Hetzner cloud server price increases in 2026: full breakdown and alternatives

• Hetzner raised prices across cloud servers and dedicated servers effective 15 June 2026, applying to new orders and rescales; this article focuses on cloud server increases across Germany/Finland, USA, and Singapore regions
• CCX (dedicated vCPU) instances increased 2.1x to 2.73x monthly in Germany/Finland and 2.1x to 2.6x in the USA; CPX (shared AMD vCPU) instances increased 2.4x to 2.75x in Germany/Finland and up to 3.1x in the USA; Singapore saw smaller increases across both families. CX (shared Intel/AMD) and CAX (Arm) instances saw smaller increases of roughly 1.3x to 1.4x monthly
• Teams running workloads directly on a single cloud provider have limited options when that provider raises prices: absorb the cost, or migrate. Deployment platforms with self-serve BYOC (Bring Your Own Cloud) support reduce that migration cost by keeping workload configuration in a platform layer rather than in provider-specific scripts
• Northflank provides a full-stack deployment platform with self-serve BYOC support across AWS, GCP, Azure, OCI, CoreWeave, Civo, on-premises, and bare-metal, giving teams the option to run workloads across multiple providers from a single control plane

Hetzner raised cloud server and dedicated server prices on 15 June 2026, with some cloud server instance families increasing by more than 2.5x monthly. The increases apply to new orders and cloud instance rescales; orders placed before 15 June 2026 retain the previous prices.

This article covers which instance families were affected, how much prices increased across regions, and what teams running workloads on Hetzner cloud should consider.

Hetzner is a German infrastructure provider offering cloud servers, dedicated servers, and storage. Its cloud server product competes with VPS and managed compute offerings from providers like AWS, GCP, and Azure, and has historically been priced lower than those providers for raw compute.

The price adjustment took effect on 15 June 2026 at 8 AM CEST. It applies to new orders and cloud instance rescales from that date. Orders placed before 15 June 2026 but delivered after retain the previous prices.

Hetzner's cloud server SKUs fall into four families:

The price adjustment affects all four families, but the scale of the increase differs significantly between them. All prices below exclude VAT and exclude IPv4 costs.

CCX instances provide dedicated AMD vCPUs, meaning the physical processor cores are allocated solely to your instance. These are the most production-critical tier in Hetzner's cloud lineup.

The table below shows old and new monthly prices for the Germany (FSN/NBG) and Finland (HEL) region in euros, excluding IPv4.

The entry-level CCX13 saw the steepest relative increase at 2.69x. The largest CCX63 instance went from €374.49 to €853.49 per month.

CPX instances use shared AMD vCPUs. The table below covers the Germany/Finland region.

CX instances use shared Intel or AMD vCPUs. CAX instances use shared Arm (Ampere) vCPUs. Both families saw smaller increases than CCX and CPX in Germany/Finland.

The price adjustment covers three regions: Germany/Finland (FSN/NBG/HEL), USA (ASH/HIL), and Singapore (SIN). The table below shows CCX13 and CPX monthly prices across all three regions in euros, old versus new. CPX32 is not available in the USA region; CPX31 is shown for USA.

Germany/Finland and USA saw the largest relative increases on CCX and CPX instances. Singapore saw smaller relative increases, partly because Singapore pricing was already higher in absolute terms before the adjustment. USA CPX instances saw some of the steepest increases across all regions: CPX31 in the USA increased 2.98x monthly, compared to CPX32 in Germany/Finland at 2.54x and Singapore at 1.51x.

For teams on new orders or rescaling existing instances, the updated prices apply from 15 June 2026. The practical cost impact scales with how many instances a team runs and which families they use.

As a worked example: a team running three CCX33 instances in Germany/Finland was paying 3 x €62.49 = €187.47 per month. At new prices, the same three instances cost 3 x €138.49 = €415.47 per month, an increase of €228 per month or approximately €2,736 per year (€228 x 12), for the same compute.

Teams on CX or CAX instances see a smaller impact. Three CX33 instances in Germany/Finland move from €19.47 to €25.47 per month, an increase of €6 per month.

The impact is most significant for teams running CCX or CPX instances, as those are the families where the increases are largest.

Teams affected by the Hetzner price increases have several options, including:

1. Use a deployment platform with BYOC (bring your own cloud) support: where your deployment configuration, pipelines, and workflows live in the platform, not in provider-specific scripts. Moving infrastructure means pointing the platform at a different cloud account and redeploying, without rewriting your stack. This also reduces exposure to the same situation if a future provider raises prices. Not all BYOC platforms are self-serve; some require a sales process before you can connect a cloud account.
2. Migrate to another raw cloud provider: this involves reprovisioning instances, reconfiguring networking, storage, and security rules, and redeploying on the new provider. The engineering time required depends on how much of the original infrastructure is scripted versus manually configured. For teams managing infrastructure directly, this migration cost falls entirely on internal engineering resources.
3. Absorb the cost: if the new pricing fits within budget, no migration work is required. For teams on CX or CAX instances where increases are smaller, this may be the most practical short-term option. For teams on CCX or CPX instances, the annual cost increase may make this harder to sustain.

Northflank is a full-stack deployment platform with BYOC as one of its offerings. It covers the full lifecycle from code commit to production: deployments, built-in CI/CD pipelines, managed databases and queues, preview environments, secure sandboxes, GPU workloads, internal developer platform (IDP) capabilities, and Kubernetes orchestration, all from a single control plane.

Northflank's BYOC is self-serve. You can connect your own AWS, GCP, Azure, OCI, CoreWeave, Civo, on-premises, or bare-metal infrastructure to Northflank without going through a sales process.

Northflank provisions and manages Kubernetes clusters within your cloud environment, handling upgrades, scaling, and maintenance. Your templates, pipelines, CLI, and API remain consistent across managed and BYOC environments. Move workloads between managed and BYOC environments without changing deployment workflows. All workloads and data remain within your cloud boundary while the Northflank control plane manages deployments, scaling, and operations.

Northflank supports EKS, GKE, AKS, and other compatible Kubernetes distributions via Bring Your Own Kubernetes (BYOK). The Pay-as-you-go tier covers 600 BYOC regions with no added cost for running in your VPC.

See how teams use Northflank's BYOC in production: Upwork Lifted runs production workloads inside their own VPC with Northflank and Ultralight migrated from AWS ECS to EKS with Northflank.

For SaaS teams whose enterprise customers require software to run inside their own cloud account, Northflank provides customer VPC deployments. You define your application once and Northflank deploys it into each customer's AWS, GCP, Azure, Oracle, Civo, CoreWeave, or on-premises environment, with CI/CD, monitoring, logging, security scanning, and disaster recovery handled across all customer environments from a single control plane.

Northflank runs CPU and GPU workloads on the same platform. GPU support covers NVIDIA L4 (24 GB, $0.80/hr), A100 (40 GB, $1.42/hr), A100 (80 GB, $1.76/hr), H100 (80 GB, $2.74/hr), and RTX PRO 6000 (96 GB, $3.00/hr) on Northflank's cloud, or on your own GPU infrastructure via BYOC across AWS, GCP, Azure, Oracle, and CoreWeave.

Northflank compute is billed per second. Pay-as-you-go rates are CPU: $0.01667/vCPU/hour and Memory: $0.00833/GB/hour. The smallest compute plan, nf-compute-10, provides 0.1 shared vCPU and 256 MB of memory at $2.70/month ($0.0038/hr).

• Sandbox free tier: always-on compute, two free services, one free database, two free cron jobs
• Pay-as-you-go: self-service, no salesperson needed. Billed per second, no seat-based pricing, teams included for free, 6+ managed cloud regions, 600 BYOC regions, CPU and GPU support
• Enterprise: custom SLAs, white-label, SSO/SAML/OIDC, audit logs, global backups and HA/DR, managed control plane in your VPC, on-prem deployments, and bring your own registry, Vault, DNS, and more

See full pricing and the pricing calculator to estimate your monthly spend.

Hetzner is a German infrastructure provider offering cloud servers, dedicated servers, and storage. Its cloud pricing has historically been lower than major providers like AWS, GCP, and Azure, making it a common choice for cost-sensitive engineering teams. The June 2026 increase raised prices on CCX cloud server families by 2.1x to 2.73x monthly in Germany/Finland, and CPX families by 2.4x to 2.75x monthly in Germany/Finland, with CPX instances in the USA increasing up to 3.1x monthly.

The CCX (dedicated vCPU) and CPX (shared AMD vCPU) families saw the largest increases. In Germany/Finland, CCX instances increased 2.1x to 2.73x monthly and CPX instances increased 2.44x to 2.75x monthly. In the USA, CPX instances increased up to 3.1x monthly. CX and CAX instances saw smaller increases of roughly 1.3x to 1.4x monthly.

Orders placed before 15 June 2026 retain the previous prices. The new prices apply to new orders and cloud instance rescales from 15 June 2026 onwards.

The price adjustment covers Germany (FSN/NBG), Finland (HEL), USA (ASH/HIL), and Singapore (SIN). Germany/Finland and USA saw the largest relative increases on CCX and CPX instances.

BYOC (bring your own cloud) is a deployment model where you connect your own cloud account to a platform, and the platform manages workload deployment on that infrastructure. Your deployment configuration, pipelines, and workflows live in the platform rather than in provider-specific scripts, so moving to a different cloud provider involves provisioning capacity in a new account and redeploying, rather than rewriting infrastructure configuration.

Teams can move to other raw cloud providers and manage infrastructure directly, or use a deployment platform with self-serve BYOC support to deploy across multiple providers from a single control plane. Northflank supports self-serve BYOC across AWS, GCP, Azure, OCI, CoreWeave, Civo, on-premises, and bare-metal. See Best options for BYOC in cloud computing for a comparison of platforms.

• Best BYOC sandbox platforms: a comparison of platforms for running isolated workloads in your own cloud account
• Best PaaS that runs in your own cloud account: how to evaluate PaaS platforms that deploy into your own cloud rather than a shared managed environment
• Best cloud hosting platforms: a comparison of cloud hosting platforms across pricing, deployment models, and infrastructure support
• Heroku pricing comparison and reduction: a side-by-side comparison of Heroku dyno pricing against Northflank compute plans, for teams evaluating Heroku as an alternative to their current cloud setup