Rta LabsHigh-density AI colocation, ready for your hardware.
Rta Labs Compute sources high-density colocation for AI hardware you own or have ordered: power, liquid-cooling support, racks, cages and larger deployments, in North America and other regions. If your GPUs are already on order, we work back from the delivery date and coordinate the facility, power, cooling and connectivity discussions.
What Rta Labs sources
- High-density power for GPU servers and clusters, from a few racks to multi-megawatt requirements
- Liquid-cooling support: direct liquid cooling or rear-door heat exchangers
- Racks, cages and larger deployments, with expansion options
- Network carriers, cross-connects, internet exchange and cloud on-ramps
What to include in your request
- Hardware: owned, ordered or planned, its type, and delivery date
- Power: total target IT load in kW, minimum viable load, maximum kW per rack
- Cooling: air, direct liquid cooling, rear-door or flexible, and CDU needs
- Resilience: A/B power and redundancy expectations
- Connectivity: carriers, cross-connects, exchange or cloud on-ramp needs
- Operations: remote hands, physical security, required certifications
- Timing and terms: target ready-for-service date, term, expansion option
- Location: country, region or metro, and any data residency requirement
For each major point, say whether it is required, preferred or flexible.
Location, residency and control
Rta Labs Compute sources worldwide, with a focus on North America. You can require a country, set a preferred region or metro, and say whether a location is required or only preferred.
You can require data residency in a specific country, or a provider controlled in a specific country, such as a Canadian-controlled provider. Rta Labs verifies these per provider and per requirement.
Physical location alone does not make a provider sovereign. Rta Labs does not label a provider sovereign on location alone.
When a workload allows it, Rta Labs also checks options in other regions.
AI Colocation Sourcing FAQ
Our GPUs are already ordered. Can you find space before they arrive?
Yes. Share the hardware, the expected delivery date, and your power and cooling needs. Rta Labs sources facilities that can be ready for service around that date and coordinates the facility, power, cooling and connectivity discussions. Ready-for-service dates are confirmed by the facility in writing.
Does Rta Labs Compute operate data centres?
No. Rta Labs Compute sources colocation from third-party facilities. Rta Labs does not own or operate third-party infrastructure unless it expressly says so.
What power density can I request?
State your total target IT load, the minimum you could start with, and the maximum kW per rack you need. Rta Labs checks provider options against those figures and confirms them before presenting an option.
Can I request liquid cooling?
Yes. The request covers air cooling, direct liquid cooling and rear-door heat exchangers, plus any CDU requirement. You can mark cooling as required or preferred.
Who signs the colocation contract?
The buyer and the provider generally contract directly, unless another written arrangement says otherwise. Rta Labs coordinates the introduction and the commercial process.
Capacity, pricing, and delivery and deployment dates are confirmed with suppliers before a formal option is presented. Suppliers deliver under their own agreements, and their identities stay private until an introduction is approved.
Tell us what you need.
A person at Rta Labs reads every request. Incomplete requirements are welcome.
Also sourced: GPU hardware · GPU compute · Inference infrastructure