NVIDIA H100, H200, B200, B300 and DGX-ready Data Centers

Many won’t, and they don’t tell you upfront. Some facilities have AUPs that specifically exclude consumer cards; others have infrastructure calibrated only for server-grade hardware. We know which facilities accept consumer and prosumer cards (RTX 4090/5090, RTX 6000 Pro) so you don’t spend a week on intake forms and emails before hitting a rejection. Tell us your exact card and we route accordingly.

For H100 and H200 deployments at 10-40 kW per rack, air cooling with containment or rear-door heat exchangers works. For B200 and B300 at 60-100+ kW, liquid (DLC or immersion) is mandatory. We match to air-cooled up to ~15-20 kW per rack and liquid for 30 kW+, so you’re not forced into a cooling model you don’t actually need. If you’re unsure, give us the GPU model and count and we’ll tell you what your hardware requires.

Committed kW is the standard billing model. You reserve a power amount and pay for it whether your hardware uses it or not. For inference workloads that idle most of the month and spike occasionally, this can be a significant overpay. We specifically look for facilities that offer true metered per-kWh billing for variable workloads. They exist, they’re just not the default. Tell us your utilization pattern and we’ll filter for it.

Yes. Most Tier 1 operators now gate small deployments behind 50-100 kW+ minimums. We route to mid-tier and regional facilities that take single-node and 1-2 cabinet GPU deployments without those minimums. The regional providers with available capacity for small NVIDIA deployments fill it through referrals, not public listings. That’s the inventory we access.

Real ranges (2026): $150-$250 per kW/month for power-only retail colo, plus $1,200-$3,500/rack/month for space and standard cross-connects. An 8x H200 HGX node (10.2 kW) typically lands at $2,500-$5,000/month all-in for power, space, and basic network in mid-tier markets. Prime markets (Ashburn, Santa Clara) run 20-40% higher. Liquid-cooled space adds $1,000-$2,000/rack premium. We send a few qualified options based on your specs, not generic ranges.

No, not strictly. H100 and H200 (700W TDP per GPU, ~10.2 kW per HGX node) can be air-cooled in facilities with rear-door heat exchangers or proper hot/cold aisle containment. Liquid is preferred for performance and density but not mandatory. For B200 (1,000W) and B300 (1,100-1,400W), direct-to-chip liquid cooling is mandatory — both for thermal reasons and to maintain NVIDIA’s warranty terms.

Air-cooled H100/H200 in power-slotted facilities: 2-6 weeks. Liquid-cooled B200/B300 in prime markets (Ashburn, Santa Clara): 3-6 months. Plan-B markets (Phoenix, Atlanta, Reno, Columbus): often 4-8 weeks even for liquid. GB200 NVL72 hyperscale racks: 3-6 months minimum. We track live capacity — the difference between a ‘qualified’ provider and one with actual power on the floor right now is what saves you 60+ days.

For steady inference and long-running training, yes — typically 40-60% lower 24-month TCO if you own the hardware. AWS p5e at ~$4.98/GPU-hr x 8 GPUs x 80% utilization is ~$700K over 2 years. Same workload on owned H200s in colo: ~$370-400K all-in. Cloud still wins for spiky, short-burst training. Hybrid (train cloud, infer colo) is the sweet spot most teams land on.

HGX H200 is NVIDIA’s reference 8-GPU baseboard that OEMs (Supermicro, Dell, Lenovo, ASUS) build into servers. DGX H200 is NVIDIA’s own complete system. Both pull ~10.2 kW per 8-GPU node, both prefer liquid but tolerate air, both want NVLink-friendly rack topology. H200 NVL is the PCIe variant: lower power (600W), air-cooled, fits standard enterprise racks at 15-25 kW.

NVIDIA’s DGX-Ready Colocation program certifies facilities meeting NVIDIA’s standards for DGX systems. Members include Equinix, Digital Realty, DataBank, and EcoDataCenter. You don’t strictly need a DGX-Ready facility unless NVIDIA support or warranty terms require it, but it’s a useful filter for AI-grade infrastructure quality. Many non-DGX-Ready providers host HGX-based systems just fine and often cheaper.

Most prime-market providers won’t take less than a full rack or a 100 kW commitment. Mid-tier and regional providers will take a single 10-15 kW cabinet on 12-month terms. We specifically maintain relationships with operators who accept small-footprint GPU deployments — a major reason single-node AI startups come to us instead of cold-calling Equinix.

400 Gbps InfiniBand (NDR) is standard for H200 multi-node training. 800 Gbps (NDR800) for B200/B300. Storage and management traffic should live on a separate 100/200G Ethernet fabric. NVLink scale-out has cable-length limits (~20-30m), so you need contiguous racks.

H100/H200 HGX: 30-40 kW/rack. H200 NVL (PCIe): 15-25 kW/rack. B200: 60-80 kW/rack. B300: up to 100 kW/rack. GB200 NVL72: 120 kW per rack-scale unit. Always provision for peak draw, not average — training workloads sustain near-max for hours.

No. We’re free to you. Providers share their existing sales commission with us, the same way they’d pay an in-house salesperson. You pay exactly what you’d pay going direct, and we save you the weeks of sales calls.

Most providers don’t publish real pricing, and the ones that do quote list prices that the real bill rarely matches (cross-connect fees, power overage tiers, install fees). ChatGPT doesn’t know who has 30 kW of liquid-ready capacity in Phoenix this week. We do because we’re actively calling them.

Yes. Most of our active conversations in 2026 involve B200 and B300 deployments. GB200 NVL72 is hyperscale-grade — fewer facilities qualify, lead times are 3-6 months minimum, and most deals are off-market. We have relationships with the operators that take these. Send specs and we’ll be straight about timelines.

Send us the SKU and quantity. We’ll back into the kW for you and confirm before quoting. Common mistake: people use nameplate wattage from the sales sheet, but real workload draw is often 70-85% of that. We size for sustained peak, which is what providers bill against.