Loading prices…
RTX 5090 rental prices
NVIDIA · Blackwell · 32 GB GDDR7 · PCIe · 2025
6 providers · 25 configurations · prices as of Sep 27, 2026, 01:12 UTC
Lowest on-demand
$0.40per GPU·h · ≈ $293/month
at Vast.aiMedian on-demand
$0.65/GPU·h
across 6 providers
Lowest spot
$0.24/GPU·h
interruptible
FP16/BF16 cost
$1.91/PFLOP·h
#21 of 70 GPUs for value
Providers
6
25 configurations
Price history
Daily RTX 5090 prices per GPU-hour. The shaded period is reconstructed from the version history of SkyPilot's public price catalogs (fewer providers than live data); after that, PetaPrice's own tracking.
Lowest on-demandMedian on-demandLowest spot
Price by provider
Cheapest On-demand RTX 5090 offer at each provider, per GPU-hour. Shorter and greener is cheaper.
Vast.ai🇺🇸$0.40best
GPU.ai🇺🇸$0.40+0%- Excess Supply$0.65+62%
Yotta🇮🇳$0.65+62%
RunPod🇺🇸$0.69+72%- Phyntec$0.70+75%
Thin vertical line = median across providers ($0.65).
Every provider we track
Pricing
Size
Host
Provider HQ
6 providers · 25 configurations
Sort
- 7× GPU · PCIe192 vCPU · 472 GB RAM · 2,834 GB disk$0.40 /GPU·h$2.81 /hOn-demand🇺🇸 United States, US● availableRent ↗
- 8× GPU · PCIe256 vCPU · 252 GB RAM · 1,151 GB disk$0.40 /GPU·h$3.21 /hOn-demand🇺🇸 United States, US● availableRent ↗
- 2× GPU · PCIe64 vCPU · 63 GB RAM · 288 GB disk$0.40 /GPU·h$0.81 /hOn-demand🇺🇸 United States, US● availableRent ↗
- Excess SupplyOn-demandHQ unknownvia Shadeformfrom · On-demand+62%$0.65/GPU·h
- 1× GPU · PCIe12 vCPU · 120 GB RAM · 900 GB disk$0.65 /GPU·h$0.65 /hOn-demand🇳🇴 NO, Central● available
- 2× GPU · PCIe24 vCPU · 240 GB RAM · 1,750 GB disk$0.65 /GPU·h$1.30 /hOn-demand🇳🇴 NO, Central● available
- 8× GPU · PCIe120 vCPU · 960 GB RAM · 6,700 GB disk$0.65 /GPU·h$5.20 /hOn-demand🇳🇴 NO, Central○ sold out
- YottaOn-demand🇮🇳 HQ in Indiavia SkyPilot catalogfrom · On-demand+62%$0.65/GPU·h
- 1× GPU14 vCPU · 119 GB RAM$0.65 /GPU·h$0.65 /hOn-demandus-east-2 (+2 regions)
- 2× GPU28 vCPU · 238 GB RAM$0.65 /GPU·h$1.30 /hOn-demandus-east-2 (+2 regions)
- 4× GPU56 vCPU · 476 GB RAM$0.65 /GPU·h$2.60 /hOn-demandus-east-2 (+2 regions)
- PhyntecOn-demandHQ unknownvia Shadeformfrom · On-demand+75%$0.70/GPU·h
- 8× GPU · PCIe64 vCPU · 256 GB RAM · 7,600 GB disk$0.70 /GPU·h$5.60 /hOn-demand🇺🇸 US, Casper, WY○ sold out
Alternatives to the RTX 5090
GPUs with similar compute and memory, and their lowest on-demand price right now.
- RTX 409024 GB · 0.165 PF FP16$0.29/h$1.78/PFLOP·h
- RTX PRO 4500 Blackwell32 GB · 0.211 PF FP16$0.72/h$3.41/PFLOP·h
Trainium32 GB · 0.19 PF FP16$1.34/h$7.07/PFLOP·h- Inferentia232 GB · 0.19 PF FP16$0.76/h$3.99/PFLOP·h
- RTX 5000 Ada32 GB · 0.261 PF FP16$0.49/h$1.88/PFLOP·h
- L4048 GB · 0.181 PF FP16$0.69/h$3.81/PFLOP·h
RTX 5090 pricing questions
How much does it cost to rent an RTX 5090?
On-demand RTX 5090 prices start at $0.40 per GPU-hour (Vast.ai). The median across 6 providers is $0.65, and the most expensive is $0.70. Running one GPU around the clock costs about $293 per month at the lowest price.
Is there cheaper spot or interruptible RTX 5090 capacity?
Yes. Spot (interruptible) RTX 5090 capacity starts at $0.24 per GPU-hour, 40% below the cheapest on-demand price. Spot instances can be preempted at any time, so use them for checkpointed training or batch jobs.
How much compute do you get per dollar with the RTX 5090?
At $0.40 per hour and 0.209 PF of dense FP16/BF16, the RTX 5090 costs $1.91 per PFLOP-hour. That ranks 21 of 70 GPUs with on-demand prices on PetaPrice. Figures are datasheet peaks; real utilization is typically 30–60%.
What are the RTX 5090's key specs?
32 GB GDDR7 with 1,792 GB/s of memory bandwidth, Blackwell architecture (2025), PCIe form factor, 575 W. Dense peak: 0.209 PF FP16 / BF16, 0.419 PF FP8, 1.68 PF FP4, 0.838 PF INT8, 0.105 PF TF32, 0.105 PF FP32, 0.0016 PF FP64.
How current are these prices?
Prices were collected on Sep 27, 2026, 01:12 UTC from provider APIs and public catalogs and refresh about every 2 hours. They are indicative list prices; always confirm on the provider's website before booking.