GPU colocation,
in the region you choose.
Your GPU cluster in a Tier III colocation facility in the region you commit to, across the US, Europe, Asia and South America. We find the facility, fit the space for the hardware, build the cluster in place and can run it for you.
- Facility
- Tier III
- Regions
- 7
- Selected by
- Lightyard
- Tenancy
- Single tenant
A Tier III facility in the region your rules require.
Data rules, customer contracts and latency decide where a cluster can sit. You commit to a region, and we shortlist and select the facility inside it.
- Facility
- Tier III
- Coverage
- US · Europe · Asia · South America
- The region is in the contractThe cluster stays in the region your contract names, for the length of the contract.
- We find the facilityShortlisted for power, cooling and connectivity, then arranged as part of the project.
- One project, one scopeThe space, the fit-out and the cluster are scoped together in one proposal.
The hall is checked against the hardware first.
A hall built for general IT can fall short on GPU racks in three places. We confirm each one with the facility before any hardware ships.
- Power per rackConfirmed against the platform you choose, at the rack positions the cluster will use.
- Cooling to the rackGB300 NVL72 racks need liquid cooling at the rack. HGX B300 nodes run on air.
- Floor loadingFully loaded GPU racks are heavy. The floor is checked for them before the layout is final.
Colocation against building your own site.
Colocation puts your capital into the cluster. The building, the power connection and the security are already there.
| Building a site | Colocation | |
|---|---|---|
| What you fund | Land, a building, a power connection and the cluster | The cluster and its fit-out in an existing hall |
| Before hardware goes in | Permits, construction and the power connection | A hall that is already built, powered and secured |
| What you commit to | The site, for its whole life | A colocation contract, with terms agreed per project |
| Adding capacity | Space you build out yourself | More racks in the same facility, where it has room |
Compared by category. Costs and terms are set in each proposal.
From a region to a running cluster.
Step 1: Commit to a region
Pick the region your customers, data and contracts need. It goes into the contract.
You getA committed regionStep 2: Select and fit the hall
We shortlist facilities in that region, select one and agree the fit-out for the cluster.
You getA hall ready for the racksStep 3: Build in place
Our team racks, cables and burns in on site, wave by wave for a large cluster.
You getA burned-in clusterStep 4: Run it
We operate it to the SLA in the contract, or your team takes it over at handover.
You getA running cluster
The systems the hall has to take.
HGX nodes are air-cooled. NVL72 racks are liquid-cooled and much denser, so the facility shortlist depends on which one you choose.
- Rack scaleNVIDIA GB300 NVL72
- Next generationNVIDIA Vera Rubin
Every solution works in colocation.
Colocation is where the cluster sits. Who owns it and who runs it is a separate choice, made in the rows below.
| Way to work with us | Owner | Design | Source | Build | Operate | Software |
|---|---|---|---|---|---|---|
| Cluster deploymentWe design, source and build it, then hand it over burned in. Your team runs it. | You | |||||
| Managed clusterYou own it, we run it. We can build it first, or take over a cluster that's already live. | You | |||||
| SoftwareYour team runs every stage, on our stack. | You | |||||
| Reserved clusterWe own and run it. You reserve it for 60 months, with rollover into the next generation. | Lightyard |
Buyers who need a region, not a building.
01Neoclouds
Capacity in the regions your customers ask for.
You bring the demand. We find the hall and build the cluster in it.
GPU clusters for neoclouds03Enterprises & sovereign
A cluster held in the region your rules name.
You bring the data rules. We pick a facility that meets them.
Private and sovereign AI06Investors
An asset in a facility, not a construction project.
You bring capital. We place the cluster in a colocation hall and run it.
GPU infrastructure for investors
Other places a cluster can live.
If you already have a data center, a powered building or land with power, one of these may fit better than a colocation hall.
- 01 Colocation · this page
A Tier III facility in the region you choose.
Questions we get on the first call.
Something missing? Ask us directly. We usually reply within a few hours.
A Tier III facility in the region you commit to. We don't name facilities on this site; your proposal sets the region.
Seven, across the US, Europe, Asia and South America: US West, US Central, US East, Europe, East Asia, Southeast Asia and South America.
We find and arrange the space as part of the project. Who holds the facility contract is agreed project by project.
The facility hosts other customers, but your cluster is single-tenant, in racks that serve only you.
Not every hall can. For GB300 NVL72 we shortlist facilities with liquid cooling at the rack, or ones that can be fitted for it.
Most clusters are 32 to 128 nodes; larger ones are scoped case by case. The facility shortlist is drawn up for the size you plan to reach.
Us, to an SLA in the contract, or your own team after handover. The facility handles power, cooling and physical security.
Solutions, sites and who we build for.
- AI factoryEvery stage, from the site to a running cluster.
- Cluster deploymentDesigned, sourced, built and handed over burned in.
- Managed clustersYou own the cluster. We build it or take it over, and run it.
- Reserved clustersWe own and run it. You reserve it for the term.
- SoftwareRun every stage yourself, on our stack.
- White-label GPU cloudYour GPU cloud under your brand, built and run by us.
- NeocloudsCapacity built to your standard, ready to sell.
- Venture capitalOne cluster, shared by your portfolio companies.
- Enterprises & sovereignYour own cluster in your region.
- Private equityGPU capacity on portfolio sites, or a neocloud as a platform.
- Asset ownersHardware, a hall or a power site, turned into a running cluster.
- InvestorsAn operated GPU asset, without building an operator.