{ "name": "aiimplementationcost.com GPU Compute Price List", "title": "GPU Compute Price List", "url": "https://aiimplementationcost.com/datasets/gpu-compute-price-list/", "csv": "https://aiimplementationcost.com/datasets/gpu-compute-price-list.csv", "markdown": "https://aiimplementationcost.com/datasets/gpu-compute-price-list/index.md", "computed_at": "2026-10-23", "license": "https://aiimplementationcost.com/terms", "headline": "The GPU Compute Price List carries 15 published figures read from 5 sources' own pages, from 0.48 to 12.42 USD, read 12 October 2026.", "quotes": [ "The GPU Compute Price List carries 15 published figures read from 5 sources' own pages, from 0.48 to 12.42 USD, read 12 October 2026.", "5 of the 24 named sources read in full publish a figure; 19 publish none, read 12 October 2026.", "GPU Compute Price List, accelerator H100 SXM: median 2.69 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator A100: median 1.09 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator H100 PCIe: median 2.01 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator H100 SXM5: median 2.57 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator HGX B200 (2 weeks-1 year commitment): median 8.87 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator P100: median 1.27 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator H100: median 3.20 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator B200 SXM6: median 9.12 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator B300: median 12.42 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator A100 80GB: median 1.48 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator H100 SXM5 (spot): median 1.98 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator A100 80GB (spot): median 1.24 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator B200 SXM6 (spot): median 3.73 USD across 1 source, read 12 October 2026.", "GPU Compute Price List, accelerator RTX 4090: median 0.48 USD across 1 source, read 12 October 2026." ], "citation": "\"aiimplementationcost.com GPU Compute Price List\", updated 2026-10-23, https://aiimplementationcost.com/datasets/gpu-compute-price-list/.", "changed_since": null, "national": { "unit": "USD", "n": 14, "median": 2.29, "min": 0.48, "max": 12.42, "p25": 1.2662499999999999, "p75": 5.015, "mean": 3.7254, "units": { "USD": 14 }, "published_sources": 5, "named_sources": 25, "read_sources": 24, "not_published": 19, "published_share": 0.208 }, "by_axis": { "accelerator": { "H100 SXM": { "USD": { "n": 1, "median": 2.69, "min": 2.69, "max": 2.69, "p25": 2.69, "p75": 2.69, "mean": 2.69 } }, "A100": { "USD": { "n": 1, "median": 1.09, "min": 1.09, "max": 1.09, "p25": 1.09, "p75": 1.09, "mean": 1.09 } }, "H100 PCIe": { "USD": { "n": 1, "median": 2.01, "min": 2.01, "max": 2.01, "p25": 2.01, "p75": 2.01, "mean": 2.01 } }, "H100 SXM5": { "USD": { "n": 1, "median": 2.5700000000000003, "min": 2.5700000000000003, "max": 2.5700000000000003, "p25": 2.5700000000000003, "p75": 2.5700000000000003, "mean": 2.57 } }, "HGX B200 (2 weeks-1 year commitment)": { "USD": { "n": 1, "median": 8.87, "min": 8.87, "max": 8.87, "p25": 8.87, "p75": 8.87, "mean": 8.87 } }, "P100": { "USD": { "n": 1, "median": 1.275, "min": 1.275, "max": 1.275, "p25": 1.275, "p75": 1.275, "mean": 1.275 } }, "H100": { "USD": { "n": 1, "median": 3.2, "min": 3.2, "max": 3.2, "p25": 3.2, "p75": 3.2, "mean": 3.2 } }, "B200 SXM6": { "USD": { "n": 1, "median": 9.12, "min": 9.12, "max": 9.12, "p25": 9.12, "p75": 9.12, "mean": 9.12 } }, "B300": { "USD": { "n": 1, "median": 12.42, "min": 12.42, "max": 12.42, "p25": 12.42, "p75": 12.42, "mean": 12.42 } }, "A100 80GB": { "USD": { "n": 1, "median": 1.48, "min": 1.48, "max": 1.48, "p25": 1.48, "p75": 1.48, "mean": 1.48 } }, "H100 SXM5 (spot)": { "USD": { "n": 1, "median": 1.98, "min": 1.98, "max": 1.98, "p25": 1.98, "p75": 1.98, "mean": 1.98 } }, "A100 80GB (spot)": { "USD": { "n": 1, "median": 1.24, "min": 1.24, "max": 1.24, "p25": 1.24, "p75": 1.24, "mean": 1.24 } }, "B200 SXM6 (spot)": { "USD": { "n": 1, "median": 3.73, "min": 3.73, "max": 3.73, "p25": 3.73, "p75": 3.73, "mean": 3.73 } }, "RTX 4090": { "USD": { "n": 1, "median": 0.48, "min": 0.48, "max": 0.48, "p25": 0.48, "p75": 0.48, "mean": 0.48 } } } }, "rows": [ { "source": "NVIDIA H100 Price Guide 2026: GPU Costs, Cloud Pricing & Buy vs Rent", "url": "https://jarvislabs.ai/blog/h100-price", "figure": "$2.69 ", "value": 2.69, "unit": "USD", "vat": null, "axis": { "accelerator": "H100 SXM" }, "retrieved_at": "2026-10-12", "quote": "An NVIDIA H100 SXM with 80 GB of GPU memory costs $2.69 per GPU-hour to rent on Jarvislabs, billed by the minute.", "source_class": "primary-provider-own-site", "text_sha256": "5ab85b0a4bd22724fff9dcd630cc8f99a62f481e13c2f1d2c780dfa39a7e140d" }, { "source": "NVIDIA H100 Pricing (September 2026): Cheapest Cloud GPU Rates", "url": "https://www.thundercompute.com/blog/nvidia-h100-pricing", "figure": "1.09 $", "value": 1.09, "unit": "USD", "vat": null, "axis": { "accelerator": "A100" }, "retrieved_at": "2026-10-12", "quote": "Provider GPU Listed $/GPU-hr 10-Hour Cost Thunder Compute A100 $1.09 $10.90", "source_class": "primary-provider-own-site", "text_sha256": "6c35ce0b29ee0487483bf48a1c6ab5f2726eb9ba18a0834c69951dab86fd833d" }, { "source": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "figure": "$2.01", "value": 2.01, "unit": "USD", "vat": null, "axis": { "accelerator": "H100 PCIe" }, "retrieved_at": "2026-10-12", "quote": "Azure charges ~$12.29/hr per GPU on their ND H100 v5 instances. On Spheron, H100 PCIe is $2.01/hr on-demand.", "source_class": "secondary", "text_sha256": "96f3b877c969809e3b3ef8b3c1e1df86138a22f089ef930686e1bcfc35d9014f" }, { "source": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "figure": "$2.50 ", "value": 2.5, "unit": "USD", "vat": null, "axis": { "accelerator": "H100 SXM5" }, "retrieved_at": "2026-10-12", "quote": "Spheron (H100 SXM5) $2.50 $1.03 8-way HGX; lowest spot rate", "source_class": "secondary", "text_sha256": "96f3b877c969809e3b3ef8b3c1e1df86138a22f089ef930686e1bcfc35d9014f" }, { "source": "GPU cloud pricing: rent NVIDIA H100, H200, and B200", "url": "https://lambda.ai/pricing", "figure": "$8.87 ", "value": 8.87, "unit": "USD", "vat": null, "axis": { "accelerator": "HGX B200 (2 weeks-1 year commitment)" }, "retrieved_at": "2026-10-12", "quote": "NVIDIA HGX B200 2 weeks – 1 year 256+ $8.87 Talk to our team", "source_class": "primary-provider-own-site", "text_sha256": "5697851822329a868a64b389a980c3ea59d9e759018d652c51116f63882fde4e" }, { "source": "Cloud Price List", "url": "https://www.oracle.com/cloud/price-list/", "figure": "$1.275", "value": 1.275, "unit": "USD", "vat": null, "axis": { "accelerator": "P100" }, "retrieved_at": "2026-10-12", "quote": "VM.GPU2.1 1x NVIDIA P100 Pascal N/A 16 GB 12 78 GB 25 Gb/sec $1.275", "source_class": "primary-provider-own-site", "text_sha256": "96412b08eb42e057cff009fb3bcd07817ce8a2886ab239fbd9b64a7bd0f8f0c3" }, { "source": "NVIDIA H100 Pricing (September 2026): Cheapest Cloud GPU Rates", "url": "https://www.thundercompute.com/blog/nvidia-h100-pricing", "figure": "$3.20 ", "value": 3.2, "unit": "USD", "vat": null, "axis": { "accelerator": "H100" }, "retrieved_at": "2026-10-12", "quote": "Thunder Compute $3.20 $32.00", "source_class": "primary-provider-own-site", "text_sha256": "6c35ce0b29ee0487483bf48a1c6ab5f2726eb9ba18a0834c69951dab86fd833d" }, { "source": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "figure": "$9.12", "value": 9.12, "unit": "USD", "vat": null, "axis": { "accelerator": "B200 SXM6" }, "retrieved_at": "2026-10-12", "quote": "Spheron (B200 SXM6) $9.12/hr $3.73/hr Live rates", "source_class": "secondary", "text_sha256": "96f3b877c969809e3b3ef8b3c1e1df86138a22f089ef930686e1bcfc35d9014f" }, { "source": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "figure": "$2.64", "value": 2.64, "unit": "USD", "vat": null, "axis": { "accelerator": "H100 SXM5" }, "retrieved_at": "2026-10-12", "quote": "H100 SXM5 (Spheron) $2.64/hr $1.98/hr Varies, compare live rates", "source_class": "secondary", "text_sha256": "96f3b877c969809e3b3ef8b3c1e1df86138a22f089ef930686e1bcfc35d9014f" }, { "source": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "figure": "$12.42", "value": 12.42, "unit": "USD", "vat": null, "axis": { "accelerator": "B300" }, "retrieved_at": "2026-10-12", "quote": "B300 (Spheron) $12.42/hr $4.19/hr Varies, compare live rates", "source_class": "secondary", "text_sha256": "96f3b877c969809e3b3ef8b3c1e1df86138a22f089ef930686e1bcfc35d9014f" }, { "source": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "figure": "$1.48", "value": 1.48, "unit": "USD", "vat": null, "axis": { "accelerator": "A100 80GB" }, "retrieved_at": "2026-10-12", "quote": "A100 80GB (Spheron) $1.48/hr $1.24/hr Varies, compare live rates", "source_class": "secondary", "text_sha256": "96f3b877c969809e3b3ef8b3c1e1df86138a22f089ef930686e1bcfc35d9014f" }, { "source": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "figure": "$1.98", "value": 1.98, "unit": "USD", "vat": null, "axis": { "accelerator": "H100 SXM5 (spot)" }, "retrieved_at": "2026-10-12", "quote": "Cheapest H100 SXM5 per hour: Spheron H100 is $1.98/hr on spot and $2.64/hr on-demand as of 12 Oct 2026.", "source_class": "secondary", "text_sha256": "96f3b877c969809e3b3ef8b3c1e1df86138a22f089ef930686e1bcfc35d9014f" }, { "source": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "figure": "$1.24", "value": 1.24, "unit": "USD", "vat": null, "axis": { "accelerator": "A100 80GB (spot)" }, "retrieved_at": "2026-10-12", "quote": "Cheapest A100 80GB per hour: Spheron A100 is $1.48/hr on-demand and $1.24/hr on spot.", "source_class": "secondary", "text_sha256": "96f3b877c969809e3b3ef8b3c1e1df86138a22f089ef930686e1bcfc35d9014f" }, { "source": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "figure": "$3.73", "value": 3.73, "unit": "USD", "vat": null, "axis": { "accelerator": "B200 SXM6 (spot)" }, "retrieved_at": "2026-10-12", "quote": "Cheapest B200 per hour: Spheron B200 is $3.73/hr on spot and $9.12/hr on-demand.", "source_class": "secondary", "text_sha256": "96f3b877c969809e3b3ef8b3c1e1df86138a22f089ef930686e1bcfc35d9014f" }, { "source": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "figure": "$0.48", "value": 0.48, "unit": "USD", "vat": null, "axis": { "accelerator": "RTX 4090" }, "retrieved_at": "2026-10-12", "quote": "Cheapest RTX 4090 per hour: Vast.ai marketplace hosts go as low as $0.31/hr; Spheron lists RTX 4090 on-demand at $0.48/hr, and Runpod Community Cloud at $0.34/hr ($0.69/hr on Secure Cloud).", "source_class": "secondary", "text_sha256": "96f3b877c969809e3b3ef8b3c1e1df86138a22f089ef930686e1bcfc35d9014f" } ], "not_published": [ { "source": "VM instance pricing", "url": "https://cloud.google.com/compute/gpus-pricing", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "Pricing - Linux Virtual Machines", "url": "https://azure.microsoft.com/en-us/pricing/details/virtual-machines/linux/", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "Amazon EC2 P5 Instances", "url": "https://aws.amazon.com/ec2/instance-types/p5/", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "Rent NVIDIA GPUs on demand: H100, H200, and B200", "url": "https://lambda.ai/service/gpu-cloud", "retrieved_at": "2026-10-12", "reason": "read in full and no accepted figure on the page" }, { "source": "CoreWeave Cloud Pricing", "url": "https://www.coreweave.com/pricing", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "GPU Cloud Pricing", "url": "https://www.runpod.io/pricing", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "nebius.com", "url": "https://nebius.com/prices", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "GPU Pricing", "url": "https://vast.ai/pricing", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "GPU Droplets Pricing", "url": "https://www.digitalocean.com/pricing/gpu-droplets", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "Cloud GPU Rental Price Index", "url": "https://aimultiple.com/gpu-index", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "Cloud GPU Pricing - Compare 79 Providers and 4,900+ Prices (2026)", "url": "https://getdeploying.com/gpus", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "Pricing", "url": "https://www.paperspace.com/pricing", "retrieved_at": "2026-10-12", "reason": "read in full and no accepted figure on the page" }, { "source": "Price list: A comparison of our Public Cloud offers", "url": "https://www.ovhcloud.com/en/public-cloud/prices/", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "GPU Instances Pricing", "url": "https://www.scaleway.com/en/pricing/gpu/", "retrieved_at": "2026-10-12", "reason": "read in full and no accepted figure on the page" }, { "source": "Server with GPU: for your AI and machine learning projects.", "url": "https://www.hetzner.com/dedicated-rootserver/matrix-gpu/", "retrieved_at": "2026-10-12", "reason": "read in full and no accepted figure on the page" }, { "source": "Plan Pricing", "url": "https://www.modal.com/pricing", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "GPU pricing for AI compute & inference", "url": "https://www.crusoe.ai/cloud/pricing", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "Civo Prices - Transparent Cloud Pricing - Civo.com", "url": "https://www.civo.com/pricing", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" }, { "source": "Pricing", "url": "https://www.together.ai/pricing#gpu-clusters", "retrieved_at": "2026-10-12", "reason": "every candidate on the page was refused" } ], "sources": [ { "url": "https://cloud.google.com/compute/gpus-pricing", "label": "VM instance pricing", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://azure.microsoft.com/en-us/pricing/details/virtual-machines/linux/", "label": "Pricing - Linux Virtual Machines", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://aws.amazon.com/ec2/instance-types/p5/", "label": "Amazon EC2 P5 Instances", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://lambda.ai/service/gpu-cloud", "label": "Rent NVIDIA GPUs on demand: H100, H200, and B200", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.coreweave.com/pricing", "label": "CoreWeave Cloud Pricing", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.runpod.io/pricing", "label": "GPU Cloud Pricing", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://nebius.com/prices", "label": "nebius.com", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://vast.ai/pricing", "label": "GPU Pricing", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.oracle.com/cloud/price-list/", "label": "Cloud Price List", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.digitalocean.com/pricing/gpu-droplets", "label": "GPU Droplets Pricing", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://jarvislabs.ai/blog/h100-price", "label": "NVIDIA H100 Price Guide 2026: GPU Costs, Cloud Pricing & Buy vs Rent", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.thundercompute.com/blog/nvidia-h100-pricing", "label": "NVIDIA H100 Pricing (September 2026): Cheapest Cloud GPU Rates", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://lambda.ai/pricing", "label": "GPU cloud pricing: rent NVIDIA H100, H200, and B200", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.spheron.network/blog/gpu-cloud-pricing-comparison-2026/", "label": "GPU Cloud Pricing Comparison 2026: H100 From $2.01/hr", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://aimultiple.com/gpu-index", "label": "Cloud GPU Rental Price Index", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://getdeploying.com/gpus", "label": "Cloud GPU Pricing - Compare 79 Providers and 4,900+ Prices (2026)", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.paperspace.com/pricing", "label": "Pricing", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.ovhcloud.com/en/public-cloud/prices/", "label": "Price list: A comparison of our Public Cloud offers", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.scaleway.com/en/pricing/gpu/", "label": "GPU Instances Pricing", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.hetzner.com/dedicated-rootserver/matrix-gpu/", "label": "Server with GPU: for your AI and machine learning projects.", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.genesiscloud.com/pricing", "label": "www.genesiscloud.com", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.modal.com/pricing", "label": "Plan Pricing", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.crusoe.ai/cloud/pricing", "label": "GPU pricing for AI compute & inference", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.civo.com/pricing", "label": "Civo Prices - Transparent Cloud Pricing - Civo.com", "license": "the publisher's own terms; quoted for reporting and comment" }, { "url": "https://www.together.ai/pricing#gpu-clusters", "label": "Pricing", "license": "the publisher's own terms; quoted for reporting and comment" } ], "methodology": "Every row is a verbatim sentence read from the named source's own page through our fetch service on the date shown, hash-pinned to the stored read (the sha of the page's visible text; the raw page is mirrored to R2). 5 of the 24 named sources that could be read published a usable figure; 19 published none and are recorded as not published, never filled in; 1 could not be read (blocked, dead or unreachable) and count nowhere. The headline figures (median, minimum, maximum, quartiles) are over ONE figure per source for each breakdown it prices, in USD (a source's median where its page yielded several accepted sentences for one breakdown), so a page that gives figures for several breakdowns gives one for each and N counts 14 figures, never sentences; 382 candidate sentence(s) were refused by a reviewer with the reason recorded. The fuller derivation this dataset is chartered to grow into (not yet what the figures above are; they are each source's own published figure as read): one row per accelerator per provider at the provider's own published on-demand hourly rate, with the reserved and spot rate where published, normalised to a price per accelerator-hour (the hyperscalers quote whole 8-GPU instances and the specialists quote single cards, so the two are not comparable as published) and to a price per unit of memory; the hyperscaler-to-specialist premium, which is the finding; repriced-since", "embed": "

0.48 to 12.42 USD

14 published figures, GPU Compute Price List, across 5 sources, read 12 October 2026

Source: aiimplementationcost.com GPU Compute Price List
" }