Decentralized networks Live data

Lium.

GPU marketplace on Bittensor Subnet 51, operated by Datura AI Corp. Independent providers list nodes and set their own prices, so rates move with competition rather than a fixed rate card. Almost all capacity is datacenter "secure" tier (a spot tier exists and can be reclaimed by its provider at any time). Renters deploy containerized pods with persistent volumes; providers earn rental fees daily plus subnet emission. Note that a third of nodes have a minimum order of more than one GPU.

At a glance

Business model: Decentralized networks
Tier: Decentralized
Payout: SN51 alpha (rental fees, daily) + TAO emission

When to pick Lium

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 Lium's top GPUs.

GPUs available on Lium

GPU	Tier	VRAM	$/hr	Range	Offers
Nvidia GeForce RTX 3090	consumer	24GB	$0.18/hr	$0.18/hr–$0.18/hr	3	Compare →
Nvidia GeForce RTX 4090	consumer	24GB	$0.35/hr	$0.30/hr–$0.58/hr	9	Compare →
Nvidia L40S	datacenter	48GB	$0.38/hr	$0.38/hr–$0.38/hr	1	Compare →
Nvidia RTX A6000	workstation	48GB	$0.46/hr	$0.42/hr–$0.49/hr	2	Compare →
Nvidia GeForce RTX 5090	consumer	32GB	$0.60/hr	$0.50/hr–$0.69/hr	9	Compare →
Nvidia RTX 6000 Ada Generation	workstation	48GB	$0.80/hr	$0.55/hr–$1.04/hr	13	Compare →
Nvidia A100 SXM4	datacenter	80GB	$1.23/hr	$1.23/hr–$1.23/hr	2	Compare →
Nvidia RTX PRO 6000 Blackwell Server Edition	datacenter	96GB	$1.29/hr	$1.19/hr–$1.50/hr	21	Compare →
Nvidia RTX PRO 6000 Blackwell Workstation	workstation	96GB	$1.33/hr	$1.20/hr–$1.46/hr	2	Compare →
Nvidia H100 SXM	datacenter	80GB	$1.55/hr	$1.55/hr–$1.55/hr	1	Compare →
Nvidia H100 PCIe	datacenter	80GB	$2.24/hr	$2.24/hr–$2.24/hr	2	Compare →
Nvidia H200	datacenter	141GB	$3.00/hr	$3.00/hr–$4.23/hr	5	Compare →
Nvidia B200	datacenter	192GB	$5.55/hr	$5.50/hr–$5.60/hr	2	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.

io.net

80% host

Cheapest: $0.14/hr · Nvidia RTX A4000 · 16 GPUs

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

Akash Network

Catalog-only — no live pricing yet

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.

Frequently asked

How is Lium 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 Lium 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 Lium 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 Lium?

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.