Written by the CloudHostAI team
Hetzner vs the hyperscalers: what BYOH really costs
CloudHostAI is a BYOH (Bring Your Own Hetzner) managed PaaS: you connect your own Hetzner Cloud account and we handle provisioning, deploys, and TLS. The pitch is Railway-style developer experience at Hetzner prices — here is the actual price comparison for a small workload, using 2025-2026 list pricing.
The workload
A small containerized app: roughly 2 vCPU, 2 GB RAM, 40 GB SSD, with modest outbound transfer. This is the shape CloudHostAI's free-tier onboarding targets (one CPX11-class server).
Three price tags for the same box
| Provider | Monthly | Spec |
|---|---|---|
| Hetzner CPX11 | €3.99/mo | 2 vCPU · 2 GB · 40 GB SSD · 20 TB out |
| AWS t3.small + 40 GB EBS | ~$22/mo | + data transfer ~$10 and up |
| Railway | ~$10/mo | + usage on top |
Estimates based on 2025-2026 list pricing. Your actual bill depends on usage — this is a starting-point comparison, not a guarantee.
Why the gap
Hetzner's list prices for compute are structurally lower than the big three hyperscalers for the same vCPU/RAM/disk shape, and Hetzner bundles a large outbound-transfer allowance (20 TB on a CPX11) rather than metering it separately the way AWS does. Railway sits in between: cheaper than raw AWS, but it marks up the underlying compute to cover its own managed layer — the same layer CloudHostAI provides directly on top of your own Hetzner account instead.
What BYOH means for the bill
With CloudHostAI you pay Hetzner directly for compute (starting at €3.99/mo) and never lose visibility into that invoice — there's no reseller markup on the server itself. CloudHostAI's own free plan covers one project; paid plans and optional add-ons (like the Lumen API cost-router at $19/mo) are billed separately and only cover the orchestration layer, not the underlying compute.