerns about the US CLOUD Act, positioning SCX as a domestic alternative for sensitive workloads.

SCX.ai’s architecture relies on SambaNova SN40L processors with air cooling to maximize compute per watt and avoid water-intensive cooling, delivering higher efficiency and 2.5x to 5.6x performance per watt versus GPU-based systems for certain workloads.