From $0.69 per GPU-hour at Vast.ai on 2026-09-18 — 53% under the $1.47 median across 30 providers.
Prices collected 2026-09-18 · every provider's own public pricing
Individual on-demand configurations, in stock, cheapest first — a listing is one instance type, SKU or marketplace offer, priced per GPU. Marketplace offers are single machines and can be gone by the time you look.
| # | Provider | Configuration | GPUs | $/GPU/hr | Month | vs median |
|---|---|---|---|---|---|---|
| 1 | L40S ×1 · offer 50582903 · Czechia, CZ | 1 | $0.69 | $500 | −53% | |
| 2 | 1x l40s · listed | 1 | $0.76 | $547 | −48% | |
| 3 | L40S · Community Cloud | 1 | $0.79 | $569 | −46% | |
| 4 | 1x l40s · listed | 1 | $0.79 | $569 | −46% | |
| 5 | 1x l40s · listed | 1 | $0.80 | $576 | −46% | |
| 6 | L40S ×2 · offer 25918211 · Taiwan, TW | 2 | $0.80 | $577 | −46% | |
| 7 | L40S ×4 · offer 35897695 · Taiwan, TW | 4 | $0.80 | $577 | −46% | |
| 8 | L40S ×1 · offer 25990407 · Taiwan, TW | 1 | $0.80 | $578 | −45% | |
| 9 | 8x l40s · listed | 8 | $0.84 | $605 | −43% | |
| 10 | 4x l40s · listed | 4 | $0.84 | $605 | −43% |
Each provider reduced to its own median for the L40S — 15 of 30 sit below the market median.
| Provider | Kind | Median $/GPU/hr | vs market |
|---|---|---|---|
| neocloud | $0.76 | −48% | |
| marketplace | $0.79 | −46% | |
| marketplace | $0.80 | −46% | |
| marketplace | $0.80 | −46% | |
| marketplace | $0.89 | −39% | |
| neocloud | $0.92 | −37% | |
| marketplace | $0.96 | −35% | |
| neocloud | $0.97 | −34% | |
| neocloud | $0.97 | −34% | |
| neocloud | $0.98 | −34% | |
| neocloud | $1.09 | −26% | |
| neocloud | $1.20 | −18% | |
| neocloud | $1.20 | −18% | |
| neocloud | $1.29 | −12% | |
| neocloud | $1.44 | −2% | |
| neocloud | $1.50 | +2% | |
| neocloud | $1.53 | +4% | |
| hyperscaler | $1.57 | +7% | |
| neocloud | $1.69 | +15% | |
| hyperscaler | $1.80 | +22% | |
| neocloud | $1.95 | +33% | |
| neocloud | $1.95 | +33% | |
| neocloud | $1.95 | +33% | |
| neocloud | $2.14 | +46% | |
| neocloud | $2.25 | +53% | |
| hyperscaler | $2.30 | +56% | |
| neocloud | $2.92 | +98% | |
| hyperscaler | $3.39 | +130% | |
| hyperscaler | $3.50 | +138% | |
| neocloud | $3.51 | +139% |
Median of the providers headquartered in each country. Where the hardware sits can differ; this is who you contract with.
| Country | Providers | Median | Cheapest there |
|---|---|---|---|
| AEUAE | 1 | $0.80 | $0.79 at GPU.ai |
| ITItaly | 1 | $0.98 | $0.98 at Seeweb |
| GBUK | 1 | $1.29 | $1.29 at Civo |
| DEGermany | 1 | $1.44 | $1.19 at Lyceum |
| FRFrance | 2 | $1.50 | $1.20 at Koyeb |
| FIFinland | 1 | $1.53 | $1.45 at DataCrunch (Verda) |
| INIndia | 2 | $1.58 | $1.20 at E2E Networks |
| NLNetherlands | 1 | $1.69 | $1.55 at Nebius |
| USUSA | 17 | $1.95 | $0.69 at Vast.ai |
On 2026-09-18 the cheapest on-demand L40S we track is $0.69 per GPU-hour at Vast.ai (L40S ×1 · offer 50582903, Czechia, CZ), about $500 a month per GPU. The market median across 30 providers is $1.47, so the cheapest listing is 53% below it.
Yes. The spot/interruptible median is $0.82 against $1.47 on-demand — 44% off — from 1 provider. The cheapest spot listing is $0.74 at Nebius. Spot can be taken away mid-job, which is why it never enters the headline median.
Because the market is three markets. Marketplaces are independent hosts competing on price with consumer-grade uptime; neoclouds sell the GPU more directly; hyperscalers bundle CPU, RAM, network and an ecosystem into the instance price. The median sits where most providers sit; the cheapest listing is usually a marketplace host or a promotional neocloud rate. 15 of 30 providers are below the median today.
Prices are re-read every six hours from each provider's own pricing and every snapshot is kept, so this page reflects the latest run and the history behind it. Recent cuts: Vast.ai -8.2% on 2026-09-18; io.net -5.6% on 2026-09-18; GPU.ai -15.7% on 2026-09-18.