Few UK businesses plan to overspend on cloud. The original promise made sense: migrate off on-premise infrastructure, pay only for what you use, and watch costs drop. But years into that migration, finance directors across the country are opening Azure or AWS invoices that look nothing like the projections they originally approved.
Individual teams spun up resources, workloads multiplied across departments, and nobody was ever tasked with switching things off. Cloud sprawl has since become one of the greatest budget drains in British enterprise, and the scale of it still catches most leadership teams off guard.
What’s actually driving the problem, and why can’t FinOps teams fix it on their own? The answer involves poor governance, missing accountability structures, a cloud market that’s growing faster than most organisations can keep up with, and a widespread assumption that someone else is keeping an eye on the bill.
Where the money actually goes
Nobody gets an alert saying “you’re wasting £150,000 a year on cloud resources you don’t use.” That waste builds gradually, and exponentially. Virtual machines stay running months after the projects they supported have wrapped up. Storage volumes that were never deleted continue racking up charges, orphaned snapshots keep backing up resources that no longer exist, and development environments hum along through weekends and bank holidays with nobody logged in.
IDC forecasts that global public cloud spending will surpass $1 trillion in 2026, growing more than 21% year on year. At that rate of acceleration, even a small percentage of waste adds up to enormous sums. IDC projects Western Europe will account for roughly $255 billion of that global total in 2026, making it the second-largest regional cloud market behind the US.
Industry estimates consistently put wasted cloud spend at between 25% and 35% of total infrastructure budgets, depending on how mature an organisation’s cost controls are. For a mid-sized London business spending £500,000 a year on cloud, that translates to somewhere between £125,000 and £175,000 vanishing into services that nobody touches.
Every new team, every new project, and every new AI experiment will add resources to the estate. Without a clear process for decommissioning them, those resources will stay active long after anyone remembers what they were for.
Why FinOps teams can’t solve this alone
FinOps has become the default answer for cloud cost management, and that reputation isn’t unearned. The FinOps Foundation’s 2026 State of FinOps report, which drew on responses from nearly 1,200 organisations globally, found that 98% of practitioners now manage AI-related cloud spend. Two years earlier, that figure was just 31%. No other category in the survey’s six-year history has moved that fast.
But knowing where the money goes and having the power to stop it going there are two very different things. FinOps teams can typically identify waste with precision. What they often lack is the authority to force changes across engineering, product, and operations teams. An analyst can flag a cluster of VMs that hasn’t been touched in 90 days, but if the engineering team that owns them doesn’t respond to the ticket, the bill keeps ticking over regardless.
FinOps will give an organisation visibility into its cloud spend. It won’t, by itself, deliver governance over that spend. Most organisations have the first part covered and are failing badly at the second.
What governance frameworks actually look like
Businesses that manage to keep cloud costs under control tend to share certain habits. They tag every resource by owner, project, and expiry date from the moment it’s created. They run automated policies that flag or shut down anything idle beyond a defined threshold. And instead of leaving cost reviews to quarterly meetings, they build them into sprint cycles and deployment pipelines so that spending gets reviewed as part of the normal rhythm of work.
For a lot of organisations, the gap between understanding what they should be doing and actually doing it comes down to internal capacity. Cloud environments are complicated, and most in-house IT teams are already stretched thin across too many priorities. That’s exactly why many of these UK firms are now bringing in Azure consultancy specialists to audit the estate, identify what’s running and what’s redundant, and build tagging and governance policies from the ground up can give a business the clarity it needs to stop bleeding money on unused infrastructure.
Tagging, policies, and accountability
Enforcing a tagging standard remains the single most effective measure any organisation can take. Once every resource carries a cost centre, an owner, and a review date, waste detection can be automated instead of relying on someone to manually dig through invoices.
Pairing that standard with Azure Policy or equivalent tooling will prevent resources from being created without the required tags in the first place, closing the loop before any waste has a chance to accumulate. The businesses getting this right don’t treat it as a one-off cleanup project. They’ve embedded it into how their teams provision infrastructure from day one.
What this means for the CFO’s next cloud review
Cloud sprawl often persists because governance hasn’t kept pace with adoption. The migration itself is the easy part. Managing cloud resources on an ongoing basis, reviewing what’s actually being used, right-sizing where possible, and decommissioning the rest is where most businesses lose control, and that failure compounds month after month as new workloads pile onto an already ungoverned estate.
If the cloud bill has been climbing steadily and nobody in the business can clearly explain why, that’s a signal that shouldn’t be ignored. A single dashboard or a quarterly cost review won’t fix it. What will make the difference is building clear rules, embedding real accountability processes, layering in automation, and making sure someone actually owns the outcome on an ongoing basis. Every month of delay adds to the eventual cleanup bill, and by then the organisation will have spent more undoing the damage than it would have spent preventing it.
Leave a Comment