# A private GPU cluster for your research

Your institution owns the servers and your data stays on them. We find a data center with the power campus can’t provide, and run the cluster for you.

## Why labs own their GPUs

- **Your data stays on your hardware.** Dedicated servers that only your team uses.
- **No queue.** No waiting for a shared cluster or cloud capacity.
- **Equipment, not a bill.** A one-time purchase that can fit a grant budget. The institution keeps the hardware.
- **Power campus doesn’t have.** An 8-GPU server draws about 15 kW. We place it where the power is.

## Built around your procurement

Purchase orders and contracts are in your institution’s name. We have no preferred vendors, so you get the best price on the market right now. We handle NVIDIA’s end-user paperwork.

Tell us your data requirements, such as HIPAA or keeping data in one state, and we’ll pick a site to match.

## Ballpark costs

- **B300 server (8 GPUs):** $550-650K
- **B200 server (8 GPUs):** $450-500K
- **H200 server (8 GPUs):** $340K
- **Networking for a cluster:** $50-90K per server
- **Data center:** $4-5K a month per B300 server
- **Our setup fee:** A flat $25K for 1-7 servers, 8 or more quoted. Not a percentage of the project.
- **Our management fee:** From $1K a month per 4 servers

## We manage. You own.

1. **Scope it.** Tell us your budget, timeline, and power. You get a plan and ballpark costs. One call.
2. **We procure it.** No preferred vendors. We source servers, networking, and data center space at today’s best price, in your name. 60-90 days typically.
3. **We run it.** Operations and maintenance (O&M), software, and customers, on an annual contract. Ongoing.

## Tell us about your lab.

- A call with our CTO about what you need
- Ballpark costs and timelines for your budget
- A straight answer if it’s not a fit

Prefer email? [hello@amcompute.com](mailto:hello@amcompute.com)

## Questions from research teams

**Whose name is the hardware in?**

Your institution’s. You can pay every vendor directly.

**Can you work with our procurement office?**

Yes. We provide specs, pricing, and vendor details for your purchasing team.

**Our building can’t power these servers.**

That’s common. We place the cluster in a U.S. colocation site, as close to you as capacity allows.

**Can several groups share the cluster?**

Yes. A scheduler such as Slurm lets groups share the GPUs. We set it up with your research computing team.

**Is owning cheaper than cloud credits?**

Over three years, owning usually costs less if the GPUs stay busy, above about 77% of the time in our own-vs-rent analysis. Cloud credits suit short or occasional projects.

**Can we use grant funds?**

Many grants allow equipment purchases. Check your award terms, and we’ll provide pricing for your budget.

## Further reading

- [Run AI on your own servers](https://www.amcompute.com/managedgpus/guides/run-ai-on-your-own-servers)
- [NVIDIA GPU differences, from Ampere to Blackwell](https://www.amcompute.com/blog/nvidia-gpu-generations)
- [Data center tiers explained](https://www.amcompute.com/blog/data-center-tiers)
- [Where to buy GPU servers](https://www.amcompute.com/blog/where-to-buy-gpu-servers)

## Hardware and hosting

- [NVIDIA B300 servers](https://www.amcompute.com/managedgpus/b300) - Ballpark pricing, power, and lead times for HGX B300 servers.
- [NVIDIA B200 servers](https://www.amcompute.com/managedgpus/b200) - Ballpark pricing, power, and availability for HGX B200 servers.
- [NVIDIA H200 servers](https://www.amcompute.com/managedgpus/h200) - Ballpark pricing and availability for HGX H200 servers.
- [NVIDIA RTX PRO 6000 servers](https://www.amcompute.com/managedgpus/rtx-pro-6000) - Ballpark pricing and power for RTX PRO 6000 servers, the lower-cost way in.
- [NVIDIA GB300 NVL72 racks](https://www.amcompute.com/managedgpus/gb300-nvl72) - What it takes to own a liquid-cooled GB300 NVL72 rack.
- [GPU colocation](https://www.amcompute.com/managedgpus/gpu-colocation) - A U.S. data center for the GPU servers you own.
- [GPU colocation directory](https://www.amcompute.com/managedgpus/colocation) - High-density and liquid-cooled data centers by metro: power, kW per rack and liquid cooling, each with a source.
- [GPU server vendors](https://www.amcompute.com/managedgpus/gpu-server-vendors) - Who sells GPU servers worldwide, from the public partner lists of NVIDIA, Supermicro, Dell, HPE, Lenovo, Cisco, Gigabyte, ASUS and AMD, with server models by maker.

## Want your own cluster? We’ll run it.

[Book a call](https://www.amcompute.com/managedgpus/research#discuss)

Source: https://www.amcompute.com/managedgpus/research
