Decentralized networks Live data

io.net.

Solana-based decentralized GPU network — host a worker node, earn IO token for completed compute jobs.

At a glance

Business model: Decentralized networks
Tier: Decentralized
Mining: Prohibited

When to pick io.net

Best for

Long-running batch jobs where the lowest spot price wins.
Token-paid workloads on networks like Akash or io.net.
Operators who want to diversify away from US cloud providers.

Avoid for

Anyone who needs a single point of support contact.
Workloads requiring specific datacenter regions.
Production inference where uptime guarantees matter.

Price history

Daily median on io.net's top GPUs.

GPUs available on io.net

GPU	Tier	VRAM	$/hr	Range	Offers
Nvidia RTX A4000	workstation	16GB	$0.18/hr	$0.18/hr–$0.18/hr	0	Compare →
Nvidia RTX A5000	workstation	24GB	$0.22/hr	$0.22/hr–$0.22/hr	0	Compare →
Nvidia GeForce RTX 3090	consumer	24GB	$0.27/hr	$0.27/hr–$0.27/hr	1	Compare →
Nvidia GeForce RTX 4090	consumer	24GB	$0.30/hr	$0.30/hr–$0.30/hr	792	Compare →
Nvidia RTX 4090 D	consumer	24GB	$0.30/hr	$0.30/hr–$0.30/hr	0	Compare →
Nvidia GeForce RTX 5090	consumer	32GB	$0.85/hr	$0.85/hr–$0.85/hr	0	Compare →
Nvidia L40S	datacenter	48GB	$0.89/hr	$0.89/hr–$0.89/hr	373	Compare →
Nvidia L4	datacenter	24GB	$0.95/hr	$0.95/hr–$0.95/hr	0	Compare →
Nvidia H100 NVL	datacenter	80GB	$1.00/hr	$1.00/hr–$1.00/hr	0	Compare →
Nvidia H100 PCIe	datacenter	80GB	$1.00/hr	$1.00/hr–$1.00/hr	14	Compare →
Nvidia A100 PCIe	datacenter	40GB	$1.39/hr	$1.39/hr–$1.39/hr	3	Compare →
Nvidia A100 SXM	datacenter	40GB	$1.39/hr	$1.39/hr–$1.39/hr	16	Compare →
Nvidia H100 SXM	datacenter	80GB	$2.09/hr	$2.09/hr–$2.09/hr	640	Compare →
Nvidia H200	datacenter	141GB	$2.99/hr	$2.99/hr–$2.99/hr	0	Compare →
Nvidia B200	datacenter	192GB	$4.99/hr	$4.99/hr–$4.99/hr	0	Compare →
Nvidia B300	datacenter	288GB	$6.73/hr	$6.73/hr–$6.73/hr	8	Compare →

Compare with peers

Other providers in the same bucket — quick way to sanity-check pricing before committing.

Salad Cloud

Catalog-only — no live pricing yet

Pays consumer PCs to run batch inference while idle. Install the Salad client, keep earning in the background. Consumer GPUs only.

Akash Network

Cheapest: $0.060/hr · Nvidia Tesla P4 · 20 GPUs

Cosmos-based deploy marketplace — run a provider node, auction compute to deployers, settle in AKT token.

Render Network

Catalog-only — no live pricing yet

P2P GPU-render marketplace. Paid per OctaneBench-hour served — best for Octane/Blender users with workstation-grade cards.

Fluence

80% host

Cheapest: $1.06/hr · Nvidia GeForce RTX 4090 · 3 GPUs

Decentralized compute marketplace — register a compute peer, accept jobs, settle via USDC + FLT token on-chain. ~20% Delegation Rate to stakers; provider kee...

Frequently asked

How is io.net billing measured?

Most providers bill per-second once the instance is running, with a small minimum (often 60 seconds). Some first-party clouds round up to the minute. Headline $/hr is the right comparison unit.

Which regions does io.net offer GPUs in?

P2P marketplaces aggregate hosts worldwide so region varies per offer. First-party clouds and hyperscalers expose explicit region pickers (US, EU, APAC). Filter on the provider's site after clicking through if region matters.

Does io.net offer an SLA?

Hyperscalers (AWS, GCP, Azure, Oracle) publish formal SLAs. First-party clouds (Lambda, CoreWeave) offer support contracts. P2P marketplaces and decentralized networks have no SLA — uptime depends on the individual host.

What's the refund / cancellation policy on io.net?

Per-second billing means you only pay for compute used — stop the instance and billing stops. Pre-paid credits and committed-use discounts have provider-specific terms; check the provider's billing docs before pre-paying.