# UpStation rack plan UPS-DSN-20261103-G9YQ — budgetary design, not a quote

Prepared 2026-11-03 for **Example AI Lab** · Head of ML Infrastructure · prices as of 2026-11-03

## 4× HGX B300 training cluster (32 GPUs) with NVMe storage and a services cluster, Montana colo

**82.51 kW provisioned** (nameplate ÷ 0.8), each rack held to 40 kW, A+B feeds (either one carries the whole rack).

(rack elevation: the SVG drawing in the HTML/PDF version of this document)

## Power per rack

| Rack | Space | Provisioned | Circuits (A+B) | Load per feed | Cords per feed |
|---|---|---|---|---|---|
| Compute 1 | 22/40 U | 35.24 kW | 1× 415V 3-ph 60A | 82% | 8 |
| Compute 2 | 20/40 U | 33.92 kW | 1× 415V 3-ph 60A | 79% | 6 |
| Services 1 | 15/40 U | 13.35 kW | 1× 208V 3-ph 50A | 74% | 10 |

## Colocation by location

| Location | $/kW/month | Per month | Term total | Sales tax | Property tax |
|---|---|---|---|---|---|
| Montana (this design) | $260 | $21,453.00 | $772,308.00 (36 mo) | none | none on the first $1M of equipment, about 0.8%/yr above |
| Texas | $225 | $18,565.00 | $668,340.00 (36 mo) | 8.25% (6.25% state + local); the data-center exemption needs $200M and 20 jobs | about 2.2%/yr of market value |
| Oregon | $175 | $14,439.00 | $519,804.00 (36 mo) | none | about 1.5-2%/yr of market value (ORS 308.290); abatements closed to data-center property since 2026 |
| Pennsylvania | $250 | $20,628.00 | $742,608.00 (36 mo) | 6% (8% Philadelphia, 7% Allegheny), or none if the facility is state-certified and the tenant holds its own exemption certificate (100 kW+, 2-year contract) before buying | none |
| Canada (Ontario, Quebec) | $225 | $18,565.00 | $668,340.00 (36 mo) | HST 13% (Ontario) or GST+QST (Quebec) on hardware and colo, recovered as an input tax credit by a registered business | none |

## How it was laid out

- GPU nodes per 40 kW rack = min(floor(40 U ÷ 10 U), floor(40 ÷ 16.96 kW)) = 2 (power binds).
- Compute 1: 22 U, 35.24 kW provisioned → 1 × 415V 3-ph 60A circuit pair (A+B), 82% of each feed; 8 power cords per feed.
- Compute 2: 20 U, 33.92 kW provisioned → 1 × 415V 3-ph 60A circuit pair (A+B), 79% of each feed; 6 power cords per feed.
- Services 1: 15 U, 13.35 kW provisioned → 1 × 208V 3-ph 50A circuit pair (A+B), 74% of each feed; 10 power cords per feed.

## Devices

| Rack | U | Name | Device | RU | kW | PSUs |
|---|---|---|---|---|---|---|
| Compute 1 | 40 | gpu-sw-1 | 32× 800G OSFP switch — GPU fabric | 1 | 1.25 | 2 |
| Compute 1 | 39 | oob-1 | 48× 1G RJ45 + 4× 10G switch — OOB | 1 | 0.07 | 2 |
| Compute 1 | 11 | gpu-02 | 8x NVIDIA B300 (HGX B300 NVL8) | 10 | 16.96 | 6 |
| Compute 1 | 1 | gpu-01 | 8x NVIDIA B300 (HGX B300 NVL8) | 10 | 16.96 | 6 |
| Compute 2 | 11 | gpu-04 | 8x NVIDIA B300 (HGX B300 NVL8) | 10 | 16.96 | 6 |
| Compute 2 | 1 | gpu-03 | 8x NVIDIA B300 (HGX B300 NVL8) | 10 | 16.96 | 6 |
| Services 1 | 40 | fe-1a | 32× 400G QSFP-DD switch — front-end pair, MLAG | 1 | 0.75 | 2 |
| Services 1 | 39 | fe-1b | 32× 400G QSFP-DD switch — front-end pair, MLAG | 1 | 0.75 | 2 |
| Services 1 | 13 | svc-03 | 1U compute node | 1 | 1.2 | 2 |
| Services 1 | 12 | svc-02 | 1U compute node | 1 | 1.2 | 2 |
| Services 1 | 11 | svc-01 | 1U compute node | 1 | 1.2 | 2 |
| Services 1 | 9 | stor-05 | 2U all-NVMe storage node | 2 | 1.65 | 2 |
| Services 1 | 7 | stor-04 | 2U all-NVMe storage node | 2 | 1.65 | 2 |
| Services 1 | 5 | stor-03 | 2U all-NVMe storage node | 2 | 1.65 | 2 |
| Services 1 | 3 | stor-02 | 2U all-NVMe storage node | 2 | 1.65 | 2 |
| Services 1 | 1 | stor-01 | 2U all-NVMe storage node | 2 | 1.65 | 2 |

## Assumptions

- GPU server: 8x NVIDIA B300 (HGX B300 NVL8), Intel, 3TB (Aivres) (UPS-GPU-B300-8-T2-AIV), the default HGX line (proposed)
- Second source: 8x NVIDIA B300 (HGX B300 NVL8), Intel (UPS-GPU-B300-8), same price and design, 30+ weeks on NVIDIA allocation. Next generation: NVIDIA Rubin (HGX Rubin NVL8, liquid-cooled racks) follows B300, not priced

**Next step.** Book the intro (https://calendly.com/alecglover/intro-call) or email alec@upstation.io with the design ID; UpStation validates the design and introduces the supplier that fits.

Prepared by UpStation from public supplier prices at ai.upstation.io. Budgetary only; the supplier confirms price, availability and lead time on order. UpStation does not resell. alec@upstation.io
