GPU MODELS / RTX 5080

RTX 5080 cloud GPU pricing

28 listings across 8 providers. Cheapest published rate: $0.150/GPU/hr at Salad. Sorted by per-GPU hourly price. Every row links to the provider's own published price.

WHAT IS THE RTX 5080

The RTX 5080 is Blackwell's upper-mainstream consumer card: GDDR7 bandwidth and FP4/FP8 tensor support in a 16 GB, 360 W package. It is genuinely fast per card but the 16 GB ceiling arrives quickly - in cloud catalogs it serves the same dev-and-small-model niche as the 4080 it replaces, just faster.

MAKER
NVIDIA
ARCHITECTURE
Blackwell
LAUNCHED
2025
MEMORY
16 GB GDDR7
MEMORY BANDWIDTH
960 GB/s
TENSOR COMPUTE
56.3 TFLOPS FP32; 5th-gen tensor cores with FP8/FP4 support
POWER DRAW
360 W
INTERCONNECT
PCIe 5.0 x16

WHAT IT IS GOOD AT

Small-model inference
Excellent - 7B-13B quantized models serve fast on 960 GB/s of GDDR7, and FP8 support stretches the memory.
Fine-tuning
LoRA on models up to ~13B. Full fine-tuning is capped by 16 GB.
Image and video generation
Strong fit - Blackwell media engines and tensor cores suit diffusion workloads.

WHO SHOULD NOT RENT THE RTX 5080

Models past ~13B parameters - 16 GB is the hard wall; an RTX 5090's 32 GB or an A100 is the step up.
Multi-GPU training - no NVLink.
ProviderConfigVRAMLocation$/hr$/GPU/hrSource
SpotGPUsSpotGPUsRTX 5080 × 116 GBNot published$0.090$0.090STALEreceipt
SaladSaladRTX 5080 × 116 GBGlobal (distributed)$0.150$0.150receipt
CloreCloreRTX 5080 × 116 GBRU$0.150$0.150receipt
QuickPodQuickPodRTX 5080 × 116 GBNot published$0.150$0.150STALEreceipt
CloreCloreRTX 5080 × 116 GBUA$0.187$0.187receipt
CloreCloreRTX 5080 × 116 GBCA$0.188$0.188receipt
CloreCloreRTX 5080 × 116 GBVN$0.200$0.200receipt
CloreCloreRTX 5080 × 116 GBFR$0.200$0.200receipt
CloreCloreRTX 5080 × 116 GBUS$0.208$0.208receipt
CloreCloreRTX 5080 × 116 GBAR$0.208$0.208receipt
CloreCloreRTX 5080 × 116 GBAU$0.220$0.220receipt
gpu.aigpu.aiRTX 5080 × 116 GBNot published$0.230$0.230receipt
Hostnot GPUHostnot GPURTX 5080 × 216 GBAsia-Pacific$0.510$0.255receipt
Hostnot GPUHostnot GPURTX 5080 × 816 GBAsia-Pacific$2.130$0.266receipt
Hostnot GPUHostnot GPURTX 5080 × 116 GBAsia-Pacific$0.270$0.270receipt
Hostnot GPUHostnot GPURTX 5080 × 116 GBUS West$0.290$0.290receipt
Hostnot GPUHostnot GPURTX 5080 × 816 GBUS West$2.340$0.292receipt
Hostnot GPUHostnot GPURTX 5080 × 216 GBUS West$0.590$0.295receipt
Hostnot GPUHostnot GPURTX 5080 × 116 GBUS East$0.320$0.320receipt
RunPodRunPodRTX 5080 × 116 GBGlobal$0.390$0.390receipt
Hostnot GPUHostnot GPURTX 5080 × 116 GBEU East$0.390$0.390receipt
GPUBrazilGPUBrazilRTX 5080 × 116 GBNorth America$0.397$0.397receipt
Hostnot GPUHostnot GPURTX 5080 × 116 GBCanada Central$0.430$0.430receipt
Hostnot GPUHostnot GPURTX 5080 × 116 GBEU West$0.430$0.430receipt
Hostnot GPUHostnot GPURTX 5080 × 216 GBUS East$0.860$0.430receipt
GPUBrazilGPUBrazilRTX 5080 × 116 GBAsia-Pacific$0.529$0.529receipt
GPUBrazilGPUBrazilRTX 5080 × 116 GBSouth America$0.601$0.601receipt
GPUBrazilGPUBrazilRTX 5080 × 116 GBEurope$0.679$0.679receipt

THE RTX 5080 MARKET, BY THE NUMBERS

PROVIDERS
8
CONFIGURATIONS
28
MEDIAN $/GPU/HR
$0.29
SPREAD
4.5x

8 providers publish 28 RTX 5080 configurations on the panel today. Published per-GPU hourly rates run from $0.150 (Salad) to $0.68 (GPUBrazil), a 5x spread between the cheapest and the most expensive published rate for the same chip. The median listing sits at $0.29/GPU/hr. 23 rows were reported in stock by the provider at last observation; the rest are published catalog rates. All rates are on-demand - committed-use and reserved pricing is excluded.

CHEAPEST RTX 5080 BY GEOGRAPHY
canadaClore$0.188/hrvietnamClore$0.200/hrfranceClore$0.200/hrunited statesClore$0.208/hraustraliaClore$0.220/hr
RTX 5080 PRICE TREND, DAILY PANEL SNAPSHOTS

Across every RTX 5080 row on the panel, the daily floor held flat at $0.150/GPU/hr over 53 days of tracking. Snapshots run daily; the full history is public in the repo.

RTX 5080 PROVIDER NOTES
Salad · 1 config · 1 region
Sets the panel floor for this model.
from $0.150/hr
Clore · 8 configs · 8 regions · 8 in stock
Sets the panel floor for this model. Widest regional footprint here - 8 regions.
from $0.150/hr
gpu.ai · 1 config
Published catalog rates.
from $0.230/hr
Hostnot GPU · 11 configs · 6 regions · 11 in stock
Widest regional footprint here - 6 regions.
from $0.255/hr
RunPod · 1 config · 1 region
Published catalog rates.
from $0.390/hr
GPUBrazil · 4 configs · 4 regions · 4 in stock
Widest regional footprint here - 4 regions.
from $0.397/hr
BUYING RTX 5080: WHAT THE PANEL SAYS

At the current floor, one RTX 5080 running around the clock costs about $110 a month (730 hours of on-demand arithmetic). Node shape matters: the single-GPU floor is $0.150/GPU/hr against $0.266/GPU/hr on an 8-GPU node (+77.5%). Per-GPU floors by config size: 1x $0.15 · 2x $0.26 · 8x $0.27. Compare the alternatives below before committing - the same budget often buys more than one chip class.

RTX 5080 VS THE ALTERNATIVES

RTX 5080 vs RTX 5090
The 5090 doubles the memory to 32 GB with nearly double the bandwidth. For anything past 13B it is the only consumer answer; below that line the 5080 is cheaper per hour.
RTX 5080 vs RTX 4090
Last generation's flagship carries 24 GB - 8 GB more than the 5080 - with comparable real-world speed. For LLM work the older card's memory often beats the newer card's architecture.

Compare live rates: RTX 5080 vs RTX 5090 · RTX 5080 vs RTX 4090 · RTX 5080 vs RTX 5070 Ti · RTX 5080 vs RTX 4080

RTX 5080 PRICING FAQ

What is the cheapest RTX 5080 cloud GPU right now?
The cheapest published RTX 5080 rate on the panel is $0.150 per GPU-hour at Salad (RTX 5080 x 1, Global (distributed)). Every price links to the provider's own published rate.
How much does a RTX 5080 cost per hour?
Across 8 providers tracked today, published per-GPU hourly rates for RTX 5080 run from $0.15 to $0.68, with a median of $0.29. Committed-use and reserved rates are excluded; everything shown is on-demand.
How many providers rent RTX 5080 GPUs?
8 providers publish 28 distinct RTX 5080 configurations on Compute Cafe. 23 of those rows were reported in stock by the provider's live endpoint at last observation.
Where is RTX 5080 cheapest?
By published per-GPU hourly rate, the cheapest geography for RTX 5080 today is canada at $0.19/hr (Clore). The spread across 5 tracked geographies runs to $0.22/hr in australia.
Are RTX 5080 prices going up or down?
Over the 53 days Compute Cafe has tracked the panel, the RTX 5080 floor is holding flat: $0.150 to $0.150/GPU/hr (+0.0%). Daily snapshots are public in the repo, so the trend is auditable.
What does a RTX 5080 cost per month?
At the current floor of $0.150/GPU/hr, a RTX 5080 running 24/7 costs about $110 per month (730 hours of on-demand arithmetic on the cheapest published rate). Committed-use contracts can price lower; the panel tracks on-demand rates only.
Is a full 8-GPU RTX 5080 node cheaper than a single card?
On today's panel: the single-GPU floor is $0.150/GPU/hr and the 8-GPU node floor is $0.266/GPU/hr - more expensive per GPU on a full node (+77.5%). Check the table for the exact configs behind each floor.
Is the RTX 5080 good for LLMs?
Up to about 13B parameters quantized, yes - fast and cheap. The 16 GB memory, not the compute, is the limit, which is why the 4090 and 5090 keep their following.
Why rent a 5080 over a 4090?
For FP8/FP4-native inference stacks and media generation it is faster per dollar. For raw LLM capacity the 4090's 24 GB usually matters more.

See also alternatives to the RTX 5080 · cheapest GPU-hour per model · inference chips · training chips · by location