# Aquanode > aquanode.io | Compute; 1 record: free tier, limits, identity cost and evidence. HTML https://fshy.ajam.dev/p/aquanode/ | JSON https://fshy.ajam.dev/api/v1/entity/aquanode.json | how to read a record: https://fshy.ajam.dev/llms-full.txt ## Aquanode - `compute/aquanode` | paid-only | evidence documented | binding volume-quota | category GPU marketplace - needs card unknown, phone unknown, ID no, institution no | verified 2026-11-03T07:36:20.827Z (live) | stale on 2026-12-04 - url https://www.aquanode.io/pricing - free tier: no free tier or free credit is stated on the pages read: billing is per second against a credit balance that must be funded before a machine can be rented. | unit: none | metered on: wall-clock | read on: 2026-11-03 | source: [1] - per-object cap: not applicable | source: [2] | note: a machine record: vCPU, RAM and local storage per GPU vary by offer - retention: a stopped pod keeps its environment and volume in managed storage at USD 0.02 per GiB per 30-day month until it is deleted | source: [2] | note: held storage bills until the pod is deleted - egress: Network egress and ingress policy is set by each on-demand provider individually and is not uniform across the marketplace. | metered: no | source: [2] - identity: email | detail: Signup needs an email address that is verified; renting needs credit on the account, added by card or by Solana. Whether the card route asks for a billing address, a phone number or an ID check is unverified: behind a sign-up. - interface: cli, ssh - limits: - volume-quota (binding): the credit balance is the quota: credit must be on the account before any machine is rented, and usage draws it down per second; list prices are live minimums refreshed every 6 hours [1] - limits note: no free allowance; the on-demand list price is a live minimum and an L4 had no on-demand supply when read - binding note: the same as limits note - interface detail: an `aq` CLI covers the console surface and `aq ssh ` writes an SSH alias; a REST API was not found on the pages read - tooling: the `aq` CLI (`aq ssh`, `aq stop`, `aq start`, `aq down`) and a web console; a stopped pod can be restarted on a different card or provider - composes with: One account and one balance across several GPU providers, billed per second, with a pod (environment plus volume) that survives a stop and can restart on another provider. The price is a live minimum across current supply, so it moves between page loads. - reliability: preemptible offers run on hardware contributed by independent hosts, so they can be taken back; on-demand offers run in operator-managed facilities - notes: A marketplace over several GPU hosts. On-demand list rates per GPU-hour read on the vendor's pages (live, refreshed every 6 hours): H100 80GB USD 2.78; A100 80GB USD 1.49; L40S 48GB USD 1.07; RTX 4090 24GB USD 0.760; L4 24GB USD 1.10 with no on-demand supply at the time (the cheapest offer overall). Including preemptible offers the per-model pages quote a lower floor: A100 USD 1.23, L40S USD 0.881, RTX 4090 USD 0.394. The pricing page lists 15 GPU models priced on-demand; a third-party index lists 38. - evidence: - [3] 2026-11-03: H100 on-demand list rate USD 2.78 per GPU-hour, live range 2.78 to 5.32 | capture: sha256:bf17981d95e100112028509f48388f9cd57a2fde56a0efec70c3b9c08df42267 | quote: The cheapest current on-demand rate for the H100 on Aquanode is $2.78/GPU/hr. Live rates range from $2.78 to $5.32 per GPU per hour - [4] 2026-11-03: A100 on-demand USD 1.49 per GPU-hour, live range 1.23 to 1.72 (the vendor pricing page lists the A100 at 80 GB) | capture: sha256:cf6ed276587eecfd09ba6a810c8b6ceff027bc185ae4bd5a5440d69c9c24c21d | quote: The cheapest current on-demand rate for the A100 on Aquanode is $1.49/GPU/hr. Live rates range from $1.23 to $1.72 per GPU per hour - [5] 2026-11-03: L40S on-demand USD 1.07 per GPU-hour, live range 0.881 to 3.81 | capture: sha256:6ca38359cd3b8c49af880b57716ac45e8a5d27ca6292a5d7b55758e9e71e6fe3 | quote: The cheapest current on-demand rate for the L40S on Aquanode is $1.07/GPU/hr. Live rates range from $0.881 to $3.81 per GPU per hour - [6] 2026-11-03: RTX 4090 on-demand USD 0.760 per GPU-hour, live range 0.394 to 1.000 | capture: sha256:fa90030bc9269a18d18135542dbcf075a148359beed3b6ae9e5663e4c3942e22 | quote: The cheapest current on-demand rate for the RTX 4090 on Aquanode is $0.760/GPU/hr. Live rates range from $0.394 to $1.000 per GPU per hour - [7] 2026-11-03: L4 at USD 1.10 per GPU-hour, with no on-demand supply at the time read | capture: sha256:d362d11a1dd7f4e17737f9014dff2b13ab97e92f44f25edc8f05f740e012c37f | quote: The cheapest current on-demand rate for the L4 on Aquanode is $1.10/GPU/hr (no on-demand supply right now; this is the cheapest offer overall). - [2] 2026-11-03: prices are live and refreshed every 6 hours, billed per second with no contract or minimum term | capture: sha256:3f40be6d4954693b82608b4e93356523ccb9969145b6556a540adbb374ceacc2 | quote: Live prices for every GPU you can rent on Aquanode, refreshed every 6 hours. Billed per second, with no contracts and no minimum term. - [2] 2026-11-03: offers are on-demand or preemptible, and preemptible hardware belongs to independent hosts | capture: sha256:3f40be6d4954693b82608b4e93356523ccb9969145b6556a540adbb374ceacc2 | quote: Every offer on Aquanode is tagged on-demand or preemptible. On-demand capacity runs in operator-managed facilities with consistent uptime expectations; preemptible compute is real, often cheaper, hardware contributed by independent hosts. - [1] 2026-11-03: signup needs a verified email address, and credit on the account before a machine can be rented | capture: sha256:733bc7c3fdd315e0f5f93595057204d38b8c6c0ffb893172180735dece959e76 | quote: and verify your email. You will need credit on the account before you can rent a machine - [1] 2026-11-03: credit is added from Billing by card or by Solana | capture: sha256:733bc7c3fdd315e0f5f93595057204d38b8c6c0ffb893172180735dece959e76 | quote: Cards and Solana both work. - [2] 2026-11-03: egress terms vary by the underlying provider of each offer | capture: sha256:3f40be6d4954693b82608b4e93356523ccb9969145b6556a540adbb374ceacc2 | quote: Network egress and ingress policy is set by each on-demand provider individually and is not uniform across the marketplace. - [2] 2026-11-03: held environment and volume storage is USD 0.02 per GiB per 30-day month | capture: sha256:3f40be6d4954693b82608b4e93356523ccb9969145b6556a540adbb374ceacc2 | quote: Holding them in Aquanode's managed storage bills at $0.02 per GiB per 30-day month, metered per second. - [1] 2026-11-03: machines are reached with the aq CLI and plain SSH | capture: sha256:733bc7c3fdd315e0f5f93595057204d38b8c6c0ffb893172180735dece959e76 | quote: That manages the keypair and writes a `~/.ssh/config` alias, so plain `ssh ` works afterwards. - sources: - [1] https://docs.aquanode.io/docs/quickstart - [2] https://www.aquanode.io/pricing - [3] https://www.aquanode.io/gpu/nvidia-h100 - [4] https://www.aquanode.io/gpu/nvidia-a100 - [5] https://www.aquanode.io/gpu/nvidia-l40s - [6] https://www.aquanode.io/gpu/rtx-4090 - [7] https://www.aquanode.io/gpu/nvidia-l4