GeForce RTX 4090 rental price
Workstation & consumerThe workhorse of the peer-to-peer rental market. Nothing else gets you this much throughput per dollar-hour if 24 GB is enough.
live · updated 17d ago
VRAM
24 GB GDDR6X
Architecture
Ada Lovelace
Segment
Workstation & consumer
Providers quoting
5
best rate
$0.121/GPU-hr at Vast.ai
24h
—
7d
—
30d
—
cheapest 8× node
$3.200/hr
The cheapest GeForce RTX 4090 you can rent right now is $0.121 per GPU-hour at Vast.ai, against $0.490 at the dearest of the 5 providers quoting it — a 4.1× spread for identical silicon. A full 8-GPU node starts at $3.200/hr at Excesssupply. Rates below refresh automatically.
Rent vs. buy
buy outright
$3,300
rent
$0.121/hr
breakeven
3.1 years
Buying a GeForce RTX 4090 at today’s street price of $3,300 costs the same as 27,331 hours (3.1 years) of renting one at $0.121 an hour. Run it less than that and renting is simply cheaper; run it more and the card starts paying itself back — but this is a floor, not a total cost of ownership. It counts only the purchase price against the rental rate, and ignores power, cooling, rack space, networking, depreciation and the cost of tying up the capital, all of which push the real breakeven further out.
GeForce RTX 4090 purchase prices and history →RTX-4090 hourly rate history
Daily low $/GPU-hour per provider, cheapest provider first — hover the chart for detail.
Where to rent the GeForce RTX 4090
Vast.ai
Marketplace
$0.121/GPU-hr
2× RTX-4090 · $0.241/hr per node · CN
- 1× RTX 4090 x1$0.121/hr
- 1× RTX 4090 x1$0.161/hr
- 1× RTX 4090 x1$0.161/hr
Excesssupply
GPU cloud
$0.400/GPU-hr
8× RTX-4090 · $3.200/hr per node · pcie · US, Casper, WY
- 1× n3-RTX-4090x1$0.600/hr
- 2× n3-RTX-4090x2 · no capacity$1.200/hr
- 4× n3-RTX-4090x4$2.400/hr
Phyntec
GPU cloud
$0.400/GPU-hr
8× RTX-4090 · $3.200/hr per node · pcie · US, Casper, WY
no capacity right now
Boostrun
GPU cloud
$0.490/GPU-hr
8× RTX-4090 · $3.920/hr per node · pcie · US, Raleigh, NC
no capacity right now
RTX-4090 cluster nodes
Multi-GPU configurations, cheapest per GPU-hour first. For distributed training the interconnect matters as much as the rate — an NVLink or InfiniBand node and the same count of loose PCIe cards are not the same machine.
| Provider | Configuration | Interconnect | $/GPU-hr | Node/hr |
|---|---|---|---|---|
| Vast.ai | 2× RTX 4090 x2 · CN | — | $0.121 | $0.241/hr |
| Excesssupply | 8× rtx4090.8x.epyc-7k62.96vcpu.252gib.bdb373e6b5 · US, Casper, WY | pcie | $0.400 | $3.200/hr |
| Phyntecno capacity | 8× RTX4090x8 · US, Casper, WY | pcie | $0.400 | $3.200/hr |
| Boostrunno capacity | 8× RTX4090x8 · US, Raleigh, NC | pcie | $0.490 | $3.920/hr |
| Excesssupplyno capacity | 2× n3-RTX-4090x2 · NO, Central | pcie | $0.600 | $1.200/hr |
| Excesssupply | 4× n3-RTX-4090x4 · NO, Central | pcie | $0.600 | $2.400/hr |