Government, regulated enterprise and AI builders need GB300-class compute on Australian soil, under Australian control, assessed for the workloads that stay onshore. TMkNEO is the sovereign cloud, built on AI factories we develop and own.
Coming soon. Register your interest to reserve a place in the queue, and we will come back to you with availability.
Sovereign capacity starts with owning the asset it runs on. We develop and own the AI factory, so the sovereignty is built into the compute, on Australian soil under Australian control.
GB300 NVL72, liquid-cooled and factory-integrated, built to NVIDIA reference architecture with tier-1 partners, orchestration included. Choose the tier that fits the workload, from on-demand experiments to cleared, classified production.
Spin up GB300-class GPU by the hour and scale down when the run is done. For experiments, bursts and fast iteration.
Multi-year committed capacity at a lower rate, with priority access. For training runs and production inference you can plan around.
Australian-owned and cleared for OFFICIAL: Sensitive, PROTECTED and above. The tier for the workloads that stay onshore.
The data and serving stack runs inside the same sovereign boundary as the compute, so your data sits next to the GPU with zero egress. Six managed lines, run for you.
Hot NVMe, warm and cold tiers, with zero egress. Your data lives next to the compute.
Serving and MLOps on managed Kubernetes, with the control plane handled for you.
Retrieval over your sovereign knowledge base, kept onshore inside the boundary.
Streaming data pipelines, managed and scaled to the workload.
Transactional and vector data in one managed database.
Data engineering and analytics at factory scale.
Every service runs on the same sovereign GPU, inside the same sovereign boundary, under Australian control.
Private, high-throughput interconnect from your network into NEO, for data-gravity workloads that move at scale.
Dedicated interconnect for government and cleared environments, on the cleared-channel process.
Sensitive and classified data, OFFICIAL: Sensitive, PROTECTED and above, hosted on Australian-owned GB300 infrastructure. We intend to seek IRAP assessment at PROTECTED. Air-gapped and cleared options are handled on the cleared-channel process, so onboarding fits your assurance requirements from the first conversation.
Classified and sovereign workloads on Australian-owned, onshore compute.
Fraud, risk and RegTech models on onshore, low-latency capacity.
Clinical models with patient data kept onshore inside the boundary.
Network-scale ML and automation on sovereign GPU.
Sensor and operational data, modelled at scale.
Training and inference capacity for builders, from local teams to frontier labs.
Your state's registers and AI, live in Australia under your law. Estonia and Luxembourg proved the model in 2018.
Yes, and that is the point of this page. Tell us the workload, the scale, the timeline and the assurance you need, and we will come back with the tier that fits and reserve your place in the queue. Government and cleared workloads onboard on the cleared-channel process.
One short form. We come back with the tier that fits and your place in the queue.
Government and cleared workloads onboard on the cleared-channel process.
We use these details only to respond to your enquiry. See our privacy notice. Prefer email? mike@tokenomik.com.