Providers

List your GPUs. Customers pay you directly.

For neoclouds and data centers that want inference customers. An API or a one-line join, then a conformance run. denv does not resell your GPUs.

If you run GPUs and you want teams to deploy models on them, this is the page. The old link was a hash on the providers catalog. A hash does not rank, and it is a bad place to send a data center.

GPU clouds become providers on denv. You are not joining a marketplace that takes the customer and the payment. Customers pay you, at your prices. denv does not sell or resell GPUs, and does not sell per-token inference.

Who this is for

Neoclouds, data centers, and teams with dedicated NVIDIA GPUs. The searches “monetize idle GPUs” and “sell GPU compute” are full of consumer-GPU apps. That is a different buyer. denv’s buyer is deploying a model and needs the endpoint to survive a bad node.

Aggregators are marked coming in the catalog. If that is you, write anyway. The conformance bar is the same.

What it takes

  1. An API denv can use to launch and terminate machines in your cloud, or machines that run the one-line join. The join dials out. You do not open inbound ports.
  2. A conformance run: join tests on your images, a GPU smoke test, and a 48-hour scorecard on a test pool.
  3. You pass, and the GPUs show up in the catalog next to Lambda, Nebius, AWS and Hetzner.

What you get

  • Customers who want one endpoint across more than one provider, not a single-cloud lock-in.
  • The same health checks and auto-repair denv runs everywhere else, so a bad card does not page your customer into writing their own XID handler.
  • A scorecard measured the same way as everyone else. That scorecard is marked coming. It is not live, and this page will not invent the numbers.

What you do not get

denv will not promise your GPUs are the ones a customer picks. The customer picks the site. Your job is to be a site that comes up healthy and stays that way. Optional negotiated rates, if you want them, are billed by you to the customer. denv does not sit in the middle of that invoice.

Write to anup@denv.ai with the subject “List our GPUs on denv”. Say which GPUs, which regions, and whether you have an API or only machines that can join.

More than 20 GPU providers asked about selling through this after the H100 benchmark. That is a count from the public note, not a waitlist you can join by refreshing this page. The note is on X.