UK firms report hiring squeeze from unexpected infrastructure costs | Magica
UK firms report hiring squeeze as study estimates £1.2bn in unexpected AI charges
Editorial Team
••📖7 min read
Research commissioned by cloud-storage supplier Wasabi estimates £3.8 billion in annual UK infrastructure overspend attributable to hard-to-forecast provider charges, including £1.2 billion for AI. Businesses report hiring and investment trade-offs, but recovering that money and turning it into jobs remain conditional.
CEBR estimates £3.8 billion in annual UK infrastructure overspend linked to unpredictable provider charges, including £1.2 billion for AI and machine learning.
Hiring was the leading budget sacrifice identified by UK respondents making cost-driven trade-offs across infrastructure.
The sponsor sells an alternative storage service; neither the survey nor its economic model establishes the savings a particular customer could achieve.
UK businesses report cutting back hiring and technology investment as infrastructure bills exceed expectations, according to research commissioned by cloud-storage supplier Wasabi Technologies. The study estimates £1.2 billion a year in unexpected AI and machine learning provider charges within a £3.8 billion UK infrastructure total.
The recruitment findings concern infrastructure pressures across technology categories. Among UK organisations making cost-driven trade-offs over the preceding year, 38% identified hiring in technology, data and AI as most likely to lose budget, according to coverage of the Wasabi-commissioned UK findings. Technology and software investment followed at 37%, and training and workforce development at 34%. Overall, 77% reported at least one such trade-off. Those figures do not isolate the employment effect of AI bills alone.
The Centre for Economics and Business Research (CEBR), a London economics consultancy established in 1992, produced the economic estimates. Wasabi, which markets storage for AI, analytics and other applications, has a commercial interest in more predictable infrastructure pricing. Its November 10 announcement promotes the research alongside customer and partner accounts of its services.
Wasabi-commissioned CEBR research identifies hiring, technology investment and training as UK budget sacrifices under infrastructure cost pressure. The findings do not isolate AI bills’ employment effects. Source: DataCentreNews.
What the £3.8 billion estimate measures
The estimate covers spending respondents attributed to opaque, unpredictable or difficult-to-forecast provider charges. It is narrower than all infrastructure overspend: higher usage, changing internal requirements and planning assumptions can also push bills above budget.
Palantir’s co-founder and CEO says he could back a Democrat in 2028, while excluding a Democratic Socialists of America member. His Democratic history complicates talk of a political conversion; the more immediate question is how his government software company will answer congressional scrutiny.
Across the eight markets studied, respondents attributed an average 32% to 36% of overspend to provider charges, depending on the infrastructure category. The UK figure for AI and machine learning was about 39%, according to the primary report. This is respondents' attribution, rather than an independent audit of suppliers' invoices.
CEBR combined survey estimates of typical expenditure, overruns and the provider-attributed share, adjusted for how many firms experienced overspend, then scaled the results using official business counts. It used median expenditure within market and business-size groups to reduce the influence of very large spenders, substituting pooled data where national samples were insufficient.
The underlying survey, conducted by Sapio Research in July and August 2026, covered 690 technology and infrastructure decision-makers selected for their involvement in those decisions. It spanned the UK, US, France, Germany, Italy, Japan, Australia and New Zealand; 76% represented large businesses. The stated ±3.7 percentage-point margin of error applies to the total sample, not the UK subset or the modelled monetary estimates.
The methods also define SMEs as firms with 10–49 employees and large businesses as those with at least 250, without explaining the treatment of firms between those bands. That leaves a question about the population coverage behind the national estimates.
Across all eight markets, CEBR estimates $63.4 billion in annual provider-attributable overspend, including $23.4 billion for AI and machine learning. This is an eight-market total, not a worldwide estimate or a measure of spending wasted on unsuccessful AI projects.
More UK AI overspenders cite licensing than token charges
AI and machine learning account for 32% of the estimated UK provider-attributable overspend. Almost three-quarters of UK respondents reported infrastructure spending above expectations over the previous year; UK coverage says 61% of those affected identified AI and machine learning as their biggest overspend area.
The reported UK breakdown of the commissioned research says 22% found AI infrastructure costs unpredictable. Among UK businesses that overspent on AI, 66% cited unexpected AI and model licensing fees, 43% cited integration and deployment expenses, and more than 36% cited higher-than-expected token or inference charges. These are contributing factors, rather than mutually exclusive shares of the bill.
For a buyer, that means the price of running a model is only part of the budgeting problem. Licensing and putting systems into service also matter. The primary report's eight-market figures differ from the UK breakdown: licensing was cited by 49% of AI overspenders, integration, deployment or model operations by 45%, and token, inference or usage-based charges by 38%.
CEBR’s Wasabi-commissioned eight-market survey ranks licensing ahead of integration and token charges. These figures differ from the UK breakdown; contributing factors are not mutually exclusive. Source: CEBR research report.
Switching is possible, but storage is only one layer
UK respondents also reported operational constraints. Some 84% had experienced delays from a lack of suitable hardware or compute capacity. Among businesses considering a supplier switch, the UK findings put additional costs from fragmented data environments at 91%, and restricted or delayed access to needed data at least once in the preceding year at 85%.
The research sponsor offers evidence that switching can work. In Wasabi's announcement, Allan Bettarel, chief technology officer of recording-service customer tl;dv, said the company rejected a cheaper storage tier because retrieval would be too slow, then moved from Amazon Web Services (AWS) to Wasabi. He said the move preserved immediate access while making annual costs predictable. This is a customer testimonial published by the supplier, rather than a comparative performance or cost study.
Moving storage also does not resolve every licensing, integration or compute-capacity problem identified in the survey. The report cautions that its observed relationships between switching difficulties, budget pressure and technical compromises cannot establish which causes which; firm size, sector or technology complexity could influence them together.
Independent evidence supports the concern about customer choice. The Competition and Markets Authority (CMA), the UK's competition regulator, concluded in its July 2025 cloud-market decision that AWS and Microsoft, the two largest providers, held significant market power. Fewer than 1% of customers switched annually. It identified data-transfer fees and technical barriers, while also noting that customers valued providers' quality and innovation and that prices had risen for some services and fallen for others. Its findings do not independently verify CEBR's overspend estimate.
CEBR’s Wasabi-commissioned model estimates potential UK investment of US$1.9 billion, training expenditure of US$0.8 billion and 8,235 additional full-time-equivalent jobs if overspend were recovered. These are conditional estimates. Source: CEBR research report.
The next test is lower bills and actual reinvestment
The regulatory picture has moved since that decision. In an October 16, 2026 update, the CMA said Microsoft and AWS were taking steps including removing transfer fees for customers changing providers and reducing them for customers using multiple clouds. It said barriers remained, kept a formal cloud strategic-market-status investigation under review, and planned a further update in spring 2027.
Whether lower barriers deliver the economic gains in Wasabi's research remains unproven. The primary report's UK scenarios estimate that recovered expenditure could support $1.9 billion in investment in R&D, AI and new technology, $0.8 billion in training, and 8,235 additional full-time-equivalent jobs. CEBR derives the employment figure from an assumed allocation of savings and employment costs, rather than counting lost jobs or recruitment commitments.
Those monetary scenarios are denominated in US dollars in the primary report. UK coverage gives the training figure as £800 million, a currency discrepancy that the retained material does not reconcile.
The model also estimates a $1.9 billion long-run UK GDP uplift from additional capital-forming investment. It excludes hiring and training from that GDP calculation and does not forecast an immediate annual gain. Across the eight markets, the central GDP scenario is $20.4 billion, with a $13.6 billion–$27.2 billion range under different investment-to-output assumptions.
Businesses' stated intentions point toward reinvestment: 49% of affected UK respondents said they would put recovered funds into AI, automation or data science. Establishing the payoff requires evidence that charges actually fall, migration and operating costs do not absorb the savings, and businesses then spend the released money on the investment and hiring envisaged.
ByteDance’s video model is reaching creators through multiple platforms with 30-second 4K output. Posted prices and draft workflows show why the choice of distributor still matters.
Rox’s $1.3 billion valuation backs AI agents for large enterprise sales teams. Its growth and cost claims face a crowded market, while integration with existing systems leaves the case for replacing Salesforce unresolved.
Google’s Gemini Decision 1 scores predefined choices across text, images, audio, video and PDFs for $0.10 per million input tokens. Its shared-input billing targets recurring workflow costs, but cheaper decision models, rival caching and low-cost generative alternatives complicate the savings claim.
Treasury and other agencies reportedly want Singapore cloud provider Megaspeed on a U.S. trade blacklist. Commerce denies a reported cabinet confrontation over the delay, while Nvidia says it has blocked shipments since June 2025.
United is restricting unauthorized AI agents from searching, booking and managing travel through its platforms and intermediaries. The airline cites security and safety risks as it reserves control over future integrations.
A controlled study found that pharmaceutical ads shifted AI recommendations toward promoted drugs without reducing overall clinical accuracy. The results expose a gap in accuracy tests, but do not establish effects on prescriptions or patient outcomes.
Former AWS computing chief Dave Brown will succeed Santosh Janardhan in early 2027 as Meta weighs its next capacity budget after a quarter of negative free cash flow.
Google is rolling out its Gemini app on Apple CarPlay, bringing voice conversations linked to Gmail and Google Calendar. It joins existing in-car AI options, while some personalized features remain restricted by age and region.
Google Labs’ Tabletop lets groups generate and use shared itineraries, polls and other interactive tools. The US adult-only experiment builds on an idea already offered by shareable AI app builders, leaving its practical advantage to be tested.
Anthropic’s IPO has reportedly attracted orders multiple times the available shares. Its planned $3 billion–$3.5 billion Japanese allocation would exceed Japan’s reported domestic IPO proceeds so far this year, but indications of interest are not completed sales.
Google warned that charging for news links would force it to stop distributing Indonesian news, while Meta and TikTok challenged other parts of the copyright bill. Existing Indonesian payment mechanisms and overseas deals complicate the platforms’ warnings about the cost to publishers.
Utrecht is investigating restrictions on camera glasses in municipal buildings, with findings expected in early 2027. The city must establish whether a ban is lawful, enforceable and compatible with assistive uses.
The USDT issuer’s 800-million-parameter BotPsy beats a larger Qwen model on five of six company-reported agent benchmarks. Its local-AI ambitions still face missing handset measurements and explicit limits on sensitive uses.
Google DeepMind’s AlphaGenome 2 improves cell-specific DNA predictions and accepts new single-cell data without fine-tuning. Company benchmarks show gains at high confidence, but limited recall, complementary tools and greater computational demands leave experimental validation essential.
Coinbase is rolling out server-enforced trading limits and portfolio-loss triggers for AI agents connected to customer accounts. Product-category permissions remain forthcoming, and pausing new trades does not establish a ceiling on losses.
JetBrains IDEs 2026.3 bundle Air for running coding agents and reviewing their changes, while the standalone Air app is in maintenance mode. Existing provider subscriptions remain usable, but billing, security controls and execution options vary by integration.
Amazon’s shares and convertible notes could be worth $271.8 billion at the top of Anthropic’s proposed IPO range. Existing accounting gains, a 12% ownership cap and Anthropic’s computing commitments shape what that paper value means.
Nvidia chief Jensen Huang argued in May 2025 that US chip restrictions were strengthening Chinese rivals. Filings show a substantial commercial stake, but lost sales and market share alone cannot establish whether the controls failed their security purpose.
Reported demand has Anthropic’s bankers considering a price above its filed $70–$78 range. The combined financing could exceed $100 billion, but existing cash, uneven operating results and long-term computing commitments complicate the valuation test.
