NVIDIA B200 — what a GPU-hour costs
Listed rates, not availability. 34 rates read between and . · Renting for a year or more? Reserved quotes ↓
In the last closed month (September 2026) the B200 like-for-like median rose 3.5%, the second month running. Over 12 months, +10%.
Size your workload
Who is cheapest today, and what a month costs.
GPUsHours per monthRegionB200 by provider
One table per market, on-demand first; cheapest first within each.
| Provider · what is sold | $ / GPU-hour | 30 days | Read | Source |
|---|---|---|---|---|
 Sestercefrom | $4.11−45% vs median | stoppedlast confirmed · not in the figures | ||
 Beamfrom | $4.11−45% vs median | flat | captured | Deploy ↗ |
 Hostingerfrom | $4.50−40% vs median | flat · 21 d | captured | Deploy ↗ |
 Bitdeer AIfrom$9.38/node ÷ 2 | $4.69−37% vs median | flat · 13 d | captured | Deploy ↗ |
 Deep Infra7d +32.5% | $4.89−35% vs median | +32.5% · 26 d | captured | Deploy ↗ |
 TensorPoolGPU only | $4.99−33% vs median | stoppedlast confirmed · not in the figures | ||
 GMI Cloudfrom | $5.00−33% vs median | +25.0% | captured | Deploy ↗ |
 DataoortsAtlas Series · 7d +27.8% | $5.10−32% vs median | +27.8% · 23 d | captured | Deploy ↗ |
 Hyperbolicfrom | $5.99−20% vs median | flat · 26 d | captured | Deploy ↗ |
 Hyperstack | $6.00−20% vs median | flat | captured | Deploy ↗ |
 DaytonaGPU only | $6.25−17% vs median | flat · 27 d | captured | Deploy ↗ |
 Lambda | $6.69−11% vs median | flat | captured | Deploy ↗ |
| Elice Cloud | $6.82−9% vs median | flat · 11 d | captured | Deploy ↗ |
 Barrack AI | $6.89−8% vs median | flat · 10 d | captured | Deploy ↗ |
 packet.aiDedicated | $6.99−7% vs median | flat · 27 d | captured | Deploy ↗ |
| HostingerfromDedicated | $7.08−6% vs median | flat · 21 d | captured | Deploy ↗ |
 HexGridSecure Cloud | $7.500% vs median | flat · 27 d | captured | Deploy ↗ |
 HPC-AI.COMfrom | $7.500% vs median | +87.5% · 10 d | captured | Deploy ↗ |
 E2E Networks | $7.99+7% vs median | +14.3% | captured | Deploy ↗ |
 GPU.aifrom | $7.99+7% vs median | +17.7% · 22 d | captured | Deploy ↗ |
 RunPodAPISecure Cloud · the API request | $7.99+7% vs median | +17.7% | captured | Deploy ↗ |
 Together AI | $8.19+9% vs median | flat | captured | Deploy ↗ |
 VerdaAPI7d +5.6% | $8.36+11% vs median | +25.1% | captured | Deploy ↗ |
 Crusoe | $8.50+13% vs median | unconfirmedlast confirmed · not in the figures | Deploy ↗ | |
 Nebius | $8.50+13% vs median | flat | captured | Deploy ↗ |
 CoreWeave$68.80/node ÷ 8 | $8.60+15% vs median | flat | captured | Deploy ↗ |
 Lyceum | $8.79+17% vs median | +35.4% · 27 d | captured | Deploy ↗ |
 Hugging Faceaws | $9.25+23% vs median | flat · 27 d | captured | Deploy ↗ |
 Edgevana$77.06/node ÷ 8 | $9.63+28% vs median | stoppedlast confirmed · not in the figures | ||
 Baseten | $9.98+33% vs median | flat | captured | Deploy ↗ |
 EmpirioLabs7d +1.7% | $12.20+63% vs median | +74.5% · 27 d | captured | Deploy ↗ |
 Fireworks AIGPU only | $13.00+73% vs median | flat · 21 d | captured | Deploy ↗ |
 Oracle Cloud (OCI)API | $14.00+87% vs median | flat | captured | Deploy ↗ |
 Amazon Web ServicesAPI$113.93/node ÷ 8 · us-east-1 | $14.24+90% vs median | flat | captured | Deploy ↗ |
 Google CloudGPU onlyAPIus-west1 · same rate in all 3 regions read | $16.11+115% vs median | flat | captured | Deploy ↗ |
| Provider · what is sold | $ / GPU-hour | 30 days | Read | Source |
|---|---|---|---|---|
| Nebiusfrom | $0.99−87% vs median | unconfirmedlast confirmed · not in the figures | Deploy ↗ | |
| Google CloudGPU onlyAPIus-east5 · cheapest of 11 regions read; dearest $9.67 | $1.45−81% vs median | −1.7% | captured | Deploy ↗ |
 VultrAPI$25.60/node ÷ 8 | $3.20−57% vs median | flat · 24 d | captured | Deploy ↗ |
| DaytonaGPU only−$2.66 (−43%) vs its on-demand $6.25 | $3.59−52% vs median | flat · 27 d | captured | Deploy ↗ |
| VerdaAPI$8.36/node ÷ 2 · −$4.18 (−50%) vs its on-demand $8.36 · 7d +5.6% | $4.18−44% vs median | +25.1% | captured | Deploy ↗ |
| CoreWeave$34.11/node ÷ 8 · North America · −$4.34 (−50%) vs its on-demand $8.60 | $4.26−43% vs median | flat | captured | Deploy ↗ |
| CoreWeave$34.87/node ÷ 8 · Europe · −$4.24 (−49%) vs its on-demand $8.60 | $4.36−42% vs median | flat | captured | Deploy ↗ |
| Lyceum−$4.30 (−49%) vs its on-demand $8.79 | $4.49−40% vs median | +36.5% · 16 d | captured | Deploy ↗ |
| Elice Cloud−$0.68 (−10%) vs its on-demand $6.82 | $6.14−18% vs median | flat · 11 d | captured | Deploy ↗ |
| Provider · what is sold | $ / GPU-hour | 30 days | Read | Source |
|---|---|---|---|---|
| RunPodAPICommunity Cloud · the API request | $5.98−20% vs median | flat | captured | Deploy ↗ |
 LiumfloorAPI | $6.00−20% vs median | thin marketfewer offers than a floor needs · last floor · not in the figures | Deploy ↗ | |
 Vast.aifloorAPI | $7.50+0% vs median | thin marketfewer offers than a floor needs · last floor · not in the figures | Deploy ↗ | |
| GPU.aifrom | $8.38+12% vs median | stoppedlast confirmed · not in the figures | ||
| LiumfloorAPI$69.84/node ÷ 8 · Multi-GPU machines | $8.73+16% vs median | thin marketfewer offers than a floor needs · last floor · not in the figures | Deploy ↗ | |
| Vast.aifloorAPI$20.00/node ÷ 2 · Multi-GPU machines · 7d +6.7% | $10.00+33% vs median | +6.6% | captured | Deploy ↗ |
| Provider · what is sold | $ / GPU-hour | 30 days | Read | Source |
|---|---|---|---|---|
| packet.aiDynamic | $3.75−50% vs median | flat · 26 d | captured | Deploy ↗ |
| Provider · what is sold | $ / GPU-hour | 30 days | Read | Source |
|---|---|---|---|---|
 Koyeb | $5.50−27% vs median | flat | captured | Deploy ↗ |
 CerebriumGPU only | $6.01−20% vs median | flat | captured | Deploy ↗ |
 ModalGPU only | $6.25−17% vs median | flat | captured | Deploy ↗ |
 fal | $7.99+7% vs median | +27.8% | captured | Deploy ↗ |
| Provider · what is sold | $ / GPU-hour | 30 days | Read | Source |
|---|---|---|---|---|
| Hyperstackfrommin. commitment commit | $5.10−32% vs median | flat | captured | Reserve ↗ |
| Together AIfrommin. commitment commit | $6.79−9% vs median | flat | captured | Reserve ↗ |
| Provider · what is sold | $ / GPU-hour | 30 days | Read | Source |
|---|---|---|---|---|
| Bitdeer AIfrom$8.72/node ÷ 2 · −$0.33 (−7%) vs its on-demand $4.69 | $4.36−42% vs median | flat · 13 d | captured | Reserve ↗ |
 Civo$35.92/node ÷ 8 | $4.49−40% vs median | flat · 22 d | captured | Reserve ↗ |
| Google CloudGPU onlyAPIus-west1 · same rate in all 7 regions read · −$4.99 (−31%) vs its on-demand $16.11 | $11.12+48% vs median | flat | captured | Reserve ↗ |
| Provider · what is sold | $ / GPU-hour | 30 days | Read | Source |
|---|---|---|---|---|
| Civo$30.32/node ÷ 8 | $3.79−49% vs median | flat · 22 d | captured | Reserve ↗ |
| Bitdeer AIfrom$8.16/node ÷ 2 · −$0.61 (−13%) vs its on-demand $4.69 | $4.08−46% vs median | flat · 13 d | captured | Reserve ↗ |
| Google CloudGPU onlyAPIus-west2 · same rate in all 7 regions read | $7.09−5% vs median | flat | captured | Reserve ↗ |
Other cards today, median rate card per GPU-hour: H100 SXM $3.53 (Price Index) · H200 $4.56 · B300 $8.99 · GB300 NVL72 $18.00 · MI325X $2.25 · RTX 4090 $0.49 · all cards
Renting the B200 for a year or more?
The cheapest committed B200 rate each provider publishes, by term. A commitment is paid whether the GPU runs or not.
| Term | Per GPU-hour | Provider | vs its on-demand |
|---|---|---|---|
| 1-year | $4.36 read | Bitdeer AI cheapest of 3 providers | −7% ($4.69) |
| 3-year | $3.79 read | Civo cheapest of 3 providers | — |
| Other commitment commit | $5.10 read | Hyperstack cheapest of 2 providers | — |
Across every card, a 1-year rate runs a median −15% against the same provider’s on-demand rate (7 providers); a 3-year rate −30% (5 providers). Every committed rate, card by card.
Get quotes for the B200. A person at GPU Economy reads your request and passes it to up to three providers that list a rate for the card, the cheapest listed first; the quotes come from them, directly to you.
We may be paid a fee if a request becomes a contract; who receives it follows the listed rates, never a fee, and it changes no published rate or the Price Index (how, privacy).
As of , the cheapest listed rate for the NVIDIA B200 is $1.45 per GPU-hour on Google Cloud (spot, us-east5). The cheapest on-demand rate is $4.11 on Beam (from). Across 34 providers, the median rate card — each provider’s cheapest on-demand rate, community capacity and marketplace floors out — is $7.50 per GPU-hour, and the middle half of those rate cards sits between $6.00 and $8.57. Between rate cards, on-demand rates for the same card run from $4.11 (Beam) to $16.11 (Google Cloud), 3.9× apart across 34 providers. By market, the cheapest listed rates are $1.45 spot (Google Cloud, GPU only), $5.98 on community capacity (RunPod), $4.11 on-demand (Beam, from), $5.10 reserved with no stated term (Hyperstack, from), $4.36 on a 1-year commitment (Bitdeer AI, from) and $3.79 on a 3-year commitment (Civo). On a single rate card, Bitdeer AI lists it at $4.69 on-demand, $4.36 on a 1-year commitment and $4.08 on a 3-year commitment: a commitment is paid whether the GPU runs or not. With 192 GB of VRAM, that is $0.021 per GB of memory per hour at the cheapest on-demand rate and $0.039 at the median.
The two cheapest on-demand rate cards for the NVIDIA B200, side by side on every card they both list: Beam vs Hostinger.
Dated pieces on the NVIDIA B200: — Dataoorts B200 went $3.99 to $5.10 on October 22, 2026; Beam $4.11 · — Bitdeer first read October 10, 2026: GB200 $6.94, H100 SXM $2.36 · — B200 rates ran $3.69 to $16.11 a GPU-hour on October 3, 2026. Each is frozen as of its date.
As of , 2 providers price the NVIDIA B200 on-demand as a whole 8-GPU node (HGX, as CoreWeave and Nebius name it in their own listings): from $68.80 per node-hour (CoreWeave, HGX B200, $8.60 per GPU) to $113.93 (Amazon Web Services, p6-b200.48xlarge, us-east-1, $14.24 per GPU), 1.7× apart. 8 GPUs at the cheapest on-demand rate ($4.11 on Beam, from) come to $32.88 an hour, where a provider rents 8 to one account and they need not share a unit. A unit price is what the provider quotes for the 8 GPUs together; whether it also rents one is not part of the quote. 1 prices it in units of 2 — the rate table below names each quote's own unit.
How it is computed
What we measure
Publicly listed rental prices for 79 GPU models (NVIDIA and AMD) at 99 providers — 1165 current quotes across seven markets — and the retail price of the cards.
Where it comes from
Each provider's own pricing page or API; retail from reseller pages. Every figure links back to its source.
How it is normalised
Node prices are divided by GPU count, per-second rates multiplied out, a price listed in another currency shown as listed, beside its USD figure at the dated official rate (ECB; the Bank of Russia for RUB and VND). A reservation published as a total for its term is divided by the hours in that term, 8,760 per year.
What is modelled, not measured
One figure: cost per million output tokens, a listed rate divided by the throughput NVIDIA published for that card, published with the run that produced it in /token-cost.json, not on this page. Risk is not a model — it is read off the market a quote is from.
What we never do
Never estimate a price, never convert a currency silently, never publish a number a reader cannot verify on the linked page.
1165 quotes · 99 providers. Seven files under CC BY 4.0; history is stored append-only and never rewritten.
Weekly history since 2025, with a like-for-like table per provider: NVIDIA B200 price history.
Retail listings
As of , the NVIDIA B200 is listed new at $55,198.85 by Buy Direct & Save — one seller’s store price, not a market median; we publish a median only with three or more sellers.
| Seller | Condition, kind | Price | Seen | Source |
|---|---|---|---|---|
| Buy Direct & Save | new store price | $55,198.85 | Listing | |
| Orange Hardwares | new store price | $73,810 | Listing |
At the cheapest on-demand rate ($4.11/hour on Beam, from), that purchase price equals about 13,430 GPU-hours of rental — roughly 18 months of continuous use. The lower spot rate ($1.45 on Google Cloud) is left out: that capacity can be reclaimed or moves with supply. Power, hosting, cooling and idle time are not counted.
Compare the NVIDIA B200 with: NVIDIA H200 · NVIDIA B300 · NVIDIA GB200 NVL72 · NVIDIA H100 SXM
Every rate on this page as data: /gpu/b200/prices.json — the same fields as prices.json, narrowed to this card. CC BY 4.0, no key, no rate limit. On your own page: embed this card’s cheapest rates (/embed/gpu/b200, an iframe, no script).
Data updated: