Hetzner vs Linode: Price-performance vs mature managed services

Raw capacity or a platform that meets you halfway?

Hetzner and Linode both sell Linux virtual machines to technical buyers, and both have been doing it for longer than most of the industry. That surface similarity hides a genuine philosophical split. Hetzner has stayed close to the metal, refining a short catalogue of servers, volumes, networks and firewalls, and passing the savings from owning its own facilities back to customers. Linode, now operating inside Akamai and marketed as Akamai Cloud Computing, has spent the same years adding a managed layer: a Kubernetes engine, managed database clusters, load balancers and a monitoring service.

The result is that a team choosing between them is really choosing how much infrastructure work it wants to keep. On Hetzner the answer is nearly all of it, in exchange for compute that goes further. On Linode a meaningful slice can be handed over, in exchange for a platform that costs more per unit and gives you less headroom on a single machine.

This page works through that trade in the places it actually bites: provisioning, storage, networking, support and long-term operations. Both providers are available through AllCloudShop, on the Hetzner account listing and the Linode account listing.

Where the catalogues overlap and diverge

Read down the Kubernetes and database rows and the essential difference is already visible. Everything else is detail around that.

The first week on each platform

A developer starting on Hetzner Cloud creates a project, generates an API token scoped to it, and from that point can drive the whole environment from hcloud or Terraform. Servers accept cloud-init user data at creation, so a fresh machine can arrive with a hardened SSH configuration, an unprivileged user and a container runtime already installed. Firewalls are their own objects that attach to servers by label, which makes rule management pleasantly declarative.

A developer starting on Linode meets StackScripts instead: shell scripts with user-supplied variables that execute on first boot, shareable across accounts and used by the Marketplace catalogue itself. The linode-cli tool covers every resource type, the Terraform provider is well maintained, and the Cloud Manager remains one of the clearer control panels in this tier. Reference material has been folded into Akamai's technical documentation, available through the Linode API reference.

The divergence appears in week two, when the application needs a database. On Linode you provision a managed MySQL or PostgreSQL cluster, copy the connection string and move on. On Hetzner you launch another server, install the engine, configure replication if you need it, set up backups and decide how to monitor the whole thing. Neither answer is wrong, but only one of them is free of ongoing work. The managed versus self-hosted comparison is a sober look at what that work involves.

Both platforms expose their full functionality through an API, so infrastructure-as-code teams will find familiar ground on either side. If you have not built that habit yet, the introduction to infrastructure as code is a good place to start, because both providers reward it.

The differences with real consequences

1. The managed control plane. LKE removes control plane operation entirely, including upgrades and etcd care, and integrates with NodeBalancers and Block Storage so that Kubernetes service and volume objects work as expected. Hetzner has excellent building blocks for a cluster but no managed control plane, which means somebody on your team becomes the cluster administrator.
2. Headroom per euro or dollar. Hetzner's dedicated vCPU plans and its Arm64 family let a fixed budget buy noticeably more compute and memory than an equivalent tier elsewhere. That matters most for workloads that are simply large: build farms, video pipelines, analytics jobs, self-hosted development platforms.
3. Reach. Akamai ownership has pushed Linode's region list well beyond what it had in 2021, including places where developer clouds rarely appear. Hetzner has expanded too, but its map is still recognisably European with a few outposts.
4. Support expectations. Linode includes phone and ticket support at any hour for every account, a legacy of its hosting roots. Hetzner answers tickets competently but does not present itself as a round-the-clock concierge, and the difference shows at three in the morning.
5. Corporate context. Hetzner is independent and moves at its own pace. Linode sits inside a listed parent with an enterprise sales organisation, edge security products and a compliance programme that larger buyers often ask about.

Reading both as a purchase decision

The finance conversation is straightforward. Both providers bill by consumption with a monthly ceiling per resource, so neither produces the per-request surprises a hyperscaler invoice can contain. The variable that matters is unit cost against operational cost: Hetzner lowers the first and raises the second, Linode does the reverse. A team with a capable platform engineer usually comes out ahead on Hetzner; a team of four product developers with no operations specialist usually comes out ahead on Linode.

Jurisdiction is the second axis. Hetzner is a German company with most of its capacity in Germany and Finland, which keeps the data protection story short for European customers. Linode operates European regions but contracts through a United States parent, and buyers with strict requirements should read the data residency and sovereignty discussion before assuming either arrangement satisfies their auditors.

• If your procurement team wants an EU contracting party, Hetzner answers that directly.
• If your procurement team wants a vendor with an enterprise account structure and formal escalation paths, Akamai answers that directly.
• If you already buy delivery or security services from Akamai, consolidating compute onto the same relationship is a real advantage.
• If you are optimising a fixed monthly infrastructure budget, Hetzner usually wins on arithmetic alone.

Hardware, storage and the network

Hetzner builds and operates its own data centre parks, and that shows in how the cloud is organised: a small number of locations, each very large, with a fast internal network that people routinely use for distributed storage and clustered databases. Cloud servers carry local NVMe capacity, Cloud Volumes add network-attached space, and Storage Box provides a cheaper destination for backups over protocols such as SFTP and Samba. Documentation for the cloud products sits at the Hetzner Cloud docs.

Linode's infrastructure question is more about the network above the instance than the instance itself. Akamai's backbone and peering relationships give the platform a delivery story no other provider in this bracket can tell, and distributed compute regions exist specifically to place workloads closer to that edge. Instances come in shared, dedicated, high memory, premium and GPU shapes, with Block Storage for attachable capacity and Object Storage for bucket-style data.

Networking is where Linode has the richer toolkit. VPC gives layer 3 isolation, VLANs give layer 2 segments for appliances or clustering protocols that expect a flat network, and NodeBalancers front traffic. Hetzner offers Private Networks with user-defined routes, cloud firewalls and both floating and primary IP addresses, which covers most designs but stops short of layer 2. Teams planning either should read the VPC networking basics primer first.

Each platform playing to type

Hetzner

Hetzner is at its best when the requirement is a lot of reliable compute, operated by people who know what they are doing, in or near Europe. Dedicated vCPU instances give consistent performance, the Arm64 line offers an efficiency option for software that runs well on that architecture, and the path to Dedicated Root Servers means a workload can graduate to physical hardware without changing vendor. Predictable traffic allowances make it attractive for anything that moves a lot of bytes.

There is also a cultural strength worth naming: the platform changes slowly and deliberately. For infrastructure you intend to keep for years, a provider that is not constantly rearranging its product line is an asset rather than a disappointment.

Linode

Linode is at its best when a small team wants managed pieces without moving to a hyperscaler console. LKE plus Managed Databases plus NodeBalancers covers the architecture most web applications actually need, and it does so with a control panel that a new engineer can understand in an afternoon. Included round-the-clock support removes a real risk for businesses without an on-call rota.

The Akamai relationship adds a dimension that matters more over time: a wide region list, edge-adjacent placement and the option to buy delivery and security services on the same paper. The Linode and Akamai Cloud page explains how far that integration has gone.

What you give up either way

Hetzner

There is no managed Kubernetes, no managed database, no platform-as-a-service, no serverless runtime and no GPU capacity. Every layer above the operating system is yours to build and keep running. New accounts begin with conservative resource ceilings that lift as the account matures, which can be awkward if you need a large fleet on day one. Coverage outside Europe is limited, and support, while technically sound, is not marketed as a round-the-clock service.

Linode

The managed catalogue is narrower than a hyperscaler's and narrower than some of its direct peers: two database engines, no first-party platform-as-a-service, no serverless functions and object storage without a bundled delivery toggle. Value per instance is good but not extraordinary, so budget-driven projects will find Hetzner stretches further. Integration into Akamai has also meant a period of renaming and repositioning that long-standing customers have had to follow.

Which one belongs in your stack

Choose Hetzner when compute volume is the constraint, when your users and your legal requirements point at Europe, and when your team already runs its own databases and clusters comfortably. Research groups, media companies, agencies running large shared infrastructure and cost-sensitive software teams all do well here. Start from the Hetzner account page.

Choose Linode when you want managed Kubernetes and managed databases without a hyperscaler learning curve, when included support at any hour is worth paying for, or when a wide region list and the Akamai edge relationship match your audience. The Linode account page covers the available variants.

If you want to triangulate further, Hetzner against Vultr adds a provider with more locations and GPUs to the comparison, while UpCloud against Hetzner tests the value argument against a storage-performance specialist.

Answers to common questions