macrostack

NVIDIA B300 (Blackwell Ultra) · prices read 17 Nov 2026

What does it cost to rent a B300?

11 companies publish an on-demand price for the B300. Read on their own price pages, the cheapest is $7.10 per GPU-hour at Modal and the dearest $17.80 at AWS (P5, P6, G6) — 2.5× apart for the same chip. The middle price is $9.50. Every price below links to the page it was read on.

Cheapest on-demand · per GPU-hour

$7.10

Modal

Middle price · per GPU-hour

$9.50

of 11 published prices

Dearest on-demand · per GPU-hour

$17.80

AWS (P5, P6, G6) · 2.5× the cheapest

What a month costs

The rental alone, at the cheapest, middle and dearest published on-demand price. Storage, network and data transfer are usually billed on top and are not included.

How muchCheapestMiddleDearest
1 GPU for 100 hours$710$950$1,780
1 GPU for a working month (160 h)$1,136$1,520$2,848
1 GPU always on (730 h)$5,182$6,935$12,995
8 GPUs always on (730 h)$41,459$55,480$103,964

Work out your own B300 bill →Any number of hours, 1 or 8 GPUs, every company side by side.

Every published on-demand B300 price, cheapest first

One row per company: its own fixed on-demand price per GPU per hour, rentable on the day it was read. Ordered by price and nothing else. Each row says how its number was made and how fresh it is — current for a week, aging to 30 days, stale after that; the site will not build with an on-demand price older than 45 days.

#CompanyTierPer GPU-hourFreshnessSource
1ModalNvidia B300Base price: GPU time only, preemptible by default. Non-preemptible runs cost 3x; CPU and memory are extra.
How this number was made

Per GPU per second x 3600, no division.

Per-second billing: $0.001972/sec x 3600. GPU only; CPU/memory extra. Base (preemptible) rate; non-preemptible = 3x.

Region: Not pinned: base price lets Modal place the job; region selection costs 1.15-1.75x

On-demand$7.10Current · 17 Nov 2026their page ↗
2Hyperstack (NexGen Cloud)NVIDIA B300 (288 GB)
How this number was made

Already per GPU

Per hour per GPU, in the On-Demand GPU Pricing table; billed per minute. The page's top banner still reads 'NVIDIA B300s are coming to Hyperstack — On-Demand in August, reserved private clusters in Q4'; the table lists the price with nothing marking it unavailable.

Region: not stated on page

On-demand$7.40Current · 17 Nov 2026their page ↗
3Civo8 x NVIDIA B300
How this number was made

8-GPU instance $59.92/hr ÷ 8

Per hour, 8 x NVIDIA B300 instance (the only B300 size listed)

Region: not stated on page

On-demand$7.49Current · 17 Nov 2026their page ↗
4CrusoeB300 288GB HGX
How this number was made

Per GPU-hour, column 'Current on-demand'. The page's 'New on-demand, starting Dec 1, 2026' column reads $9.90; the price in force on the read date is recorded.

Region: not stated on page

On-demand$8.96Current · 17 Nov 2026their page ↗
5RunPodB300 288GB (low = Secure Cloud Pod; high = Instant Clusters, multi-node)
How this number was made

none — page prices per GPU

Secure Cloud Pod $8.99 to Instant Clusters $10.79, both on-demand with no commitment. Read on runpod.io/pricing 17 Nov 2026 (page: 'Updated November 12, 2026').

Region: not stated (30+ regions)

On-demand$8.99–$10.79Current · 17 Nov 2026their page ↗
6NebiusNVIDIA HGX B300
How this number was made

Per GPU-hour, column 'On-demand, GPU-hour'.

Region: not stated on page

On-demand$9.50Current · 17 Nov 2026their page ↗
7Together AINVIDIA HGX B300
How this number was made

Printed per GPU per hour, no division.

GPU Clusters, on-demand, pay as you go; hourly. 'All prices are per GPU per hour.'

Region: Not stated on the page

On-demand$11.49Current · 17 Nov 2026their page ↗
8DataCrunch (now Verda)1x B300 SXM6 268GB (high = B300 CC, confidential computing)
How this number was made

Already per GPU (1x instance)

Per hour, 1x B300 SXM6 268GB $12.53; the confidential-computing 1x B300 CC SXM6 268GB $12.78. Read on verda.com/pricing 17 Nov 2026 (USD shown).

Region: not stated on page

On-demand$12.53–$12.78Current · 17 Nov 2026their page ↗
9Fireworks AIB300 288 GB
How this number was made

Printed per GPU, no division.

Billed per GPU second; Fireworks prints $0.250/min and $15.00/hr; the hourly figure is used.

Region: Not pinned (region-restricted deployments cost 1.5x, via sales)

On-demand$15.00Current · 17 Nov 2026their page ↗
10Oracle Cloud (OCI)B300 263GB (BM.GPU.B300.8, bare metal, 8 GPUs)
How this number was made

Oracle prices per GPU (metric 'GPU Per Hour'), PAY_AS_YOU_GO, no division. BM.GPU.B300.8, part B112237. The page's price cells are filled from Oracle's public price-list API (https://apexapps.oracle.com/pls/apex/cetools/api/v1/products/?partNumber=B112237&currencyCode=USD, lastUpdated 2026-11-02), which is where the figure was read.

Pay-as-you-go.

Region: Not region-specific (Oracle's USD price list carries one rate per part number)

On-demand$15.00Current · 17 Nov 2026their page ↗
11AWS (P5, P6, G6)B300 (p6-b300.48xlarge: 8 NVIDIA Blackwell Ultra GPUs, 2,144 GB HBM3e in total; AWS does not use the word SXM)
How this number was made

p6-b300.48xlarge on-demand Linux $142.416/hr ÷ 8 GPUs. Read from the price file AWS's on-demand page loads (b0.p.awsstatic.com/pricing/2.0/meteredUnitMaps/ec2/USD/current/ec2-ondemand-without-sec-sel/US East (N. Virginia)/Linux/index.json, published 2026-11-14T01:29:26Z); the page's table is rendered from it.

Billed per second.

Region: us-east-1 (US East, N. Virginia)

On-demand$17.80Current · 17 Nov 2026their page ↗

Listed apart, never ranked

Cheaper, but not a price you can count on (6)

Spot, preemptible, community and marketplace listings. They can be interrupted, or the price is set by an independent host and moves with them.

Reserved, committed and promo prices (5)

A contract term or a limited offer, not a price for renting by the hour. Published ones are listed here; negotiated rates are lower and are not published by anyone.

Listed, but not comparable

  • CoreWeave rents it without a public price — Listed with on-demand 'Contact sales'; only a spot price is published.. their page ↗
  • Vultr rents it without a public price — Cloud GPU listing reads 'Contact Sales' ('Contact sales to reserve capacity now!').. their page ↗

Questions

How much does it cost to rent a B300 per hour?
Between $7.10 and $17.80 per GPU-hour on the published on-demand prices of 11 companies, read on their own price pages 17 Nov 2026. The middle price is $9.50: about half the companies charge less, about half more. Modal was the cheapest and AWS (P5, P6, G6) the dearest — 2.5× apart for the same chip.
What is the cheapest place to rent a B300?
On fixed on-demand prices, Modal at $7.10 per GPU-hour, read 17 Nov 2026. Spot, community and marketplace listings can go lower, but they can be interrupted or depend on someone else's machine, so they are listed separately and never ranked. Reserved and committed contracts are lower again — and nobody publishes what large buyers actually pay.
How much does a B300 cost per month?
Running one B300 around the clock (730 hours) costs $5,182 at the cheapest published on-demand price, $6,935 at the middle price and $12,995 at the dearest. A machine of eight costs eight times that. Storage, network and data transfer are usually extra and are not included.
Why is the same GPU so much cheaper at one company than another?
Because you are not only buying the chip. The cheapest end is often brokered or resold capacity, priced per GPU with everything else extra. The dearest end is a hyperscaler list price for a whole 8-GPU machine with its CPUs, memory and storage, which exists mainly to make the committed rate look like a discount. Between them sit differences in support, networking, region, billing and contract terms — and in who the counterparty is.

How these numbers were made, and how to reuse them

Macrostack does not rent clusters and does not benchmark them. Every figure is a published price read on the company's own price page on the date in its row. Committed, reserved and negotiated rates are lower than anything shown as a price here, and the real ones are not published by anyone.

Each price read on the company's own public price page on the date shown. On-demand per GPU per hour in USD; multi-GPU machines divided by GPU count (see normalization); per-second billing converted to per hour. Not measured by Macrostack. Committed, reserved and negotiated rates are lower and are not published by anyone. The full table — every chip, every company, 25 B300 rows among them — is free to reuse under CC BY 4.0 with credit and a link.

Cite as

Macrostack (2026). B300 rental prices, read 17 Nov 2026. CC BY 4.0. https://www.macrostack.net/compute/gpu-prices/b300

Every chip side by side: the GPU cost calculator · Other chips: H100 SXM rental prices · H200 rental prices · B200 rental prices · A100 80GB rental prices · L40S rental prices · L4 rental prices · MI300X rental prices · RTX PRO 6000 rental prices · A100 40GB rental prices · T4 rental prices · A10 rental prices · The chip itself: NVIDIA B300 (Blackwell Ultra) specs · Who these companies are: Layer 2.5 — Compute.

Prices read 17 Nov 2026.

The Macrostack brief

New swaps, worth your inbox.

A short, occasional email when we add a high-intent alternative or ship a new head-to-head. No spam, no selling your address — unsubscribe in one click.