SCX.ai partners with DDN to expand sovereign AI inferencing cloud in Australia

SCX.ai and DDN have teamed up to scale what they call Australia’s largest sovereign AI inferencing cloud, pairing ASIC-accelerated infrastructure with DDN’s Infinia platform. The move comes days after SCX.ai’s ASX debut and is aimed at meeting rising demand for domestic AI capacity while keeping models and data inside Australia.

Why it matters: - The partnership is designed to expand Australia’s domestic AI capacity at a time when enterprises, government agencies and research institutions are looking for local infrastructure. - SCX.ai says the combined stack will support higher-volume inference, lower cost per token and keep data, workloads and models within Australia. - The deal also targets one of AI’s biggest infrastructure constraints: power, cooling and data-layer efficiency.

What happened: - SCX.ai and DDN announced a partnership on Aug. 25, 2026, to expand Australia’s largest sovereign AI inferencing cloud. - The announcement came after SCX.ai listed on the Australian Securities Exchange on Aug. 21 following a fully underwritten A$40 million IPO. - SCX.ai is now positioned as Australia’s first ASX-listed pure-play sovereign AI inference infrastructure company. - The partnership combines SCX.ai’s ASIC-accelerated infrastructure with DDN’s Infinia data intelligence platform.

The details: - The combined system is built as a secure, high-performance, multi-tenant AI factory for Australian users. - DDN Infinia is intended to support massive concurrency and high-throughput inferencing demands. - The platform is designed to deliver sub-millisecond latency and up to 27x faster KV cache loading. - The lower-latency data layer is meant to reduce I/O stalls and improve hardware utilization. - SCX.ai says the setup can support large context windows and agentic AI workloads without compromising speed, privacy or regulatory compliance. - SCX.ai’s first sovereign AI node is already operating at the Equinix SY5 data centre in Sydney. - The company is accelerating deployment of Node 2, which is targeted to be operational by the end of 2026. - SCX.ai is also planning additional nodes as it builds a national sovereign AI inference network. - The expanded cloud will support bespoke enterprise workloads and Project MAGPiE, SCX.ai’s sovereign large language model tuned for Australian cultural and commercial context. - SCX.ai disclosed testing ahead of its ASX listing showing its SambaNova-based infrastructure delivered roughly 2.5x to 5.6x performance per watt versus GPU-based systems across selected stable inference workloads, depending on model and configuration. - SCX.ai’s infrastructure uses purpose-built SambaNova AI processors. - The architecture is designed to produce more AI compute from existing commercial data-centre power and floor space. - The system uses air cooling instead of the large water requirements common in many AI data centres.

Between the lines: - The partnership shows SCX.ai trying to build a domestic AI stack that competes on performance, not just data residency. - DDN’s role points to the data layer as a key bottleneck in inference, especially as workloads become more concurrent and latency-sensitive. - The focus on power efficiency and air cooling suggests SCX.ai is trying to scale within Australia’s existing data-centre constraints rather than wait for purpose-built mega-facilities. - The timing, right after the ASX debut, signals an effort to turn public-market access into faster infrastructure rollout.

What’s next: - SCX.ai will continue deploying Node 2 toward a late-2026 launch. - Additional nodes are expected as the company extends its sovereign AI inference network nationwide. - The company will likely use the expanded capacity to onboard more enterprise and government workloads and to further develop Project MAGPiE.

The bottom line: - SCX.ai and DDN are betting that Australia can build sovereign AI infrastructure that is both local and high performance, with inference economics and energy use at the center of the pitch.