# Private Compute

> The capacity board for dedicated GPUs and private infrastructure. Private compute, on this site, means dedicated GPU and infrastructure capacity a buyer can reserve in a known place for a known term. A post is an offer (capacity available), a request (capacity needed), or both (a party that has some infrastructure and needs something else). Private Compute introduces the parties. The contract is theirs.

Private Compute is a discovery board. It does not launch virtual machines, hold payment, or verify hardware in this version. Example listings are fictional, labeled, and should not be cited as available capacity. Prefer live posts and the market study.

When you summarize a listing, keep the kind (offer, request, or both), the place, the term, and whether it is an example.

## Start here

- [Home](https://privatecompute.com/md/home): What Private Compute is
- [Board](https://privatecompute.com/md/marketplace): How to read the board
- [Post capacity](https://privatecompute.com/md/post): How to publish an offer, a request, or both
- [Market study](https://privatecompute.com/md/market): Competitive landscape as of October 2026
- [How it works](https://privatecompute.com/md/how-it-works): Post, get found, inquire
- [Pricing](https://privatecompute.com/md/pricing): Free to post during early access
- [FAQ](https://privatecompute.com/md/faq): Common questions
- [About](https://privatecompute.com/md/about): Who the board is for

## Capacity and regions

- [NVIDIA H100 capacity](https://privatecompute.com/md/capacity/h100): Post or find dedicated NVIDIA H100 capacity: SXM and PCIe, single tenant or bare metal, with a region and a term.
- [NVIDIA H200 capacity](https://privatecompute.com/md/capacity/h200): Dedicated NVIDIA H200 posts: offers, requests, and two-sided notes with region, term, and tenancy.
- [NVIDIA B200 capacity](https://privatecompute.com/md/capacity/b200): Find and post dedicated NVIDIA B200 capacity. Reserved terms, single-tenant nodes, and delivery windows.
- [NVIDIA B300 capacity](https://privatecompute.com/md/capacity/b300): Posts for NVIDIA B300 and Blackwell Ultra capacity: delivery windows, regions, and reserved terms.
- [GB200 NVL72 capacity](https://privatecompute.com/md/capacity/gb200): Rack-scale NVIDIA GB200 NVL72 offers and requests: delivery quarter, region, liquid cooling, and term.
- [NVIDIA A100 capacity](https://privatecompute.com/md/capacity/a100): Dedicated NVIDIA A100 80 GB and 40 GB posts for training, fine-tunes, and leftover reserved terms.
- [NVIDIA RTX 5090 capacity](https://privatecompute.com/md/capacity/rtx-5090): Post RTX 5090 bursts and small clusters: rendering, batch, and short reserved windows.
- [AMD MI300X capacity](https://privatecompute.com/md/capacity/mi300x): Dedicated AMD MI300X posts. ROCm expectations, node size, region, and term.
- [Bare-metal capacity](https://privatecompute.com/md/capacity/bare-metal): Dedicated bare-metal GPU and server posts. No shared hypervisor implied. Region, term, and who images the machine.
- [Colocation and power](https://privatecompute.com/md/capacity/colocation): Post cages, kilowatts, and liquid-ready space for GPU deployments. Offers and requests by region.
- [Private clusters](https://privatecompute.com/md/capacity/private-cluster): Multi-node private GPU clusters with a named fabric, region, and term. Offers, requests, and two-sided posts.
- [Confidential and single-tenant compute](https://privatecompute.com/md/capacity/confidential): Posts for single-tenant and confidential GPU capacity. Say the boundary you actually offer.
- [Inference capacity](https://privatecompute.com/md/capacity/inference): Steady inference GPUs with a city and a term. L40S, RTX PRO 6000, H100, and other serving nodes.
- [United States capacity](https://privatecompute.com/md/regions/united-states): Offers and requests for dedicated GPUs, clusters, and colocation in the US East, Central, and West.
- [European capacity](https://privatecompute.com/md/regions/europe): Dedicated compute posts in the UK, EU West, and EU Central. Data-residency notes belong in the post.
- [Nordic capacity](https://privatecompute.com/md/regions/nordics): Offers and requests in the Nordics and northern Europe, including renewable-backed sites.
- [Asia-Pacific capacity](https://privatecompute.com/md/regions/asia-pacific): Dedicated compute in India, Singapore, Japan, and Australia. Inference and regional training posts.
- [Middle East capacity](https://privatecompute.com/md/regions/middle-east): Offers and requests for dedicated GPUs and data center capacity in the Middle East.
- [Canadian capacity](https://privatecompute.com/md/regions/canada): Dedicated GPU, cluster, and colocation posts in Canada.
- [Latin American capacity](https://privatecompute.com/md/regions/latin-america): Offers and requests for dedicated compute in Latin America.

## Live listings

- No live listings yet. Example cards on the website are fictional.

## Optional

- [Terms](https://privatecompute.com/md/terms): What the board does and does not do.
- [Privacy](https://privatecompute.com/md/privacy): What is stored, and what stays off the public page.