UNSW Sydney's Katana HPC was chronically GPU-oversubscribed while its Intel CPU capacity sat idle. By migrating inference to HELIX CPU pods, UNSW freed GPU capacity for its highest-value work and avoided $1.5–4M in GPU CAPEX — with a sub-3-month payback and no new hardware.
The deterministic decision-transformer substrate paper is released.
ReadHELIX v1.7 (Build #58) goes live at UNSW — audited on the official vLLM GuideLLM benchmark: 940 ms p50 TTFT, 100% valid JSON, and 8.95–12.89× faster than stock Ollama.
ReadAfter a successful pilot on its Intel-based Katana HPC, UNSW Sydney adopts Verificate HELIX — our flagship sovereign-inference case study.
ReadHELIX v1.4 is deployed for the UNSW evaluation on Intel HPC.
The University of New South Wales starts a pilot of HELIX CPU inference on its Katana HPC.
Our inference-level hallucination-prevention method — a per-token truthfulness manifold — is released.
ReadVerificate becomes an IBM Business Partner, with access to IBM Fusion HPC infrastructure and Granite model integration.
Peer-facing research and audited technical reports. Measured results carry a M mark and are substantiated on our Legal & claims page.