The Cloud's Dirty Little Secret: "Infinite Scalability" Has a Waiting Room

We spend our days placing talent inside some of the UK's most critical data centre and infrastructure operations. And right now, a conversation keeps coming up that deserves a wider audience.

Azure UK regions are at capacity.
Not "a bit busy." Not "slightly constrained." At. Capacity.

If you're trying to provision additional compute, storage, or networking resource in Azure's UK South or UK West regions right now, you may well hit a wall. The path forward? Raise a support ticket with Microsoft, request a quota increase, make your case, and then wait. And quite often, the answer that comes back is no.

The "infinite scalability" promise was the whole pitch.

For the better part of a decade, the cloud evangelists told us to hand back the raised floor space and ditch the capex headache. Why own capacity when you can rent as much as you need, whenever you need it? That framing drove an enormous wave of migration and led businesses to deprioritise in-house infrastructure expertise, shrinking internal teams on the assumption that physical infrastructure was someone else's problem now.

Except those same businesses are now discovering that elasticity has very real physical constraints behind it. Someone has to build the halls, run the power, and rack the kit. And right now, in the UK, that build-out hasn't kept pace with demand.

So what does this mean in practice?

For businesses running time-sensitive workloads or DR scenarios that depend on rapid Azure provisioning, this is a genuine operational risk. The assumption that capacity will simply be there is no longer safe. And for organisations that have exited their own data centre footprint entirely? Options are limited. You can't re-rack servers you sold three years ago.

The uncomfortable truth from where we sit.
The talent that understands physical infrastructure has been in short supply for years, partly because businesses told themselves they no longer needed it in-house. That calculation is being revisited, fast. Hybrid infrastructure, colocation strategy, sovereignty requirements, and now cloud capacity constraints are all pushing demand back toward people with real, hands-on data centre expertise.

The cloud isn't going anywhere. But neither is physical infrastructure. There are buildings, power feeds, and cooling systems underneath every workload, and when those things are constrained, the elasticity stops.
If your cloud strategy has no contingency for the day your hyperscaler says "sorry, no quota available," it might be worth revisiting.