NVIDIA RTX 6000 Ada
Ada · PCIe · 48 GB VRAM
As of , the cheapest listed rate for the NVIDIA RTX 6000 Ada is $0.52 per GPU-hour on DataCrunch (spot). The cheapest on-demand rate is $0.61 on Vast.ai (floor). Across 4 providers, the median on-demand rate is $0.84 per GPU-hour. To buy, the median new-condition listing is $7,299 across 8 sellers. With 48 GB of VRAM, that is $0.011 per GB of memory per hour at the cheapest rate and $0.017 at the median.
Buy
Median new-condition listing: $7,299 across 11 listings from 8 sellers, most recent September 8, 2026.
| Seller | Condition | Price | Seen | Source |
|---|---|---|---|---|
| MBPC | new | $3,999.99 | Listing | |
| SHI International | new | $6,000 | Listing | |
| Newegg.com - Tech Trends | new | $6,399 | Listing | |
| ServerSupply.com | new | $6,555 | Listing | |
| ServerSupply.com | new | $6,958.75 | Listing | |
| ASA Computers | new | $7,299 | Listing | |
| DirectDial | new | $8,092 | Listing | |
| ServerOrbit.com | new | $8,593.20 | Listing | |
| PCNation | new | $10,711.09 | Listing | |
| ServerSupply.com | new | $11,352.50 | Listing | |
| PCNation | new | $11,742.95 | Listing |
At the cheapest tracked rental rate ($0.52/hour on DataCrunch), that purchase price equals about 14,037 GPU-hours of rental — roughly 19 months of continuous use. Power, hosting, cooling and idle time are not counted.
Rent
| Provider | USD / GPU-hour | USD / GB VRAM-hour | Type | Region | Min. commitment | Source |
|---|---|---|---|---|---|---|
| DataCrunch | $0.52 | $0.011 | spot | — | — | Price page |
| Vast.ai | $0.61 | $0.013 | on-demand floor | — | — | Price page |
| RunPod | $0.74 | $0.015 | on-demand | Community Cloud | — | Price page |
| RunPod | $0.74 | $0.015 | spot | Community Cloud | — | Price page |
| RunPod | $0.84 | $0.017 | on-demand | Secure Cloud | — | Price page |
| RunPod | $0.84 | $0.017 | spot | Secure Cloud | — | Price page |
| DataCrunch | $1.04 | $0.022 | on-demand | — | — | Price page |
| Paperspace | $1.57 | $0.033 | on-demand | GPU Droplets | — | Price page |
Data updated: