Connect with us

AI

UK’s AI Storage Costs Keep Draining Budgets Despite Fee Pledges

UK firms funnel 62% of AI budgets into data and storage, yet a light-touch fix from Britain’s competition regulator has not lifted returns.

Published

on

UK businesses now put 62% of their AI budgets into data, storage and processing power, versus 36% for software. Only a quarter report a positive return on investment (ROI) from that spending, according to the 2026 Global Cloud Storage Index from Wasabi Technologies, a cloud storage provider that competes directly with Amazon, Microsoft and Google.

That gap between spending and payoff is compounded by a cloud market Britain’s competition regulator has just spent three years investigating, and found wanting.

The Data Layer Is Swallowing AI Budgets

The imbalance shows up first in how AI money gets divided. Wasabi’s research, conducted with research firm Vanson Bourne among 1,700 IT decision-makers worldwide including 200 in the UK, found data, storage and compute now claim 62% of AI budgets, with software a distant second at 36%.

Kevin Dunn, Wasabi’s vice president and general manager for EMEA (Europe, the Middle East and Africa), has pointed to why that split matters. Storage costs are “consuming a growing share of AI budgets,” he has said, arguing “a more flexible approach to data management is needed to turn investment into meaningful returns.”

Part of the shortfall starts before storage bills even enter the picture. AI models depend on vast, often unstructured datasets that need extensive cleaning and preparation, and that groundwork is frequently where projects stall first.

Businesses have not abandoned the bet despite the weak returns so far. In a separate finding, 48% say they are confident of a positive ROI once infrastructure improves, and most plan to keep raising AI infrastructure spending regardless of today’s results. The stall echoes a pattern playing out elsewhere in enterprise AI: India’s global capability centers face a similar pilot-to-production gap keeping promising projects from reaching the scale where they pay for themselves, as detailed in India’s GCCs Have an AI Innovation Window and a Pilot Problem.

The scale of the imbalance shows clearly across Wasabi’s UK findings:

Metric UK Figure
Share of AI budget spent on data, storage and compute 62%
Share of AI budget spent on software 36%
Businesses reporting positive ROI from AI projects 25%
Storage spend that goes to fees rather than capacity 48%
Businesses that exceeded their cloud storage budget in 2025 46%
Businesses using a hybrid, on-premises plus public cloud, storage strategy 72%

Every one of those figures traces back to the same mechanism: how cloud providers bill once data starts moving, not just sitting in place.

Where Does the Rest of the Storage Bill Go?

Mostly to fees, not capacity. Wasabi found 48% of UK cloud storage spending in 2025 went toward charges for moving data and calling application programming interface (API) operations to access it, rather than the cost of the storage itself, and that fee structure bit hardest as AI workloads scaled up.

  • Egress fee – the charge a cloud provider bills when data leaves its network, whether to another cloud, an on-premises system or an end user; it applies on top of, not instead of, the monthly cost of storing that data.

AI workloads generate an unusual amount of that kind of movement. Training runs pull datasets repeatedly, retrieval systems query stored data constantly, and each request can carry its own charge under some hyperscaler pricing models. Wasabi found 46% of UK businesses exceeded their cloud storage budget in 2025. Of those, 84% blamed fee charges specifically, separate from the basic cost of storage capacity.

The Regulator That Called the Market Broken

Britain’s Competition and Markets Authority (CMA), the country’s antitrust regulator, spent three years examining this exact corner of the cloud market. It closed a three-year investigation into UK cloud competition on March 31, 2026, concluding that competition there “is not working well.”

Investigators identified three barriers: market concentration, since Amazon Web Services (AWS) and Microsoft Azure each hold an estimated 30% to 40% of the UK cloud market; technical and commercial hurdles including egress fees and incompatible systems; and Microsoft’s software licensing terms, which the regulator said hurt AWS and Google’s ability to compete.

The CMA’s own investigators had recommended a binding Strategic Market Status (SMS) designation, the toughest tool available under the UK’s newer digital markets powers. The regulator chose not to use it.

What We Know

  • Market concentration: AWS and Microsoft Azure each hold an estimated 30% to 40% of the UK cloud market.
  • Voluntary deal: AWS and Microsoft agreed to lower egress fees and improve interoperability instead of facing a binding order.
  • Open case: A separate CMA investigation into Microsoft’s software licensing practices is still active.

What’s Unconfirmed

  • Real savings: Whether the voluntary commitments will actually lower UK businesses’ bills, since they carry none of the legal force an SMS designation would.
  • Timing: When lower fees reach customer invoices, since Wasabi’s 46% budget-overrun figure was drawn largely from 2025 billing, before the commitments took effect.

A parallel pattern is playing out with a different giant. Google is separately contending with a billion-dollar EU fine over its Android advertising practices, agreed under an entirely different competition order, evidence that European regulators are pressing the same handful of platforms on several fronts at once this year.

A Deadline Hyperscalers Cannot Negotiate Away

Europe is applying pressure through a different lever entirely. The EU Data Act sets a hard deadline of January 12, 2027, after which cloud providers serving EU customers cannot charge any switching costs at all, including egress fees. Until that date, providers may still charge only the direct costs of the switching process itself.

AWS, Google Cloud and Microsoft Azure already announced free data transfer for customers leaving their platforms entirely, back in 2024, under early pressure from the same law. The offers carry conditions. Azure requires a complete exit and cancellation of all subscriptions. AWS excludes services like CloudFront and Direct Connect from the free transfer.

The law targets only the cost of leaving a provider altogether, leaving alone the routine egress charges businesses pay every day just to serve stored data out to users.

The stakes keep rising because AI infrastructure spending shows no sign of slowing. Gartner has raised its 2026 global IT spending forecast to $6.31 trillion, up 13.5%, citing stronger than expected demand for AI compute and data center capacity. IDC separately projects worldwide AI infrastructure spending will reach $487 billion in 2026, up 53% from the prior year, on its way past $1 trillion by 2029.

Why Hybrid Storage Is Winning by Default

Faced with unpredictable fees on one side and a slow-moving regulatory fix on the other, most UK businesses have already chosen a workaround. Wasabi’s index found 72% now run a hybrid strategy, splitting workloads between on-premises data centers and public cloud rather than committing everything to one environment.

  • Sensitive or frequently accessed data stays on company-owned hardware, closer to compliance and audit requirements.
  • AI models and bulk training datasets move to public cloud, where capacity can expand and contract with demand.
  • Workloads shift between the two based on cost, latency and regulatory rules, instead of sitting fixed in one place.
  • Spreading storage across providers reduces reliance on any single vendor and gives businesses more negotiating power over fees.

The approach only works if a business understands exactly how each provider bills for moving data between sites, since transfer charges between environments can erase the savings a hybrid split is supposed to deliver.

It is also colliding with a separate infrastructure constraint already reshaping European AI buildouts, the power and grid capacity now constraining Europe’s AI buildout, since both halves of a hybrid split still need physical data centers to run.

The Vendor Pitching Its Own Fix

Wasabi has an obvious stake in this story. The company sells cloud storage priced at $6.99 per terabyte per month with no advertised egress fees, positioned as roughly 80% cheaper than the major hyperscalers, the exact charges it says are draining AI budgets elsewhere.

That does not make its underlying numbers wrong. Wasabi is a paid research sponsor rather than a neutral observer, and its diagnosis happens to match the product it sells, a detail worth holding alongside the survey data rather than instead of it.

UK and EU businesses still absorbing 2025’s fee increases have one certain deadline to hold onto: hyperscaler switching costs fall to zero across the European Union on January 12, 2027, CMA voluntary commitments or not.

Frequently Asked Questions

Is the UK’s AI ROI Problem Worse Than the Global Average?

Slightly, yes. Wasabi’s survey found 25% of UK businesses report positive ROI from AI projects, compared with roughly one-third worldwide across the full sample Wasabi and Vanson Bourne polled, suggesting UK businesses lag the global average on realized returns.

Does the EU Data Act Apply to UK Businesses After Brexit?

Only indirectly. The Data Act is EU law, so its binding January 2027 deadline on switching and egress fees covers providers and customers operating inside the European Union. The UK left the EU in 2020 and is relying instead on the voluntary commitments the CMA secured from AWS and Microsoft in March 2026.

What Is Strategic Market Status Under UK Competition Law?

Strategic Market Status (SMS) is the binding designation the CMA can apply to a firm with substantial and entrenched market power, letting the regulator impose legally enforceable conduct rules. The CMA’s own investigators recommended it for cloud providers in this case, but the regulator opted for voluntary commitments instead.

Logan Pierce is a writer and web publisher with over seven years of experience covering consumer technology. He has published work on independent tech blogs and freelance bylines covering Android devices, privacy focused software, and budget gadgets. Logan founded Oton Technology to publish clear, no nonsense tech news and reviews based on real hands on testing. He has personally tested and reviewed dozens of mid range and budget Android phones, written extensively about app privacy, and built and managed multiple WordPress publications over the past decade. Logan holds a bachelor's degree in English and studied digital marketing at a certificate level.

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Trending