News
DealsCorvex agrees USD 103 million take-or-pay deal to perform AI inference with Mid-Atlantic GPUsSovereignNeysa scales AI infrastructure in India with AMD Instinct MI350X GPUs on TytanInferenceMicrosoft Foundry adds Baseten, DeepSeek AI inference (MSFT:NASDAQ)ClustersDataOne and Rowan University sign engineering co-op program agreement for DataOne's Vineland AI factoryClustersLook inside Cal Poly's $3 million AI Factory (Video)DealsOpenAI Taps AMD Chips for 457 MW of AI Capacity in Six-Year TensorWave DealClustersCal Poly launches an AI data center on campus. Take a look insideInferenceArista takes aim at the next AI network battleground: distributed inferenceDealsOpenAI Signs 457 MW AI Data Center Deal as AMD Expands Its Challenge to NvidiaClustersBehind Dell AI Factory’s NVIDIA & AMD Enterprise ExpansionDealsCorvex agrees USD 103 million take-or-pay deal to perform AI inference with Mid-Atlantic GPUsSovereignNeysa scales AI infrastructure in India with AMD Instinct MI350X GPUs on TytanInferenceMicrosoft Foundry adds Baseten, DeepSeek AI inference (MSFT:NASDAQ)ClustersDataOne and Rowan University sign engineering co-op program agreement for DataOne's Vineland AI factoryClustersLook inside Cal Poly's $3 million AI Factory (Video)DealsOpenAI Taps AMD Chips for 457 MW of AI Capacity in Six-Year TensorWave DealClustersCal Poly launches an AI data center on campus. Take a look insideInferenceArista takes aim at the next AI network battleground: distributed inferenceDealsOpenAI Signs 457 MW AI Data Center Deal as AMD Expands Its Challenge to NvidiaClustersBehind Dell AI Factory’s NVIDIA & AMD Enterprise Expansion
EveryNeocloud
Lium

Lium

GPU rental marketplace with per-second billing

Lium is a GPU compute marketplace operated by Datura AI Corp. It connects independent hardware providers with renters who need on-demand graphics processing power for artificial intelligence tasks. The platform offers per-second billing and supports various Nvidia chips.

Share on WhatsApp

  • Providers keep 95% of rental fees
  • Billing occurs per second with no minimum
  • Supports Nvidia H100, H200 and Blackwell chips
  • Operated by Datura AI Corp in Saint Kitts and Nevis
  • Offers CLI and SDK for AI agent integration

What it is

Lium operates as a marketplace for renting graphics processing units. Independent providers list their hardware, and users rent these machines as containers called pods. The service charges by the second with no minimum time limit. Prices are set by the hardware owners, who keep ninety-five percent of the rental fee. The platform is run by Datura AI Corp, which is based in Saint Kitts and Nevis. Users can access the service through a web interface, command line tools or application programming interfaces.

The compute

The marketplace supports a range of Nvidia accelerators. Available models include the H100, H200 and B200 series, along with consumer cards like the RTX 4090 and professional Blackwell editions. Providers offer these chips in individual pods that run on Linux systems. The platform does not guarantee high-speed interconnects such as InfiniBand for all nodes. Users can choose between secure tiers and spot instances, which are cheaper but may be reclaimed by the provider. Each pod includes persistent storage that survives reboots.

Where the GPUs are

Compute resources are distributed across a global network of independent providers. Listings show machines located in places such as Singapore and the United States. The platform does not own data centres but relies on third-party hardware owners to host the GPUs. Power and physical infrastructure are managed by these individual providers rather than a central entity. The marketplace enforces export controls and blocks access from sanctioned regions.

Customers and deals

The service targets developers and artificial intelligence agents that need flexible compute power. Tools are provided for coding assistants like Claude Code to manage GPU jobs automatically. Renters prepay for usage via credit card or cryptocurrency. The platform provides invoices for business users who save their tax details. Support is handled through email and a community Discord server.

Money and ownership

Datura AI Corp operates the platform and collects a five percent fee from each transaction. Providers receive their share of the revenue daily. Idle pay is available for hardware owners when their GPUs are not rented. The company processes payments through Stripe and supports various cryptocurrencies for top-ups. There is no public information on external funding or valuation.

Its place in AI

Lium positions itself as a tool for AI agents to acquire compute resources programmatically. It allows automated systems to start and stop jobs within a set budget. The open nature of the marketplace means prices and availability change in real time. This model contrasts with fixed-price cloud services by letting the market determine the cost of GPU time.

Questions and criticism

Users must maintain a prepaid balance to cover potential costs. Spot instances carry the risk of sudden termination if the provider needs the hardware back. The reliance on independent operators means reliability scores vary between nodes. Some machines may reboot occasionally, which could interrupt long-running tasks. Export control policies restrict access for users in certain countries.

Questions people ask about Lium

How does billing work on Lium?

Lium charges for GPU usage by the second. There is no minimum rental period. Users must preload a balance that covers at least fifteen minutes of runtime for each active pod. Billing stops immediately when the pod is deleted or reaches its scheduled end time.

Who owns the hardware on the platform?

Independent providers own and list their own machines. Lium acts as a marketplace connecting these owners with renters. The providers set their own hourly prices and are responsible for maintaining the physical hardware and internet connectivity.

What types of GPUs are available?

The marketplace lists various Nvidia cards. These include data centre chips like the H100 and H200, as well as Blackwell series processors. Consumer-grade options such as the RTX 4090 are also available. Specific models depend on what independent providers have listed at any given time.

Good reading and sources

    Where this page comes from

    There is no English Wikipedia article on Lium yet, so the facts on this page come from the sources listed above: lium.io/about, lium.io/. Some are the subject's own pages, so its claims about itself are reported as its claims. EveryNeocloud wrote the text in its own words, with AI assistance, and compares it with the sources to make sure no passage is copied.

    It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.

    The weekly digest

    Every Monday: the provider of the day, new pages on EveryNeocloud, and the conferences and anniversaries of the week ahead.