在哪里出租 你的 AI GPU。

Clore.ai

Marketplace 实时

P2P GPU marketplace. Host lists a server, pays out in CLORE, BTC, or USD; renters pay per-hour for on-demand or spot capacity.

RunPod

Marketplace 实时

Community Cloud lets independent hosts list GPUs. Consumer-friendly onboarding, weekly USD payouts via Stripe Connect.

Akash Network

Decentralized

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

GPU marketplace — independent providers list machines and set their own prices, so rates move with competition rather than a fixed rate card. Renters top up by card or crypto and are billed per second, deploying containerized pods with persistent volumes. Almost all capacity is datacenter "secure" tier (a spot tier exists and can be reclaimed by its provider at any time). Note that a third of nodes have a minimum order of more than one GPU.

Vast.ai

Marketplace 实时

P2P marketplace — list consumer or datacenter GPUs at your own per-hour bid. Huge renter demand, weekly BTC/USD payouts.

io.net

Decentralized 实时

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

First-party AI cloud — Lambda owns and operates the GPUs. Transparent on-demand per-GPU pricing on B200, H100, GH200, A100. No marketplace commission, no host onboarding (Lambda is the operator).

Indian GPU cloud run by Cyfuture India Pvt Ltd (Noida). On-demand and 1-, 6- and 12-month reserved instances from 1 to 8 GPUs — NVIDIA H100 SXM, A100, L40S, V100 and RTX PRO 6000, AMD MI300X and MI325X, Intel Gaudi2, plus reserve-only B300 — billed per hour with no platform or egress fees. Cyfuture quotes prices in Indian rupees; RentGPU converts them to USD at the current exchange rate.

Fluence

Decentralized 实时

Decentralized compute marketplace — register a compute peer, accept jobs, settle via USDC + FLT token on-chain. ~20% Delegation Rate to stakers; provider keeps ~80%.

First-party GPU cloud billed per minute. A virtualized GPU tier (RTX A6000, L40, L40S, A100 80GB, H100) undercuts the dedicated passthrough tier on every card, and instances attach straight to VS Code or Cursor via the tnr CLI. Storage is billed separately per GB-hour and data egress is free. Crypto mining is prohibited by their terms of service.

DeepInfra

First-party cloud 实时

First-party AI cloud — DeepInfra also rents B200 GPU instances on demand (1x/2x/4x/8x configs). Same provider as the hosted-inference API, just a different surface.

EC2 GPU instances (p4, p5, g5, g6) on the largest cloud network. Reserved + Savings Plan discounts available; spot prices fluctuate by region. Best fit when you need broader AWS services (S3, SageMaker, VPC peering) alongside the GPU.

Cudo Compute

Marketplace

Descended from Cudo Miner — hosts install the Cudo Compute Provider agent and earn on completed compute jobs. USD or BTC payouts.

Compute Engine A2/A3 (A100, H100), G2 (L4), and TPU instances. First-class integration with Vertex AI for managed training and serving. Sustained-use discounts auto-apply.

ND/NC/NV-series VMs covering A100, H100, H200, MI300X. Tight integration with Azure ML and OpenAI Service. Reserved Instances and Spot VMs offer up to 90% off list price.

AI and agent cloud platform providing on-demand GPU rentals for training and inference workloads with high performance and cost efficiency.

Bare-metal and VM GPU shapes (A10, A100, H100, H200) with RDMA cluster networking that often beats other hyperscalers on $/GPU-hour for large training jobs.

Render Network

Decentralized

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

Salad Cloud

Decentralized

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

Spheron Network

Decentralized

Decentralized compute marketplace with community-contributed GPU nodes. Stake + operate a provider, settle in crypto.

TensorDock

Marketplace

P2P marketplace with 100+ GPU models across home + datacenter hosts. Often the cheapest on-demand rates in the long tail.

Cloud GPU + Bare Metal GPU instances. Globally distributed datacenters (32 regions), per-hour billing, transparent published pricing. Sweet spot for teams wanting hyperscaler-style coverage without AWS-level complexity.