Genecigs Corporate · AUS Data Centre & AI Centre
We design, construct, and operate high-density compute facilities across Australia and the Asia–Pacific — engineered around the AI accelerators you choose, from NVIDIA to AMD and beyond. Own the asset outright, or rent ready-to-run capacity by the rack.
Two ways to take delivery
Greenfield or retrofit. We handle site selection, electrical and mechanical design, structural fit-out, and commissioning — then hand over a fully tested, asset-on-your-books data centre or AI hall.
Skip the capex. Reserve power-ready racks or dedicated GPU clusters in our partner facilities, provisioned with the hardware you spec and billed on a term that fits your roadmap.
Vendor-neutral by design
We aren't tied to a single supplier. We source, integrate, and tune the accelerator platform that matches your workload, your budget, and your availability — and we tell you honestly when a cheaper option will do the job.
NVIDIA
GB200 NVL72, H200, H100, L40S — training & inference
AMD
Instinct MI300X / MI325X accelerators
Intel
Gaudi accelerators & Xeon host platforms
Selected deployments
A liquid-cooled AI hall in Singapore built to host edge-AI inference at scale. We delivered the power train, direct-to-chip cooling loop, and GPU cluster integration, then tuned the fabric for low-latency serving across the region. The site reached steady-state operation inside the planned window with no derating of design density.
Greenfield AI training facility with rear-door heat-exchange halls and an InfiniBand back-end fabric, handed over as a client-owned asset.
Dedicated NVIDIA L40S inference nodes provisioned in a partner facility on a reserved annual term — live in weeks, not quarters.
What ships in 2026
Direct-to-chip cooling plumbed from day one, so racks above 100 kW don't need a retrofit later.
Rack-scale reference builds for the newest NVIDIA platforms, fully cabled and burn-in tested.
On-site solar, battery buffering, and demand-response options to take pressure off the connection.
Per-rack power, thermal, and PUE metrics streamed to your team in real time.
Follow-the-sun NOC coverage across Australian and Southeast Asian sites.
Designs documented for ISO 27001 and SOC 2 audit scopes from the outset.
Common questions
Both. We build owned facilities end-to-end, and we also offer ready-to-run rented capacity — by the U, the rack, the cage, or as a full dedicated GPU cluster. Many clients start by renting and move to a build once their demand is proven.
We're vendor-neutral. NVIDIA is the most common request, but we also integrate AMD Instinct and Intel Gaudi platforms, and we'll recommend the option that genuinely fits your workload and budget rather than the most expensive one.
Standard air-cooled halls handle typical enterprise racks comfortably. With rear-door heat exchange we support higher densities, and with direct-to-chip liquid cooling we support AI racks above 100 kW — up to around 132 kW per rack on supported platforms.
It depends on scale and whether it's greenfield or a retrofit. Pod-scale fit-outs can move in a few months; multi-megawatt campuses run longer. Rented capacity is the fastest path — often live in weeks. We give you a committed program before any work starts.
Our home base is Australia, with delivered projects across the Asia–Pacific including Singapore, Jakarta, and New South Wales. If you need capacity in a specific jurisdiction, ask — we'll tell you what's feasible.
Yes. We document facility designs and operational controls so they map to ISO 27001 and SOC 2 audit scopes, and we work alongside your compliance team to keep the evidence audit-ready.
Where the site allows, we design in on-site solar, battery buffering, and demand-response so facilities lean on the grid less at peak. Energy efficiency is tracked live through per-hall PUE telemetry.
Whether you're planning a build or just need capacity online quickly, send us your power, density, and timeline targets and we'll come back with options.
