• Committed power: 10–80kW per rack, billed at 80–100% utilization; A/B 3‑phase 208–415V feeds.
• Cabinet or cage: 42–52U, with appropriate depth and weight rating for GPU/HPC chassis.
• Standard cooling:

1. Air / hot‑aisle containment up to ~20–30kW per rack.
2. Rear‑door heat exchangers in some facilities to push to ~30–40kW.

Network port: 1–10Gbps included; sometimes 10G commit or burst‑able.

• Cross‑connects: $100–$400 per month per x‑connect (to cloud on‑ramps, carriers, IX).
• Remote hands: $150–$300 per hour for GPU swaps, cable tracing, etc.
• Install & turn‑up: $500–$3,000 for rack/stack, cabling, and initial testing.
• HD cooling upgrades: RDHx or “liquid‑ready” manifolds adding $500–$2,000 per rack per month at 40kW+.
• Burst bandwidth / 95th percentile: Overage charges if you go beyond your commit on heavy data movement.

​

Real‑world difference between the “marketing” quote and the actual monthly run‑rate often lands 20–60% higher, depending on cross‑connect count, support usage, and cooling path.

Option A: Google / GPT + direct sales

You search “30kW rack colocation,” fill 10 forms, and hear “we support HD” until they admit their ceiling is 8–10kW with no RDHx. Weeks lost, no clear $/kW.

Option B: Directories & maps

Datacenter maps list hundreds of sites but not which ones energize 20–80kW racks or have real liquid/RDHx in place. You still end up on sales calls with providers that can’t meet your density.​

Option C: RFPs / consultants

Great for 5MW campuses, overkill for 10–80kW. Slow process, high fees, and you still have to interpret $/kW, cooling, and power paths yourself.

Option D: QuoteColo (power‑first broker)

Tell us 10–80kW per rack, markets, and cooling. We scan 500+ sites and send 3–5 that actually support your density, with $/kW, PDU/containment details, and timelines. Our fee is baked into provider sales costs, and clients typically save ~10% vs going direct.

10kW rack colocation:

• Often achievable with standard air cooling and simple containment.
• Typical for dense but not extreme mixed CPU workloads.

20kW rack colocation:

• Boundary between “standard” and “high density”.
• Often requires stronger containment and careful PDU design (3‑phase 208V, larger breakers).

30–40kW rack hosting:

• Usually needs RDHx (rear‑door heat exchangers), improved airflow, and detailed thermal planning.
• Not every “high density colocation” provider can sustain this level 24/7.

50–80kW rack colocation:

• Solidly into liquid‑cooling territory (direct‑to‑chip or immersion) for most designs.
• Requires different power risers, CDUs, and often custom containment; only a subset of facilities can offer it today.

Air / hot‑aisle containment: Reliable up to ~20–30kW per rack if designed well. Beyond that, fan energy and temperature differentials become problematic.

Rear‑door heat exchangers (RDHx): Extends air cooling up into the 30–40kW per rack range by removing more heat at the rack itself. Good compromise if you aren’t ready to go fully liquid.

Liquid cooling (30–80kW): Direct‑to‑chip or immersion, often adding $1,000–$2,000 per rack per month but enabling 60–80kW rack colocation with lower PUE and more predictable temperatures.

PDUs / power paths:

• 10–20kW: 3‑phase 208V PDUs with 30–60A breakers per feed may be enough.
• 30–40kW: Larger 3‑phase PDUs, careful balance across phases, and more outlets per rack.
• 50–80kW: Higher‑voltage distribution (sometimes 415V), CDUs for liquid, and strict redundancy planning.

Containment systems:

• Cold‑aisle / hot‑aisle containment is a standard baseline.
• At 30–40kW, you need tuned pressure/airflow, sensors, and sometimes blanking panels installed religiously.
• At 50–80kW, containment helps but is secondary to proper liquid cooling and power distribution.