Request capacity
FOR AI TENANTS

Colocation for GPU clusters, scoped right the first time.

You bring the GPUs. We find the space, power, and cooling to run them. LD Advisors works with neoclouds, AI labs, and physical AI companies deploying anything from a single rack to 20 MW+ across the US.

Request capacityWhat to send us
WHO IT'S FOR
  • GPU clouds expanding into new US regions
  • AI labs moving training off public cloud
  • Robotics and physical AI teams that need compute near their data
  • Enterprises bringing GPUs on-prem
WHAT WE HANDLE
  • Site search across US operators with capacity that fits
  • Spec translation (GB300 / NVL72, B300, H100/H200) into cooling and power requirements
  • Power date and ownership checks
  • Term negotiation: deposits, power pass-through, ramp schedules, expansion rights
  • Network, interconnect, and storage planning
HOW IT WORKS

One requirement in, vetted options out.

01

Scope

We build a full spec with you: power, density, cooling, location, timing, term, tier, network.

02

Match

We take it to operators with capacity that fits, anonymously if you prefer.

03

Vet

We confirm power dates, cooling, and site ownership before any introduction.

04

Close

We help negotiate terms and stay involved through move-in.

GB300 READINESS

What a GB300 NVL72-class rack asks of a site.

Most existing data halls were built for air-cooled racks. Before we bring you a site, we check it against the rack's published requirements.

  • ✓Power: 132 kW nominal per rack, with busway provisioned for 192 kW (EDPp).
  • ✓Cooling: direct-to-chip liquid cooling through a CDU tied into the facility water loop.
  • ✓Weight: about 3,300 lb fully loaded. Floor loading and the delivery path both need checking.
  • ✓Power whips: 400 V or 480 V, 60 A.
  • ✓Footprint: 600 mm wide by 1,068 mm deep.
  • ✓Energization: a confirmed date, in writing, for the hall you would occupy.

Source: HPE QuickSpecs, NVIDIA GB300 NVL72 by HPE.

Diagram of a GB300 NVL72-class liquid-cooled rack with its coolant distribution unit
FROM REQUEST TO MOVE-IN

What happens after you send a requirement.

  1. 01

    Reply

    We respond the same business day and set up a mutual NDA if you want one.

  2. 02

    Scope

    One call to pin down IT load, density, cooling, markets, timing and term.

  3. 03

    Shortlist

    Vetted options in about five business days, each with a confirmed power date.

  4. 04

    Diligence

    Tours, cooling and power checks, and site ownership confirmed before you commit.

  5. 05

    Terms and move-in

    We help negotiate the contract and stay involved through installation.

WHAT TO SEND US

A complete spec gets you real options faster.

Send what you have. If some of it isn't settled yet, we'll work it out with you on the first call.

Request capacity
IT load
MW or rack count
Density
kW per rack
Cooling
Air, rear-door, direct-to-chip
Location
Preferred markets or latency needs
Timing
Target energization date
Term
Length and ramp schedule
Tier
Tier and redundancy
Network
Carrier diversity
Access
Security and access requirements
FAQ

Straight answers.

What does a data center colocation advisor do?

A colocation advisor finds and negotiates data center space and power for a tenant. LD Advisors scopes your GPU deployment, takes it to operators with matching capacity, confirms power dates and cooling, and helps negotiate terms through move-in.

What do I need to send to get started?

Send your IT load (MW or rack count), kW per rack, cooling type, preferred markets, target date and term. If you don't know some of it yet, we'll help you work it out on the first call.

Can you find liquid-cooled space for GB300 NVL72 racks?

Yes, that's our focus. GB300 NVL72-class racks need direct-to-chip liquid cooling and over 100 kW per rack. We only bring back sites that can support that density, with a confirmed energization date.

What size deployments do you work on?

Anything from a single rack to 20 MW+, for neoclouds, AI labs, and physical AI companies.

Which US markets do you cover?

We source capacity nationwide. Tell us your latency, power, and budget constraints, and we'll suggest markets that fit.

Will you share my company's name with data centers?

Not without your permission. We can take your requirements to operators anonymously and share details only under NDA.

What's the difference between colocation, powered shell, and GPU-as-a-service?

Colocation rents you racks, power, and cooling for hardware you own. Powered shell is an energized building you fit out yourself. GPU-as-a-service rents you the GPUs too. We focus on colocation and powered shell for teams that own their hardware.

REQUEST CAPACITY

Tell us what you need to deploy.

We'll come back the same business day. Details stay under NDA, and we can take your requirement to operators anonymously.

Have capacity to fill? List your capacity

Got it. Your requirement is in.

We'll reply the same business day. If you'd like an NDA in place first, say so in your reply and we'll send ours.

That didn't go through. Email contact@ldadvisors.ai and we'll pick it up from there.