AI VPS

Run AI agents, automations and small models on a server you control.

A KVM VPS with NVMe and root for CPU based AI workloads and tooling.

AI VPS from

$6per month

vCPU
1
Memory
2 GB
NVMe
40 GB
Deploy AI VPS

Same plans and prices in every location.

What you get

A AI VPS is a full KVM server, set up and documented for this job.

What runs well on a CPU VPS

These plans are CPU and memory servers, not GPU machines, so they suit the glue around AI rather than training large models. That covers a lot: agent frameworks, automation that calls hosted model APIs, retrieval pipelines, vector databases, and small quantized models that run acceptably on CPU.

If you need heavy GPU compute for training, pair a VPS for the orchestration with a separate GPU provider for the model, and keep your keys and data on the server you control.

AI VPS plans

Start small and widen your lane when you need more room.

Lane 2

$12 per month

Each lit lane is one vCPU

vCPU
2
Memory
4 GB
NVMe storage
80 GB
Monthly traffic
4 TB
Deploy Lane 2

Every plan includes a dedicated IPv4 address, IPv6, full root or administrator access, snapshots and your choice of Linux or Windows.

Ready to deploy?

Pick a location, size your AI VPS and connect in minutes.

Deploy now

AI VPS questions

Do these plans have a GPU?

No. They are CPU and memory servers, ideal for agents, automations and small CPU models rather than large scale training.

Can I run an AI agent around the clock?

Yes. A VPS keeps agents and automations running all day, with your keys stored on the server.

Can I run a local model?

Small quantized models run on CPU. For large models, use a VPS for orchestration alongside a GPU provider.