Safeera
Three Centurions on a planted hospital rooftop at sunrise; close-ups of the compute inside, a patient room with monitoring equipment, and one unit glowing in a night field.
Centurion: ambient film, no narration.

Centurion.

Up to 64 B300 GPUs for the hospital or campus, Vera Rubin-ready.

GPUs
Up to 64
Power
100–120 kW
Height
3.2 m

Tell us about the site.

Where will it live?

Privacy

Talk to Safeera

Next: a site assessment, an electrical and network review, and a written proposal. No order fee. First deliveries planned from January 2027; timing subject to change.

Est. total $6,600,000 at 64 GPUs · Final price confirmed before delivery.

Due to scarcity of compute components, the price may change in the event of memory or NVIDIA price increases.

Specification

A hospital’s own cortex. Up to 64 B300 Blackwell Ultra GPUs today, ready for Vera Rubin. Installed on the roof or standing in the park.

Compute
Up to 64 B300 Blackwell Ultra GPUs, with an upgrade path to Vera Rubin
Power
100–120 kW from the building’s supply
Site
Rooftop after structural review, or a ground pad in the park
Size
About 3.2 metres tall
Data
Never leaves the campus
1.75 m 3.2 m tall ≈ 2.5 m across ⌀ ≈ 2.5 m base ⌀ ≈ 1.3 m door ≈ 0.9 m ≈ 0.5 m floor ≈ 3.9 × 3.5 m
Centurion: 3.2 m tall, ≈ 2.5 m across. Stands on a ≈ 1.3 m base, anchored to the slab. Clearance ≈ 0.9 m at the door, ≈ 0.5 m on the other sides; floor space ≈ 3.9 × 3.5 m.

Economics

Capacity value is computed from public cloud on-demand list prices for equivalent GPU capacity, the median of the 4 providers below as of September 11, 2026, at 40, 70 and 90% utilization. It is gross revenue potential, not a forecast.

Capacity value

What the same capacity rents for in the public cloud: 64 B300 Blackwell Ultra GPUs at the full configuration, at published on-demand list prices. This is what an owner would otherwise pay, or could recover by making spare capacity available.

AWS
$1,139.20 per hour — P6-B300 (p6-b300.48xlarge, 8x B300), $17.80 per GPU-hour
Microsoft Azure
$1,730.56 per hour — ND GB200 v6 (Standard_ND128isr_NDR_GB200_v6, 4x GB200 Blackwell GPUs), $27.04 per GPU-hour
Oracle Cloud
$960.00 per hour — BM.GPU.B300.8 (bare metal, 8x B300), $15.00 per GPU-hour
CoreWeave
$550.40 per hour — B200 bare metal (8x B200), $8.60 per GPU-hour
Per hour
$1,049.60 — median of the providers above
Per year at 40%
$3,677,798
Per year at 70%
$6,436,147
Per year at 90%
$8,275,046

Method: median list price per GPU-hour × 64 GPUs × 8,760 hours × utilisation. Based on published on-demand list prices for equivalent GPU capacity as of September 11, 2026; excludes networking, storage and support; not a forecast of resale income.

Owner’s calculator

Financing for Centurion is arranged per project, so the figures stop at net before financing; the payback on the full estimate assumes an outright purchase at the est. total.

Monthly gross revenue
$509,528 est.
Monthly electricity
$11,151 est.
Internet
$250 est.
Estimated monthly operating cost
$11,401 est.
Net before financing
$498,127 est.
Payback on the full estimate
14 months est.
How this is calculated
  • Powered hours a month = 730 × (1 − downtime). Hours sold = powered hours × utilization.
  • Monthly gross revenue = sell price per GPU-hour × 64 GPUs × hours sold.
  • Monthly electricity = (120 kW × hours sold + 120 kW × powered hours not sold) × price per kWh. The top of the 100-120 kW range, charged in every powered hour; no published breakdown.
  • Estimated monthly operating cost = electricity + internet. Net before financing = gross revenue − operating cost.
  • Payback on the full estimate = est. total ÷ net before financing, in whole months.

Defaults: utilization 70% and downtime 5% are assumptions to adjust. The sell price, $16.40 per GPU-hour, is the median of the 4 list prices above. Electricity, $0.134 per kWh, is the 2025 U.S. average retail price for the commercial sector (U.S. Energy Information Administration, Electric Power Monthly, Table 5.3; checked October 11, 2026). Internet, $250 a month, is a published business fiber list price for a 2 Gbps line, plus taxes and fees (checked October 11, 2026).

Revenue potential depends on demand and on the price Safeera’s distribution channels achieve. The figures exclude electrical service upgrades, internet upgrades, maintenance, insurance, taxes and time offline beyond the downtime entered. They are estimates, not a forecast or a guarantee.

The equipment price is an estimate and subject to change; the final price is confirmed before delivery.

Made for

  • Hospitals

    Pathology, imaging and clinical models trained inside the security perimeter.

  • Pharma and labs

    Molecular design at the scale of the labs, with the same upgrade path the national labs chose.

  • Plants and campuses

    Digital twins and a shared intelligence for robot fleets, air-gapped if you choose.

Field notes

The clinic in the unit: private medical AI for hospitals

Mayo Clinic's on-premises Blackwell system compresses four weeks of pathology analysis into one across 20 million digitized slides, with physicians still making the final call. The strongest hospital AI runs on hardware the hospital owns. Read the note

Robots in healthcare: from surgical arms to bedside helpers, and the brain they'll share

Surgical arms, delivery robots and exoskeletons are hospital infrastructure; Diligent's Moxi alone has made over 1.25 million deliveries. Each is a networked machine with cameras among vulnerable people. Their shared intelligence belongs on the hospital's own roof, where it can be audited and governed by the hospital itself. Read the note

Physical AI on the factory floor

Humanoids from Figure, Apptronik and Agility are on plant floors, and Amazon runs more than a million robots. A fleet does not need to carry all its intelligence on its back; the safest place for its shared brain is on-site, behind one switch. Read the note

Talk to Safeera

Also

Googie, for the homeSparks, for the officeThe thesisDrones, cameras, and street lights: giving AI a body, safely