Affiliate links present. Disclosure
Hetzner
Honest European hardware at honest prices
When something breaks, there is no default path — recovery depends on your decisions.
This only makes sense if you accept the trade-offs above.
Why teams accept these trade-offs
- Teams optimising cost accept European-only geography for unmatched resource density
- Transparent pricing with no renewal cliffs — the cost you model is the cost you pay
- No hand-holding, but teams with Linux competence get more per euro than anywhere else
Hetzner doesn't oversell its infrastructure. The company operates large-scale physical data centers in Germany and Finland, runs them efficiently, and passes that efficiency to customers as compute pricing that most cloud providers cannot match at equivalent specs. The product is the hardware. The pricing is the argument. Everything above the OS is the customer's responsibility. Outside Europe, Hetzner effectively doesn't exist. And inside Europe, if something breaks at the stack level, the resolution is entirely yours.
What you're actually getting
Details may vary by plan and region
Profile
These scores describe capability — not how easy this will be to operate.
How This Infrastructure Actually Works
Hetzner operates KVM-based VPS instances with local NVMe storage, modern multi-core CPUs, and 1Gbps uplinks on shared cloud VPS and up to 10Gbps on dedicated servers. Monthly bandwidth allocation is generous — 20TB included on most cloud VPS plans — which effectively eliminates bandwidth as a cost variable for the vast majority of workloads. You provision an instance, choose an OS, and receive root access. The server is yours to configure.
Hetzner Cloud's additional primitives — Firewall, Load Balancers, Floating IPs, Private Networks, Volumes (block storage), Snapshots, and Object Storage (S3-compatible) — are available as add-ons through the same Hetzner Cloud dashboard and API. These are infrastructure primitives, not managed services. You configure them; Hetzner provides the infrastructure beneath. The Cloud API is well-documented and supports Terraform, making infrastructure automation straightforward.
Hetzner also offers Dedicated Root Servers — physical bare-metal hardware — at price points that compete with many providers' premium VPS tiers. For compute-intensive workloads that require physical resource isolation or high memory density, dedicated servers are available in pre-configured options or via auction (used hardware at significantly reduced prices). This puts Hetzner in a unique position: the only major European provider where the choice between VPS and dedicated bare metal is a practical consideration rather than a cost-prohibitive one.
Core Philosophy
Hetzner's philosophy is honesty in both hardware and pricing. The company does not inflate specifications, maintain opaque pricing models, or wrap its infrastructure in managed-service overhead that customers pay for without using. What you see in the pricing table is what you get. There are no egress charges designed to create surprise billing. Bandwidth allocations are generous by market standards. The contract terms are straightforward.
This honesty extends to what Hetzner doesn't provide. The company makes no claims about developer experience, managed services, or support depth beyond what its infrastructure-level ticket system covers. The product is a well-built server at a low price, delivered with the expectation that the person buying it knows how to use it. For developers and teams with server administration capability, this directness is the point. There is no managed layer you're paying for but not using.
The European positioning is also deliberate. Hetzner is a German company, operates primarily in Germany and Finland, and has built deep peering relationships with European network infrastructure. For organizations with EU data residency requirements or whose user base is concentrated in Europe, Hetzner's network performance within the continent is a practical advantage rather than a compliance checkbox.
Performance & Behavior
Hetzner's dedicated CPU cloud instances deliver consistent compute performance. Physical cores are not shared between tenants on dedicated CPU plans, which eliminates the performance variance that affects shared-tenancy environments under load. Shared CPU instances use a credit-based burst model — appropriate for variable workloads, less appropriate for applications with sustained high CPU demand.
NVMe local storage is fast on sequential I/O benchmarks. The constraint is that local storage is not distributed — disk performance is subject to the dynamics of the physical host, which under heavy contention from neighboring VMs can show variance. For most web application and database workloads, this is not a practical issue. For applications with extreme I/O sensitivity under concurrent load, the distributed storage architecture of providers like UpCloud provides structural protection Hetzner's local NVMe model doesn't replicate.
Network performance within Europe is strong. Hetzner peers directly with major European internet exchanges (DE-CIX, AMS-IX, LINX), which delivers low latency to European users and strong interconnection to major European network providers. Inter-continental latency to US or Asian users is higher, constrained by Hetzner's limited non-European presence.
Pricing Logic
Hetzner's compute pricing is among the lowest in the European market for equivalent specs. A cloud VPS with 4 dedicated vCPU, 8GB RAM, and 160GB NVMe SSD runs around €12–15/month — a price point where most providers offer shared CPU and half the storage. Dedicated servers with 6 physical cores and 64GB RAM are available under €50/month in many configurations. The price-to-hardware ratio is consistently difficult to match from a European provider.
The 20TB monthly bandwidth inclusion eliminates a cost variable that adds to bills at providers charging per-gigabyte overages. Additional bandwidth beyond the allocation is charged at a low per-gigabyte rate, but the allocation is generous enough that most workloads never reach it. Load Balancers, Object Storage, and Volumes are priced separately at competitive rates for European infrastructure.
Trade-offs
You gain the best price-to-hardware ratio in European cloud infrastructure, generous bandwidth inclusion, a clean API and dashboard, and a bare-metal option that stays within a realistic budget. The infrastructure is well-built and the pricing is honest. For teams with server administration capability, Hetzner is extremely difficult to displace on cost efficiency.
You give up managed services. There is no managed Kubernetes in the same sense as DigitalOcean's DOKS, no managed database service with automatic failover, and no support tier that covers application-level assistance. If something goes wrong at the OS or application layer, the path to resolution is the customer's own expertise. You also give up geographic presence outside Germany and Finland — a real constraint for applications requiring low-latency deployment outside Europe.
When It Fits
- Development teams with server administration capability who want maximum compute value in European infrastructure
- Applications with a European-primary user base where the network proximity advantage is real
- Cost-sensitive projects where compute budget is the binding constraint and managed services are not required
- Organizations evaluating dedicated server hardware at cloud-adjacent prices for compute-intensive workloads
- Teams using Infrastructure-as-Code who want a well-documented API without paying for managed overhead
When It Breaks
Hetzner's constraints are structural, not incidental:
- When infrastructure needs to serve users in North America or Asia with low latency — Hetzner's non-European presence is minimal
- When managed databases with automatic failover are required and the team doesn't want to operate them independently
- When storage I/O consistency under sustained concurrent load is a hard application requirement — UpCloud's distributed MaxIOPS architecture provides structural guarantees Hetzner's local NVMe model doesn't
- When the team lacks server administration expertise and needs a managed or guided environment
Alternatives
DigitalOcean is the most common alternative for teams that want a managed services catalog alongside their compute. The price differential is real — DigitalOcean charges more per gigabyte of RAM and storage — but the managed databases, Kubernetes, and object storage reduce operational overhead for teams that use them. The DigitalOcean vs Hetzner comparison makes the trade concrete.
UpCloud provides a reliability-first alternative for teams whose I/O consistency requirements exceed what local NVMe storage provides. The MaxIOPS distributed storage and 100% uptime SLA address a specific class of workload that Hetzner's architecture doesn't serve as well. The premium over Hetzner is real. See UpCloud vs Hetzner.
Linode (Akamai Cloud) competes in a similar developer cloud tier with a broader global footprint and Akamai's edge delivery integrated at the account level. For teams that need European cloud plus global CDN from a single provider, Hetzner vs Linode clarifies where each platform's advantage is real.
Verdict
Hetzner makes sense for technically capable teams building European-primary infrastructure where compute budget efficiency is the primary constraint. The price-to-hardware ratio is genuinely exceptional, the bandwidth inclusion is generous, and the infrastructure is solid. It doesn't make sense for teams that need managed services, global geographic coverage, or a provider that covers application-level support. Within its scope — European infrastructure for teams that manage their own stack — Hetzner is very difficult to displace.
Go to Hetzner →You can start small — no commitment needed.
In practice
"Hetzner wins on price. DigitalOcean wins on sanity. The question is which one your team actually needs."
Where to go next
Closest alternatives to this model.
© 2026 Softplorer