Best VPS for AI agents.
Prices checked
For AI agents, Leaseweb VPS 2 · 6 vCPU / 16 GB · Europe at $9.04 a month is the plan to start with, of 300 that clear the minimums, checked .
Top 10
all 527 plans- 1
VPS 2 · 6 vCPU / 16 GB · Europe
Leaseweb · 6C · 16 GB · 200 GB NVMe$9.04/mo - 2
VPS 1 · 4 vCPU / 6 GB · Europe
Leaseweb · 4C · 6 GB · 100 GB NVMe$5.02/mo - 3
VPS 2 · 6 vCPU / 16 GB · UK
Leaseweb · 6C · 16 GB · 200 GB NVMe$9.63/mo - 4
VPS 2 · 6 vCPU / 16 GB · US
Leaseweb · 6C · 16 GB · 200 GB NVMe$9.89/mo - 5
VPS 3 · 8 vCPU / 24 GB · Europe
Leaseweb · 8C · 24 GB · 300 GB NVMe$14.07/mo - 6
VPS 4 · 12 vCPU / 48 GB · Europe
Leaseweb · 12C · 48 GB · 400 GB NVMe$26.13/mo - 7
VPS 5 · 16 vCPU / 64 GB · Europe
Leaseweb · 16C · 64 GB · 500 GB NVMe$34.18/mo - 8
VPS 2 · 6 vCPU / 16 GB · Singapore
Leaseweb · 6C · 16 GB · 200 GB NVMe$11.18/mo - 9
VPS 1 · 4 vCPU / 6 GB · UK
Leaseweb · 4C · 6 GB · 100 GB NVMe$6.02/mo - 10
VPS 1 · 4 vCPU / 6 GB · US
Leaseweb · 4C · 6 GB · 100 GB NVMe$6.29/mo
What matters
- RAM first
- A single headless Chromium for a browser-using agent takes 0.5 to 1 GB. Add a vector store, a queue and the runtime, and 4 GB is the realistic floor, with 8 GB comfortable.
- Dedicated vCPU helps
- Agent workloads are bursty (tool calls, parsing, embeddings). On shared cores a noisy neighbor turns tail latencies into timeouts mid-run.
- NVMe for local stores
- SQLite, Chroma or pgvector on fast disk keeps retrieval snappy. Inference happens at your API provider, so you don't need a GPU.
- Bandwidth rarely the bottleneck
- Token streams are small. Unless agents move large files, any plan's allowance is fine.
Or run it on a platform
Cheapest place to run an n8n automationA VPS hands you the whole machine and its updates; an app platform runs the app for you and meters the rest. The same job on a platform is Cheapest place to run an n8n automation, which asks for 1 GB of memory or more, on a runtime the provider states stays on. 35 plans clear that, and the cheapest sized runtime is Koyeb eSmall · 0.5 vCPU · 1 GB at $5.36 a month.
One-click stacks
Full stack-support matrixEverything here runs on any VPS you set up yourself. These providers also offer it as a one-click install, verified 20 Jul.
- Docker
- One-click on Hetzner, Vultr, DigitalOcean, Linode, OVHcloud, Scaleway, contabo, Hostinger, IONOS
- n8n
- One-click on DigitalOcean, OVHcloud, contabo, Hostinger, IONOS
- PostgreSQL
- One-click on DigitalOcean, Linode, Hostinger
- Node.js
- One-click on Vultr, DigitalOcean, Linode, AWS Lightsail, Hostinger
How we count
Agent frameworks are RAM-hungry and bursty: headless browsers, tool sandboxes, vector stores and queues all run alongside the orchestrator, while the LLM calls themselves go out to APIs. We only include plans with 4 GB+ RAM, 2+ vCPU and a disk included in the price. Today 300 of the 527 plans qualify, and the top pick costs $9.04/mo. They're ranked by how much of the hardware this workload relies on you get for each dollar.
Every price is the standard monthly rate for a plan with a disk included, before VAT. A plan whose provider was last read more than three days ago, or that the provider shows as sold out, keeps its place but cannot be first.
Frequently asked
- What is the cheapest VPS that qualifies for AI agents?
- Leaseweb VPS 1 · 4 vCPU / 6 GB · Europe at $5.02/mo (excl. VAT) meets the minimum requirements as of 31 Oct 2026.
- How are these requirements chosen?
- The minimums (4 GB RAM, 2 vCPU) reflect what the workload actually stresses, as explained under "what matters", and the ranking weights those dimensions rather than raw price.
- How fresh is this list?
- Prices are read again from all 25 providers every 3 days and this ranking recomputes automatically. Last refresh: 31 Oct 2026.