GPU MODELS / RTX 4090 D

RTX 4090 D cloud GPU pricing

1 listings across 1 providers. Cheapest published rate: $0.268/GPU/hr at Gitee AI. Sorted by per-GPU hourly price. Every row links to the provider's own published price.

WHAT IS THE RTX 4090 D

The 4090D is the export-compliant variant of the RTX 4090 sold in restricted markets: slightly reduced compute, the same 24 GB of fast GDDR6X. Where it appears in catalogs - mostly Asia-Pacific providers - it serves exactly like a 4090, because the memory that matters for LLMs is unchanged.

MAKER
NVIDIA
ARCHITECTURE
Ada Lovelace
LAUNCHED
2023
MEMORY
24 GB GDDR6X
MEMORY BANDWIDTH
1,008 GB/s
TENSOR COMPUTE
~49 TFLOPS FP32 (export-restricted variant of the 4090)
POWER DRAW
425 W
INTERCONNECT
PCIe 4.0 x16

WHAT IT IS GOOD AT

LLM inference to ~13B
Same as the 4090 - 24 GB and 1 TB/s serve 13B-class models well.
Fine-tuning
LoRA/QLoRA to ~30B quantized. Marginally slower than the full 4090, practically identical.

WHO SHOULD NOT RENT THE RTX 4090 D

Multi-GPU training - no NVLink, same as the 4090.
Buyers who can get the unrestricted part at the same price.
ProviderConfigVRAMLocation$/hr$/GPU/hrSource
Gitee AIGitee AIRTX 4090 D × 124 GBChina$0.268$0.268receipt

THE RTX 4090 D MARKET, BY THE NUMBERS

PROVIDERS
1
CONFIGURATIONS
1
MEDIAN $/GPU/HR
$0.27
SPREAD
1.0x

1 providers publish 1 RTX 4090 D configurations on the panel today. Published per-GPU hourly rates run from $0.268 (Gitee AI) to $0.27 (Gitee AI), a 1x spread between the cheapest and the most expensive published rate for the same chip. The median listing sits at $0.27/GPU/hr. Rows are published catalog rates; stock flags appear per row where the provider reports them. All rates are on-demand - committed-use and reserved pricing is excluded.

RTX 4090 D PRICE TREND, DAILY PANEL SNAPSHOTS

Across every RTX 4090 D row on the panel, the daily floor moved from $0.280 to $0.268/GPU/hr over 8 days of tracking (-4.4%). Snapshots run daily; the full history is public in the repo.

RTX 4090 D PROVIDER NOTES
Gitee AI · 1 config · 1 region
Sets the panel floor for this model.
from $0.268/hr
BUYING RTX 4090 D: WHAT THE PANEL SAYS

At the current floor, one RTX 4090 D running around the clock costs about $195 a month (730 hours of on-demand arithmetic). Compare the alternatives below before committing - the same budget often buys more than one chip class.

RTX 4090 D VS THE ALTERNATIVES

RTX 4090 D vs RTX 4090
The D variant loses a few percent of compute and nothing else that matters for LLM work. Rent whichever is cheaper on the panel.
RTX 4090 D vs RTX 5090
The 5090 adds 8 GB of much faster GDDR7 and Blackwell tensor cores - a clear upgrade where its price premium is small.

Compare live rates:

RTX 4090 D PRICING FAQ

What is the cheapest RTX 4090 D cloud GPU right now?
The cheapest published RTX 4090 D rate on the panel is $0.268 per GPU-hour at Gitee AI (RTX 4090 D x 1, China). Every price links to the provider's own published rate.
How much does a RTX 4090 D cost per hour?
Across 1 providers tracked today, published per-GPU hourly rates for RTX 4090 D run from $0.27 to $0.27, with a median of $0.27. Committed-use and reserved rates are excluded; everything shown is on-demand.
How many providers rent RTX 4090 D GPUs?
1 providers publish 1 distinct RTX 4090 D configurations on Compute Cafe. The panel shows published rates; stock flags appear per row where the provider reports them.
Are RTX 4090 D prices going up or down?
Over the 8 days Compute Cafe has tracked the panel, the RTX 4090 D floor is falling: $0.280 to $0.268/GPU/hr (-4.4%). Daily snapshots are public in the repo, so the trend is auditable.
What does a RTX 4090 D cost per month?
At the current floor of $0.268/GPU/hr, a RTX 4090 D running 24/7 costs about $195 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 the 4090D worse than the 4090?
Marginally on compute, not at all on memory or bandwidth - for LLM serving and fine-tuning the difference is invisible.

See also alternatives to the RTX 4090 D · cheapest GPU-hour per model · inference chips · training chips · by location