Map your requirements.

Site, go-live date, and terms from what you specify. You bring the GPUs.

What you tell us.

Send it on a call, by email, or as a spec sheet. Above 10 MW, we scope it with your team.

MW. 1 MW to 10 MW, in one block or phased by hall.
Rack density. kW per rack. Sets which sites can power and cool it.
Cooling. Air, rear-door, or direct-to-chip liquid.
Redundancy. N, N+1, or 2N power, and whether the hall is cabled for a training fabric.
Location. A region, a metro, or a latency target.
Go-live date. The date your GPUs need power. Every path is scored against it.
Your GPUs. The hardware you deploy and when it arrives. We find the space, power, and cooling around it.
Term and budget. Contract term and the monthly budget the capacity has to fit.

Every step you need is placed on a timeline with a lead time and a dependency. The longest chain of dependent steps sets the date we give you.

Lead time. Every step, from contract to cross-connect, carries a sourced lead time. Nothing is scheduled without one.
Dependencies. Some steps wait on others. Racks wait on power handover; fabric cabling waits on racks; fiber waits on the cross-connect order.
Critical path. The longest chain of dependent steps is the critical path. Every other chain has slack and does not move the date.
Parallel tracks. Contract, credit, and fiber orders run while the hall is prepared. Nothing waits that does not have to.
Fastest source. Each item is quoted from more than one source. The model keeps whichever lands first.
Your GPUs set the target. Tell us when your GPUs arrive. The schedule works back from that date, so power and cooling are ready before the hardware lands.

Change any input and the model reruns.

From first call to a signed LOI.

First call. We review your requirements and the existing sites that fit them.
NDA. Signed before we share site details.
Credit application. Our intake on your company, so terms can be set.
LOI. Non-binding or binding, depending on the deal.