Back your portfolio
with GPU capacity.
Invest capacity alongside capital: a GPU cluster the fund holds or reserves, shared by your portfolio companies on allocations or credits you set. We build it, run it and keep each company's share separate.
- Used by
- Your portfolio companies
- Access
- Allocations or credits you set
- Held by
- The fund, a vehicle or us
- Run by
- Lightyard, to an SLA
Capital and capacity, from one fund.
The fund owns GPU capacity and invests it alongside cash. Companies draw on it as they join the portfolio, and the capacity stays with the fund when they move on.
- Smaller cash roundsWith part of the investment made as compute, a company needs less cash for the same runway. The fund decides whether that means a smaller round for the same stake, or less dilution for the founders.
- Cheaper money for the hardwareGPUs can be financed. Held in a vehicle, they can be paid for by LPs or lenders looking for an infrastructure return, not a venture one.
- Capacity that staysWhen a company winds down, its share goes to the next company instead of being written off.
How one cluster is shared between companies.
One set of hardware, divided between companies that never see each other's work. The fund decides who gets how much, and changes it as the portfolio changes.
- Company ATraining16 nodes
- Company BInference8 nodes
- Company CFine-tuning4 nodes
- FreeHeld back by the fund4 nodes
- IsolationEach company runs on its own nodes, with its own accounts and access. Companies never share a node.
- AllocationShares change as companies raise, grow or wind down, without touching the hardware.
- ControlThe platform team decides who joins, how much each company gets and on what terms.
One long commitment for the whole portfolio.
GPU pricing rewards long commitments. A seed or Series A company can't sign one, so it pays short-term rates or goes without. The fund can sign one for the whole portfolio.
| Each company buying alone | The fund's cluster | |
|---|---|---|
| Commitment | Short terms or hourly, priced for the provider's risk | One long commitment, held by the fund or its vehicle |
| Price | Short-term rates, with the provider's margin | Set by the fund, from what the cluster costs to hold and run |
| Getting capacity | Whatever the provider has free that month | A fixed cluster, sized to the portfolio's plans |
| When a company fails or pivots | Its unused commitment is stranded | Its capacity goes to the next company |
Compared by category. Companies can still use public clouds for bursts.
Sized to today's demand, grown in waves.
The first cluster covers what portfolio companies need now. New capacity comes up in monthly waves of 32 nodes as the portfolio grows, on the same fabric.
- Start with real demandSize the first cluster to the companies that need GPUs now, not to the fund's whole plan.
- Spare capacityCapacity the portfolio doesn't need can be sold to outside customers, under the fund's name.
- Room to graduateA company that outgrows its share can take a cluster of its own from us.
Reserve it, own it or have it built.
Reserve a cluster we own and hold no hardware. Own it, directly or through a vehicle, and have us run it. Or have us build it and hand it to your own platform team.
| 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 |
What we do and what the fund decides.
- 01We build
- Sizing against portfolio demand
- Site and design
- Hardware procurement
- Build and burn-in
You get An accepted cluster
- 02We run
- Operations to an SLA
- A tenant per company
- Allocations and credits
- Usage reports
- A console for the platform team
You get A compute program the platform team runs
- 03The fund decides
- The structure and its capital
- Who gets what
- Terms with each company
- Whether to sell spare capacity
Questions we get on the first call.
Something missing? Ask us directly. We usually reply within a few hours.
Yes, if the fund structures it that way. The commitment becomes a credit balance on the cluster, and our software meters it. How it's valued and documented is for the fund and its counsel.
Companies need less cash when part of their compute comes from the cluster, the hardware can be paid for with infrastructure capital instead of venture equity, and capacity moves on when a company winds down.
No. Each company runs on its own nodes with its own accounts, and sees only its own share and usage.
The fund, a vehicle it sets up, or neither if it reserves the cluster from us. Companies pay the fund for what they use, or draw down credits.
We size the first cluster to current demand and add waves as it grows. Spare capacity can be sold outside the portfolio under the fund's name. We don't promise utilization.
No. The fund holds the contract with us, and companies use their allocation under the fund's own terms.
Yes. Allocations can be used for training, fine-tuning or inference, and the platform layer covers endpoints, jobs and containers.
An owned cluster is quoted per project, with operations as a term contract. Reserving one is priced per deal, for the 60-month term.
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.