Private AI on servers you own
Your models and data stay on your hardware. We host the servers and run them.
Why companies own their AI servers
Your data stays put
Servers only your company uses.
A fixed bill
The same every month, however many tokens you use.
Any open model
Qwen, DeepSeek, GLM or Kimi.
Power your office lacks
10-16 kW per server. We place it where the power is.
Built around your procurement
Orders and contracts are in your company’s name. We have no preferred vendors, so you get the best price on the market right now. We handle NVIDIA’s paperwork.
Need HIPAA, or data kept in one state? We pick a site to match. See which models fit which server.
Ballpark costs
- H200 server (8 GPUs)
- $340K
- B200 server (8 GPUs)
- $450-500K
- B300 server (8 GPUs)
- $550-650K
- Data center
- $240-325 per kW a month
- 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 guide. You decide.
Scope it
Tell us your workload, budget, and timeline. You get a plan and ballpark costs.
One call
We build it
No preferred vendors. We find the best price on the market right now. You pick, and we handle the rest in your name.
60-90 days typically
We run it
Monitoring, repairs, and warranty claims, on an annual contract.
Ongoing
Tell us about your AI plans.
- 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
Questions from companies
- Whose name is the hardware in?
- Your company’s. You can pay every vendor directly.
- Which models can we run?
- Any open-weight model, such as Qwen, DeepSeek, GLM or Kimi. A 27B model needs one GPU, and one 8-GPU B300 server runs the largest, like Kimi K3.
- Do we need our own AI team?
- Someone picks the models and connects them to your apps, often your IT team or a software vendor. We keep the servers running.
- Can it connect to our network?
- Yes, over a VPN or a private connection from the data center to your offices or cloud.
- Is it cheaper than an API?
- With steady, heavy use, yes. For a few requests an hour an API costs less. The cost is fixed either way once you own the servers.
Further reading
Hardware and hosting
- NVIDIA B300 serversBallpark pricing, power, and lead times for HGX B300 servers.
- NVIDIA B200 serversBallpark pricing, power, and availability for HGX B200 servers.
- NVIDIA H200 serversBallpark pricing and availability for HGX H200 servers.
- NVIDIA RTX PRO 6000 serversBallpark pricing and power for RTX PRO 6000 servers, the lower-cost way in.
- NVIDIA GB300 NVL72 racksWhat it takes to own a liquid-cooled GB300 NVL72 rack.
- GPU colocationA U.S. data center for the GPU servers you own.
- GPU colocation directoryHigh-density and liquid-cooled data centers by metro: power, kW per rack and liquid cooling, each with a source.
- GPU server vendorsWho 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.
