Do we actually need more GPUs, or just the right one?

RubberDuckOps — Wed, 29 Apr 2026 13:58:04 +0000

Last week I was at a tech meetup in Berlin where we got into something: are teams actually making deliberate infrastructure decisions, or just reacting to AI hype? Three practitioners shared their real experience. Here's what stuck with me.

Don't lock in before you understand your workload

Ömer from #Youzu talked through their migration off hyperscalers after getting trapped by credits and tight service coupling. Not a hypothetical, they went through it.

His takeaway: decouple early so you can move workloads freely. Know roughly where you're heading before you build, then migrate toward full control progressively. Portability isn't a nice-to-have, it's insurance.

Question the GPU arms race

David from #SteliaAI made a point that a lot of teams need to hear right now: most people are provisioning for scale they don't have yet.

His suggestion was almost counterintuitively simple:

Start with half compute, half control plane
Get customers
Then revisit

Don't optimize for a scale you haven't reached yet, because the shape of your workload will change by the time you get there.

Observability is not optional

Felix from #Cloudeteer made the case that GPU utilization metrics alone don't tell the full story. You can be running at 100% capacity and still be producing wrong outputs.

Traces — not just metrics — are what let you catch problems before they fail silently. If your AI stack doesn't have tracing today, you're flying blind with a full tank.

The thread running through all three talks

AI hype is driving infrastructure decisions that don't match actual workload needs. Every speaker arrived at the same place from a different direction: start lean, stay observable, don't couple yourself to a provider before you understand what you're building.

Now, over to you

Are you provisioning GPUs reactively or from a clear workload map?
Have you ever scaled back after realizing you over-provisioned?
Where does observability fit in your AI stack today?

Drop your experience in the comments.

Looking for a European dedicated server or VPS? Here's what to consider.

RubberDuckOps — Tue, 03 Mar 2026 13:29:53 +0000

Hardware costs are moving across the industry. RAM, SSDs, AI infrastructure demand - it's affecting everyone. If you're evaluating your infrastructure options right now, here's an honest look at what matters and where we fit in.
Disclaimer: I'm on the infrastructure team at Leaseweb. EU-native, Netherlands-owned.

What actually matters when choosing a provider

Price per spec is the obvious starting point but it's rarely the whole story. A few things worth thinking through:
Where is your data? If you're in a regulated industry or just care about GDPR, EU-native infrastructure matters. Not EU region of a US company. Actually EU-owned and operated.

How predictable is your bill? Hourly cloud pricing looks cheap until you're running sustained workloads. At high utilisation, dedicated or longer-term contracts almost always win on cost. The maths changes fast above 60-70% utilisation.
What happens when something breaks? Support SLAs vary wildly in this space. Worth checking what you're actually getting before you need it.

What we offer

VPS from €3.59/month (2 vCPU, 4GB RAM, 80GB NVMe). Good for staging environments, isolated workloads, smaller production setups.
AMD EPYC dedicated servers from €112/month. Full root access, unmanaged, API-ready for IaC. EU sovereign data centres, DDoS protection standard.
Contract terms from 1 month up to 3 years. The longer the commitment, the better the rate - up to 25% off. You lock in your pricing upfront, no surprises for the duration of your term.
We're not the cheapest option in every category.
On price-performance for sustained EU workloads, we're worth a spot on your shortlist. One thing worth checking: if inter-node throughput is critical to your workload, compare our network specs against your requirements before committing

Is it worth switching if you're happy where you are?

Honestly, maybe not. Switching has friction and if your current setup is working, the disruption cost is real.
But if you're evaluating options, spinning up a test environment costs nothing.

Drop a comment with your current setup: vCPU count, RAM, storage, region. Happy to spec match or answer any questions you have!

DEV Community: RubberDuckOps