Data Centre Colocation
Rack space and power for high-density deployments, in established facilities across the region.
- High-density racks, air or liquid cooled
- Cross-connects and carrier access
- Flexible terms from a few racks upward
Products
Everything your AI workloads run on
Start with one service or combine them. We size each deployment to the workload and the timeline, and stay vendor-neutral on the underlying providers.
AI infrastructure
GPU Compute
Dedicated GPU clusters for training and inference, on reserved or term-based capacity.
- Bare-metal clusters sized to your model
- Capacity sourced across providers
- Planning through to ready-for-service
Token Service
Access leading AI models through one API and pay only for the tokens you use.
- Pay-as-you-go, billed per million tokens
- Choice of open and commercial models
- Usage reporting by project or team
Cloud & Hybrid
Public cloud resources combined with dedicated infrastructure, managed as one environment.
- Public cloud account setup and billing
- Hybrid links between cloud and colocation
- Ongoing cost and capacity reviews
Data, network & security
Data & Analytics
Turn operational and usage data into dashboards and reports your teams can act on.
- Data pipelines and warehousing
- BI dashboards and scheduled reports
- Usage and cost analytics for AI and cloud
SD-WAN
Software-defined networking that links offices, data centres and cloud over the best available path.
- Centrally managed policies across sites
- Multiple links with automatic failover
- Traffic priority for business applications
Security
Protection for your network, applications and data, from the edge to the workload.
- Firewall and secure access for remote users
- DDoS and web application protection
- Encryption in transit and at rest
Facilities
Built on established data centres
We place your infrastructure in professionally operated facilities with redundant power, cooling and connectivity, chosen to match your location and density needs.
Token Service
Use AI models without running the hardware
Models read and write text in small units called tokens. With our Token Service you call models through a single API and pay for the tokens processed, with no GPUs to buy or manage.
Pay for what you useNo upfront commitment. Billing follows input and output tokens.
One API, several modelsSwitch between models as your use case changes.
Grow into dedicated capacityWhen volume is steady, move to reserved GPUs with the same team.
Network & Security
Connect every site securely to your AI and cloud
AI workloads depend on reliable, protected links between users, offices, data centres and cloud. We design and deliver the network and security around your infrastructure.
SD-WAN across sitesOne managed network across offices, data centres and cloud.
Secure accessFirewall and zero-trust access policies for users and applications.
Protection at the edgeDDoS and web application protection for public-facing services.
Getting started
What we need to size your deployment
Rough figures are fine. With these, we can usually come back with options quickly.
| Workload | Training, fine-tuning, inference or token usage |
|---|---|
| Power | Total IT load MW and density kW/rack |
| Cooling | Air, direct-to-chip liquid, or hybrid |
| Compute | GPU model and node count, or expected tokens per month |
| Data | Data sources, volume TB and the reports you need |
| Network | Sites to connect, bandwidth Gbps, security needs |
| Location | Preferred market and facility tier |
| Timeline | Go-live date RFS and contract term |
How we work
From requirement to go-live
Understand
We take your workload, scale and timeline, and agree what success looks like.
Design
We propose the right mix of facilities, compute, data, network and security, with clear pricing.
Contract
We help settle terms and service levels with the providers involved.
Deliver
We coordinate installation, testing and handover, then stay on for support.