I’m going to say something that might sound controversial in tech circles: most companies using AWS, Azure, or GCP are massively over-provisioned and under-utilizing what they’re paying for.

There. I said it.

Before the replies start — no, I’m not saying hyperscalers are bad. They’re engineering marvels. If you’re Netflix, running a global CDN serving 200 million users simultaneously, you absolutely need that kind of scale. If you’re training the next large language model on 10,000 GPUs, yes, go with the big guys.

But if you’re a European company running a web application, a handful of APIs, and some background jobs? Let’s talk.

The complexity tax

The biggest cost of hyperscale cloud isn’t the invoice (though we’ll get to that). It’s the complexity.

Have you looked at the AWS service catalog lately? As of this writing, there are over 200 services. IAM alone has a learning curve that could be a university course. Setting up a simple VPC with proper security groups, routing tables, NACLs, and endpoints is a multi-day project for experienced engineers.

And here’s the thing: you need to understand all of this just to do it right. Get one security group rule wrong, and you’re on the news. Forget to set a budget alert, and your weekend experiment costs as much as a used car.

This complexity has spawned an entire industry. AWS consultants, FinOps platforms, cloud cost management tools, certification programs. You’re paying for the cloud, then paying more people to help you use the cloud, then paying for tools to monitor what you’re paying for the cloud.

At some point, you have to ask: is this actually simpler than the alternative?

The invoice problem

Let’s talk money. A typical mid-size European company running on AWS might have:

• A few EC2 instances for their application
• An RDS database
• S3 for storage
• A load balancer
• CloudFront for CDN
• Route 53 for DNS
• Various supporting services

Monthly bill: anywhere from 2,000 to 15,000 euros, depending on size. And that’s before data transfer costs, which are AWS’s quiet profit machine. Egress fees — the cost of data leaving AWS — can add 20-30% to your bill in ways that are genuinely hard to predict.

Now, the same workload on properly-sized European infrastructure? Often 40-60% less. Not because we’ve found some secret — but because you’re not paying the “global hyperscaler” premium for infrastructure you don’t use at global hyperscale.

What you actually need

Most businesses need surprisingly simple things from their cloud provider:

• Reliable servers that stay up and perform consistently
• Good networking with low latency to their users (who are mostly in Europe)
• Security basics done well — firewalls, DDoS protection, encryption
• Backups that actually work when you need them
• Support from humans who respond in hours, not days

That’s it. You don’t need 200 services. You don’t need machine learning-powered auto-scaling that kicks in for traffic spikes you’ll never have. You don’t need a multi-region active-active setup across three continents for your B2B platform with 500 daily users.

You need infrastructure that works, in a location that makes sense for your business, at a price that doesn’t require a dedicated finance person to understand.

“But what about scaling?”

This is the objection I hear most. “What if we grow? What if we go viral? What if we need to scale to millions of users overnight?”

Honest answer? You probably won’t. And if you do, that’s a great problem to have — the kind of problem that comes with revenue to solve it.

Building for hypothetical scale is one of the most expensive mistakes in tech. Companies spend thousands per month on infrastructure “just in case,” for traffic that never comes. Meanwhile, that money could go toward actually building the product that might attract those users.

Start with what you need. Scale when you need to. A well-architected setup on straightforward infrastructure can handle far more than people assume.

The European advantage

There’s another angle here. When your infrastructure is in Europe, specifically in the Netherlands:

• Latency to European users is excellent. Amsterdam is one of the best-connected cities in the world, with AMS-IX being the largest internet exchange by member count.
• Legal clarity. Your data is under EU law. Full stop. No CLOUD Act complications.
• Local support. When something goes wrong at 3 AM, you’re talking to someone in your timezone who speaks your language.
• Pricing in euros. No currency conversion surprises when the dollar moves.

These aren’t flashy features. They’re practical ones. The kind that matter when you’re actually running a business, not just evaluating feature comparison charts.

Making the switch

If you’re currently on a hyperscaler and this resonates, switching doesn’t have to be dramatic. Many companies start by moving non-critical workloads — staging environments, development servers, internal tools — to see how it feels. No big bang migration required.

The technical lift is usually smaller than expected. Containers help a lot here — if your application runs in Docker or Kubernetes, it runs anywhere. That’s literally the point of containers.

The right tool for the job

I’m not here to tell you that hyperscalers are bad or that everyone should leave AWS tomorrow. That would be silly. Some workloads genuinely need what they offer.

But for a lot of European businesses, the honest answer is: you’re paying a premium for capabilities you don’t use, complexity you don’t need, and global reach you’ll never require. And you’re getting less control over your data in return.

Sometimes the right cloud isn’t the biggest one. It’s the one that fits.

---

grn.cloud offers European cloud infrastructure starting at EUR 0.99/month — with Kubernetes, firewalls, load balancers, VPN, and DDoS protection included at no extra cost. Because cloud shouldn’t be complicated.