Capacity mandate · CAP-01

Data center capacity, sourced and negotiated on your behalf.

Retail colocation, wholesale halls, high-density AI-ready suites and the power behind them. We find the capacity, benchmark the number against what this market is actually paying, and negotiate the terms — without your requirement ever being circulated to the market.

01

What a capacity mandate covers

  • Requirement definition

    Megawatts, density per rack, cooling method, latency tolerance, redundancy tier, ramp curve and expansion horizon — settled before anyone is contacted, because a vague requirement gets priced as a risky one.

  • Market shortlist

    Operators who genuinely hold what you need, in the regions where power is actually available on your timeline. You see the reasoning behind every name, including the ones we ruled out.

  • Commercial benchmarking

    What the market is paying per kW per month for comparable density, in that region, at that term length. Without this, the first number you are shown becomes the anchor.

  • Negotiation and close

    We advise from behind, lead alongside your team, or run the table outright. You contract directly with the operator and own that relationship.

02

What actually moves in a colocation negotiation

Price per kW is the headline. It is rarely where the money is.

Buyers walk into colocation negotiations focused on the rate and walk out having conceded most of the terms that determine total cost over the life of the contract. These are the clauses we work, roughly in order of what they are worth.

  1. The ramp schedule. When you begin paying for capacity you cannot yet fill is often worth more than the headline rate. A twelve-month ramp against a six-month ramp on the same rate is a materially different contract.
  2. Contracted versus installed capacity. Paying on contracted kW when your actual draw is a fraction of it is the most common quiet overspend in this market. Where the operator's economics allow it, there is room here.
  3. Redundancy tier against actual need. Training workloads and inference workloads do not need the same resilience. Buying 2N across the whole footprint when N+1 would carry half of it is an avoidable premium.
  4. Cross-connect and interconnection fees. Recurring, rarely benchmarked, and in carrier-dense facilities they compound. Caps and included allowances are negotiable and almost never volunteered.
  5. SLA credits with teeth. A credit capped at one month's fee on a workload that costs you far more than that per hour of downtime is not a remedy, it is a gesture. What counts as an outage, how it is measured, and what the credit is worth all move.
  6. Term length itself. Colocation is underwritten on five to ten years because that is what the operator’s financing case needs. That requirement is real, which makes term one of the few things you have that the operator genuinely wants — length is worth trading for rate, ramp or expansion rights rather than conceding for free.
  7. Escalators. A fixed annual escalator versus one indexed to inflation is a different number over a ten-year term. So is whether the escalator applies to power pass-through.
  8. Power pass-through mechanics. Whether you pay metered, at a blended rate, or with a PUE factor applied — and who carries the risk when the regional power price moves.
  9. Expansion rights. Right of first refusal on adjacent space, held at a pre-agreed rate, for a defined window. Cheap to secure at signature and expensive to buy later.
  10. Exit and assignment. Termination for convenience, assignment on a change of control, and what happens if your own capacity plan changes. Most standard forms are written to make leaving difficult.

None of this is exotic. It is simply what a counterparty who negotiates these contracts every week knows and a buyer negotiating one every few years does not.

Which market you are negotiating in changes what is winnable on every one of these. The same requirement priced in the United States, EMEA and APAC produces three different sets of movable terms, and where nothing existing fits the requirement it becomes a powered shell or build-to-suit negotiation instead. The Capacity Desk covers what the desk sources at density.

03

Where the capacity actually is

Power availability, not demand, decides where AI capacity lands. The right market is often not the one nearest you.

North America

Northern Virginia remains the deepest market and the most constrained. Dallas–Fort Worth and the wider ERCOT footprint, Phoenix, Atlanta, Chicago and Columbus carry more available power on shorter timelines, and secondary markets frequently beat all of them on time-to-energisation.

Nordics

Iceland, Norway and Sweden — low-cost renewable power, free cooling for much of the year, and the strongest cost-per-megawatt case in Europe for latency-tolerant training work.

UK & Ireland

London and Slough for latency-bound inference. Ireland where grid position allows it, which is increasingly the constraint rather than the land.

Continental Europe

Frankfurt, Amsterdam and Paris for the FLAP-D core, with Nordic and Iberian alternatives when FLAP-D power is unavailable at the density you need.

APAC

Singapore, Johor and wider Malaysia, and Japan. Capacity has migrated across the strait from Singapore, and pricing reflects it.

Latin America

Brazil, Chile and Mexico — renewable-heavy grids, improving subsea connectivity, and a cost basis that rewards workloads without a regional latency requirement.

Questions

Asked on this mandate.

01

Do you charge the buyer for capacity sourcing?

No. Winning business has a cost and it sits with the operator long before you enter the picture — that is where we are paid from. You contract directly with the operator at the rate we negotiated, and nothing about our involvement lands on your side of the deal.

02

Will our requirement stay confidential?

Yes. Nothing about your requirement is circulated to find capacity — we approach operators directly, one at a time, and only the ones you approve learn anything specific. A position that becomes common knowledge is a weaker position, so protecting it is part of the work rather than a courtesy.

Anonymous first contact is available, an MNDA can be in place before anything specific is shared, and a live mandate can run to completion without the wider market learning you were in it.

03

What density can you actually source?

Up to 150 kW per rack, with air, rear-door heat exchanger and direct-to-chip liquid cooling depending on the facility and the region. Above roughly 40 kW per rack the list of operators who can genuinely deliver narrows sharply, and knowing which claims survive a site visit is most of the value.

04

How quickly can capacity be identified?

First response within one business day, and a credible shortlist inside the first week for most standard requirements. Delivery timelines are set by the operator and the utility, not by us — for anything above a few megawatts, energisation date is usually the binding constraint rather than availability of space.

05

Can you help if we already have a proposal in hand?

Yes, and that is a common entry point. We benchmark what you have been offered, tell you where it sits against the market, and identify which terms are actually movable. Mid-negotiation is not too late.

Tell us the megawatts, the density and the date.

We will come back with a shortlist, a benchmark on the number, and a view on which terms are worth fighting for.

Book 30 minutesRequest the qualification →