What is io.net
io.net is a decentralised GPU cloud that aggregates independent compute capacity for training, fine-tuning and inference workloads.
On-demand pricing
As of August 5, 2026, io.net bills on-demand compute per second with no minimum contract. Its official April 2026 guide lists RTX 4090 at US$0.18 per GPU hour, RTX 3090 at US$0.28, L40S at US$0.75, A100 40GB at US$1.20, A100 80GB at US$1.49, H100 PCIe at US$1.49 and H100 SXM at US$2.20. Inventory and market pricing can change, so the live deployment quote remains authoritative.
Included services and enterprise
Standard use includes GPU compute, access to common open-source frameworks, monitoring, SSH and API access, container image storage, inter-GPU networking and the first 1TB of monthly data transfer. Published persistent storage is US$0.05 per GB per month and additional egress is US$0.05 per GB.
Enterprise reserved capacity uses custom terms and can provide volume discounts and paid support. New-user promotions, GPU availability and exact real-time rates should be confirmed in the cloud dashboard before purchase.
Official sources: io.net GPU pricing guide, io.net live cloud inventory.