Craigslist for GPUs’ launches amid crushing Nvidia AI GPU shortages — allows you to rent AI GPU clusters by the hour

Key Points:

  • The launch of allows for renting GPU clusters by the card and by the hour
  • Listings on offer a range of GPUs, including A100 and H100, for rent with flexible terms
  • Prices on range from $2 to $3.50 per GPU per hour, making it a potentially cost-effective option compared to building a datacenter


A new online marketplace,, has emerged as a platform where individuals can rent GPU clusters by the card and by the hour. Described as the “Craigslist for GPU clusters,” the service allows users to access the power of multiple AI GPUs without the need to build a data center from scratch. Launched recently, the platform boasts listings from small to medium servers offering hundreds of GPUs for rent at affordable rates.


The site was introduced by Nat Friedman, an AI startup investor and former CEO of GitHub, along with Daniel Gross, the duo behind the Andromeda Cluster supercomputer. Operating similarly to Craigslist, features listings created by individuals with GPU clusters, detailing system specifications and rental costs per GPU per hour. Users can connect with sellers to arrange rental agreements on the platform.


Surprisingly, despite the high cost of building GPU clusters, there are already multiple listings on the platform offering A100 and H100 GPUs for rent. While some listings provide flexibility in terms of rental duration and quantity of GPUs, others require long-term commitments and large quantities. The rental prices per GPU per hour range from approximately $2 to $3.50, which might seem reasonable compared to the high prices of individual GPUs like the H100, estimated at around $40,000.


Comparing to other services like, it becomes evident that focuses on offering more powerful GPUs like the H100 and A100, providing users with access to a wider selection of data center cards for rent. The platform allows for renting varying quantities of GPUs, potentially up to 1024 H100s, surpassing the offerings of other similar services.


While the impact of on the accessibility of AI computing power remains uncertain, the platform presents an opportunity for data center owners to monetize idle GPU clusters. Whether the service will thrive depends on the balance of supply and demand within the market for rented GPU clusters. Overall, offers a unique solution for users seeking temporary access to powerful AI GPUs without the upfront costs of building their own data centers.



Prompt Engineering Guides



©2024 The Horizon