GPU clusters for data center,
power and hardware owners.
A data hall with space, a power or mining site, or GPUs already bought. We design the cluster around what you have, build what's missing and operate it to an SLA on our software. The asset stays yours.
- You bring
- A hall, power or hardware
- We do
- Design, build, operate
- Owner
- You
- Operator
- Lightyard
A running cluster without becoming a GPU operator.
Operating GPUs is its own trade, with its own failure modes, spares and tooling. We take that on under contract, and the asset stays on your side.
- Run to an SLAWe watch the cluster, replace failed parts and keep firmware current, to the SLA in your contract.
- The asset stays yoursNo sale, lease or transfer to us. We work under an operations contract.
- ReportingRegular reports on health, failures and repairs, in the format the contract sets.
Where we start depends on what you have.
Each asset is missing something different. The first conversation is about what's there today.
| You have | What's usually missing | What we do |
|---|---|---|
| A data hall with power | A cluster designed for the room's power per rack, cooling and floor loading | Design the cluster around the room, source and build it, then run it |
| A power or mining site | A building fit for GPU density | Retrofit the building, or put a modular data center on the land |
| GPUs already bought | A site, a fabric and an operator | Review the hardware, find a site, build the cluster around it and operate it |
| A cluster that's already live | An operator | Review it, then take over operations to an SLA |
What we review before we take anything on.
Before a design or a handover, we look at what's there, so the contract describes the site and the hardware as they are.
- Power and coolingPower available per rack, how it's distributed, and whether the cooling can take liquid-cooled racks.
- The buildingFloor loading, space, access for racks and room for the network.
- Hardware and networkFor hardware you hold or a live cluster: its condition, firmware, fabric and tooling.
From your asset to a running cluster.
Step 1: Tell us what you have
The site, the power, any hardware, and anything already running on it.
You getA site reviewStep 2: Design around it
The cluster, plus any retrofit or module the site needs.
You getOne design for site and clusterStep 3: Build what's missing
Site work, hardware, racking, network and burn-in, in one scope.
You getAn accepted clusterStep 4: Operate it
Monitoring, failed-part replacement and support on our software.
You getA cluster run to the SLA
Most owners choose a managed cluster.
You keep the asset and we run it, with the build first if the cluster doesn't exist yet. Owners with their own operations team choose a deployment.
| 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 |
Your site, upgraded if it needs to be.
On-prem when the hall is ready for GPUs, a retrofit when the building needs work, modular when there's land and power but no building, colocation when you hold hardware but no site.
Questions we get on the first call.
Something missing? Ask us directly. We usually reply within a few hours.
No. Ownership doesn't change. We design, build and operate under contract.
Yes. We check the hardware first, then find or prepare a site, build the cluster and run it.
Often, yes. We either retrofit what's there or put a modular data center on the land.
No. We build and operate the cluster; who uses the capacity, and on what terms, is your decision.
Yes, after a review of the hardware, the network and the tooling. Operations then move to us under a contract with an SLA.
Power per rack and how it's distributed, cooling, floor loading, space and network access, plus the condition of any hardware you hold.
Site work and the build are quoted per project after the review. Operations are priced in the managed contract.
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.
- ColocationA Tier III facility in the region you choose.
- On-premYour data center, with the cluster designed around it.
- RetrofitA building or mining site upgraded for GPU density.
- ModularA prefabricated site on land with power.
- 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.