Your GPUs. Our buyers.
- Rented
- 0
- Available
- 0
- Utilization
- 0%
Idle GPUs are stranded revenue.
- An unsold GPU-hour never comes back.
- HZ brings buyers to the hours you list.
- You set the price.
Sell GPU-hours, not listings.
- You choose the nodes and the price.
- HZ brings the workloads.
- You're paid for what runs.
HZ brings buyer workloads to your capacity. No listing to market.
Every job is covered by a prepaid hold, then settled to runtime.
Set a price per GPU-hour for each node you put on HZ.
Each buyer gets a dedicated GPU, kept apart from your host.
Every site, node and GPU in one dashboard.
Matching, placement, metering and billing.
From application to payout.
- Five steps.
- No listing to manage once you're live.
- 01Apply
Tell us about your fleet.
- 02Review
HZ checks GPUs, site and network fit.
- 03Connect
Add nodes, set prices, run the agent.
- 04Go live
Nodes pass health checks and take work.
- 05Get paid
Usage is metered, settled and paid out.
Isolated workloads, near bare-metal performance.
- Signed agent, enrolled over mTLS.
- One dedicated GPU per buyer.
- Private networks and cloud metadata blocked.
One control plane. Every site.
- Every connected facility joins one pool of GPU capacity.
- Buyers filter capacity by region.
- Each workload lands on a healthy GPU that fits.
Every site. One dashboard.
- Health, GPUs and workload for every site.
- Bulk enroll a rack or a whole hall.
- Marketplace or enterprise tier, per node.
| Site | Region | Nodes | Rented | Health | Earned · 7d |
|---|---|---|---|---|---|
| DC1 · Oregon | Americas | 8 | 46 / 64 | Healthy | $14,200 |
| DC2 · Ashburn | Americas | 12 | 81 / 96 | Healthy | $24,900 |
| DC3 · Dallas | Americas | 8 | 38 / 64 | 1 draining | $11,300 |
| DC4 · Frankfurt | Europe | 6 | 35 / 48 | Healthy | $10,800 |
| DC5 · Stockholm | Europe | 4 | 17 / 32 | Healthy | $5,100 |
| DC6 · Singapore | Asia-Pacific | 8 | 52 / 64 | Healthy | $16,400 |
| DC7 · Tokyo | Asia-Pacific | 4 | 26 / 32 | Healthy | $8,000 |
Apply to supply.
- A few questions about you and your fleet.
- Our team replies by email.
Questions operators ask.
How are buyer workloads kept off my hosts?
Every workload runs in its own isolated environment on a dedicated GPU, with private networks and cloud metadata blocked.
How does billing work?
Before a job starts, HZ places a hold on the buyer's prepaid credit. When the job ends, it settles to the actual runtime at your GPU-hour rate, with a 60-second minimum.
How do I connect my GPUs?
Add a node in your console and set its price. Each node needs Ubuntu, NVIDIA GPUs and a routable public IPv4 address. HZ issues an enrollment token, and you run the agent installer. The node starts taking work once it's online and passes health checks.