Why Peregrine
Australian infrastructure for the AI era.
AI infrastructure is becoming critical national digital infrastructure. Peregrine is building the capability to assess, build, operate and supply it — in Australia.
Our thesis
Compute, energy and sovereignty are converging.
AI infrastructure is becoming critical national digital infrastructure. Australia needs additional GPU capacity to support its research institutions, enterprises, start-ups and public sector — and that capacity increasingly needs to be planned alongside power.
Peregrine combines GPU compute, distributed physical infrastructure, renewable-energy opportunities, software orchestration and secure workload delivery into one offering: one partner from the GPU to the model endpoint.
Peregrine Compute is a member of the NVIDIA Inception Program.
The compute challenge
AI demand is growing faster than traditional infrastructure.
Public cloud remains an excellent fit for many AI workloads, and Peregrine is designed to work alongside AWS, Azure and Google Cloud rather than replace them. But organisations increasingly encounter constraints that a complementary Australian infrastructure partner can help address.
GPU availability
Supply of current-generation accelerators varies by architecture and timeframe.
Cloud GPU costs
Consumption pricing is flexible, but sustained workloads can be hard to forecast.
Unpredictable capacity
Teams cannot always secure GPUs when a project needs them.
Data sovereignty
Many organisations need data and models to remain in Australia.
Research queue times
Shared institutional clusters can leave researchers waiting.
Capital requirements
Owning infrastructure requires upfront investment and specialist skills.
Power availability
GPU densities place real demands on power and cooling.
Scaling inference
Production models need capacity that grows with usage.
Differentiation
GPU capacity is only one part of production AI infrastructure.
A GPU cloud provides capacity. Production AI also needs architecture, networking, storage, scheduling, model serving, monitoring, MLOps and capacity planning — and someone accountable for all of it.
Peregrine brings planning, infrastructure, platform operations and GPU capacity together, so customers can engage at whichever layer they need.
GPU cloud
CAPACITY
Peregrine
PLANNING + INFRASTRUCTURE + PLATFORM OPERATIONS + GPU CAPACITY
Centralised + distributed
Two architectures, used where each fits.
Centralised, high-density facilities suit tightly coupled training where GPUs must communicate at very high bandwidth. Distributed GPU locations are useful for a different set of workloads. Distributed nodes do not replace hyperscale data centres; they extend where and how capacity can be delivered.
CENTRALISED
High-density, tightly coupled training clusters with high-speed interconnect, planned power and cooling.
DISTRIBUTED
- Inference
- Fine-tuning
- Batch processing
- Rendering
- Computer vision
- Distributed research
- Edge workloads
- Capacity expansion
Sovereign AI
Keep Australian AI workloads closer to home.
Infrastructure can be designed to support customer security, governance and compliance requirements.
- Data and models can remain in Australia
- Australian-hosted GPU infrastructure
- Dedicated and isolated capacity options
- Local engineering and support
- Reduced dependence on offshore capacity
- Alignment with Australian data-residency expectations
- Infrastructure designed around governance requirements
- Options for government, research and regulated sectors
Six reasons
Infrastructure designed around the workload — not the other way around.
Australian Infrastructure
GPU capacity hosted in Australian data centres, with options for data to remain onshore.
Flexible Capacity Models
On-demand, reserved, dedicated and private cluster models, matched to how your workloads actually run.
Dedicated GPU Options
Single-tenant servers and private clusters for organisations that need isolation and consistent performance.
Distributed Architecture
ComputeGrid is being developed to connect capacity across data centres and other suitable sites.
Energy-Aware Infrastructure Strategy
Planning that considers power availability and renewable-energy-enabled locations as capacity grows.
Technical Partnership
Engineers who work with your team from planning through commissioning and ongoing operation.
Let's design your compute environment.
Tell us about your workloads and deployment preferences. A compute specialist will review your requirements and discuss the appropriate infrastructure model.