Building the future of sovereign AI.

SCX Labs is the global R&D layer of Australia's first sovereign AI datacenter, SCX.ai. We research, build, benchmark, and deploy AI systems that are open, environmentally responsible, and vendor-agnostic.

Explore SCX ACE View MAGPiE
Compare benchmark work ->

SCX ACE creates a fourth path for enterprise AI adaptation

ACE builds a compact, explainable playbook of operational rules that models can use at inference time.

SCX MAGPiE: reasoning with Australian context

Australia 1k and local alignment evaluation

Environmental commitment

Useful AI per unit of energy.

SCX Labs treats sustainability as an architecture problem: efficient silicon, right-sized inference, measurement, and deployment choices that reduce wasted tokens, water, and power.

ASIC-first

Inference-first silicon

Route workloads toward efficient accelerators for useful work per watt.

Air-first

Cooling strategy

Moderate-density deployments reduce dependence on water-intensive cooling.

Measured

Per-token reporting

Track kWh/token, L/token, and gCO2e/token at workload level.

Updates

Release notes and lab news.

Placeholders for upcoming papers, release notes, benchmark reports, and partnership announcements.

Research

SCX ACE release note

Framework notes, demo results, and enterprise adaptation examples.

-> Read more Benchmark

Australia 1k report

Evaluation methodology for Australian cultural, language, history, and media alignment.

-> Read more News

SCX MAGPiE technical note

How Australian context can be embedded into model reasoning rather than only prompted.

-> Read more