Cloud Platform Comparison
An honest, side-by-side look at six cloud and infrastructure options. No affiliate links. No preferred vendors. Just the trade-offs.
| AWS | Azure | GCP | Hetzner | OVH | Colo | |
|---|---|---|---|---|---|---|
| Compute cost | $$$ | $$$ | $$$ | $ | $$ | $ |
| Managed DB | Extensive | Extensive | Extensive | Limited | Moderate | Self-managed |
| Egress cost | High | High | High | Low | Low | None |
| Lock-in risk | High | High | Moderate | Low | Low | None |
| Compliance | Broad | Broad (Gov+) | Broad | EU only | EU-focused | Your choice |
| Ops complexity | Managed | Managed | Managed | Moderate | Moderate | High |
| Best for | Complex, elastic workloads | Microsoft-heavy orgs | Data/ML workloads | Predictable compute | EU hosting needs | Maximum control |
AWS
- Compute cost
- $$$
- Managed DB
- Extensive
- Egress cost
- High
- Lock-in risk
- High
- Compliance
- Broad
- Ops complexity
- Managed
- Best for
- Complex, elastic workloads
Azure
- Compute cost
- $$$
- Managed DB
- Extensive
- Egress cost
- High
- Lock-in risk
- High
- Compliance
- Broad (Gov+)
- Ops complexity
- Managed
- Best for
- Microsoft-heavy orgs
GCP
- Compute cost
- $$$
- Managed DB
- Extensive
- Egress cost
- High
- Lock-in risk
- Moderate
- Compliance
- Broad
- Ops complexity
- Managed
- Best for
- Data/ML workloads
Hetzner
- Compute cost
- $
- Managed DB
- Limited
- Egress cost
- Low
- Lock-in risk
- Low
- Compliance
- EU only
- Ops complexity
- Moderate
- Best for
- Predictable compute
OVH
- Compute cost
- $$
- Managed DB
- Moderate
- Egress cost
- Low
- Lock-in risk
- Low
- Compliance
- EU-focused
- Ops complexity
- Moderate
- Best for
- EU hosting needs
Colo
- Compute cost
- $
- Managed DB
- Self-managed
- Egress cost
- None
- Lock-in risk
- None
- Compliance
- Your choice
- Ops complexity
- High
- Best for
- Maximum control
AWS
The market leader with the broadest service catalogue. Best for organisations with complex, elastic workloads that genuinely need managed services at scale. But the pricing is opaque, egress costs are punitive, and many teams end up paying for services they don't need. If your workload is predictable, you're almost certainly overpaying.
Azure
The natural choice for Microsoft-heavy organisations. Strong enterprise integrations, good compliance coverage (especially government), and competitive if you're already in the Microsoft ecosystem. Outside of that ecosystem, it rarely wins on price or developer experience.
GCP
Google's strength is data and ML infrastructure — BigQuery, Vertex AI, and networking are genuinely world-class. For general-purpose compute and storage, it's broadly comparable to AWS and Azure in both capability and cost. Lock-in risk is somewhat lower due to Kubernetes-native tooling.
Hetzner
The price-performance leader for predictable compute workloads. Dedicated servers at a fraction of hyperscaler cost. Limited managed services — you'll need more ops capability. EU-only data centres (Germany, Finland). Ideal for teams that can manage their own infrastructure and want to stop subsidising services they don't use.
OVH
European cloud provider with competitive pricing and growing managed service offerings. Stronger than Hetzner on managed databases and object storage. Good option for organisations with EU data residency requirements who want more support than bare-metal providers offer.
Colocation
Maximum control and lowest per-unit cost — but highest operational complexity. You own or lease the hardware, you manage everything. Makes sense for large, stable workloads where the team has the skills and the business can commit to 3-5 year hardware cycles. Companies like 37signals have shown this can cut costs by 60-80%.
Not sure which is right for you?
The right platform depends on your workload, team, and business constraints. Our Platform Fit Verdict models your specific situation across all six options and gives you a clear recommendation.
Book the Platform Fit Verdict