Research opportunity · Concentration, Stack & Sovereignty
Did UK public bodies switch cloud providers after egress fees were waived?
Question
When UK public-sector cloud contracts came up for renewal before and after the 2024 egress-fee waivers and the CMA cloud investigation, how often did buyers move workloads to a different provider or add a second one, and did switching differ for workloads tied to Microsoft software licensing?
Not yet measuredMaster's thesisLow costNo study found
Why it matters
Regulators in the UK and EU have identified egress fees, interoperability barriers and licensing as switching constraints, and the three largest providers now offer approval-based egress credits to customers who leave entirely. Whether anything changed in actual buyer behaviour is the test of these remedies. Public procurement is one of the few places where cloud buying is visible in open records.
The debate lists this among what we still do not know: measured post-remedy cloud switching
Evidence that motivates it
- The UK CMA found AWS and Microsoft hold significant market power in UK cloud and identified egress fees, interoperability barriers and Microsoft software licensing as constraints on switching and multi-cloud.
The CMA identified egress fees, interoperability barriers and Microsoft licensing as constraints on switching; the claim does not license "solved switching", and measured post-action switching is named as what would change it. - On 25 June 2026 the European Commission preliminarily considered AWS and Azure important cloud gateways that should be designated under the DMA, citing lock-in, switching costs and the growing role of AI in cloud procurement.
The European Commission's preliminary view also cites lock-in and switching costs, so a measured UK baseline would inform how a final EU decision could be evaluated. - Synergy estimated AWS 28%, Microsoft 20% and Google 15% of global cloud infrastructure services revenue in Q2 2026, a combined 63%.
AWS, Microsoft and Google hold about 63% of global cloud infrastructure services revenue; this gives the market context but says nothing about switching, which is what the study measures.
A possible design
First check that the data can answer the question: draw about 100 cloud-related awards at random and record in what share the hosting provider can be identified from the notice, the contract documents or the supplier's own published customer lists. If the share is too low to build renewal chains, narrow the study to a set of buyers and ask them directly through freedom-of-information requests. Then assemble UK public-sector cloud contract awards from 2018 to the latest available year from Contracts Finder, Find a Tender and Crown Commercial Service framework (G-Cloud) records. Identify the hosting provider behind each award (directly, or through resellers and integrators using contract text and supplier lists) and link successive contracts of the same buyer for the same service into renewal chains. Outcomes: incumbent retained, full switch, or second provider added. Compare renewal decisions before and after the 2024 egress-fee waivers with an interrupted time-series design, adjusting for buyer type and contract size, and compare workloads tagged as Microsoft-stack (for example licensing-related terms in the notice) with others. Supplement with 10–15 interviews of public-sector cloud buyers on why they did or did not move.
Data needed
Open UK procurement data: Contracts Finder and Find a Tender publish their notices as Open Contracting Data Standard records through public APIs and bulk downloads, and Crown Commercial Service has published G-Cloud sales with supplier and customer names (whether its current releases keep that detail must be checked at the start). These records name the contracted supplier, which is often a reseller or integrator rather than the hosting provider, so a hand-coded sample is needed to validate provider identification; interview access to buyers.
Feasibility
Master's thesis. Low: public data and interview time; optional commercial procurement datasets would speed up cleaning but are not required. Provider identification through resellers is the main data-quality problem; the waivers applied market-wide, so the before/after comparison is descriptive unless a credible comparison group (for example licensing-tied workloads) is shown to differ. Public buyers are not the whole market, and results should be described as public-sector switching. The free-exit programmes announced by Google (11 January 2024), AWS (5 March 2024) and Microsoft (13 March 2024) require approval and give credits for complete exits within a fixed window; the CMA found they do not cover ongoing multi-cloud use and, for AWS and Microsoft, not partial switches. The treatment is therefore weaker than "fees removed", and the waivers do not apply at all to adding a second provider.
Has this been done?
The search found no study that does the core design (first-pass screen of 12 October 2026, checked again on 13 October 2026).
No study measuring actual switching between cloud providers after the 2024 egress-fee waivers or the CMA investigation was found. Economic work on egress fees is theoretical (Biglaiser, Cremer, de Corniere and Mantovani); the empirical inertia study found (Jin) measures switching within one provider. Tussell (2022) describes public cloud award values but not switching. The CMA's final decision describes low uptake of free-exit programmes, and its March 2026 decisions added new Microsoft and AWS switching commitments and a licensing investigation, without measuring switching. Parliamentary records in 2026 indicate government does not track its cloud dependence. Between-provider switching in UK public procurement before and after the 2024 waivers and the 2026 commitments remains unmeasured.
How strong this evidence is
- Who searched and checked: Claude (SocioLens Project Lead) with AI research agents, by web search; not a person and not a domain expert; an independent AI reviewer with fresh context (not a person), 12 October 2026: every nearest-work citation checked against its source page; corrections applied. Every check so far was done by AI; no person or domain expert has checked it yet.
- What was read: 1 of the 8 nearest works below was judged from an abstract, summary or search listing only; it is marked. Not every entry says how much of the work was read.
- Where the search could not reach: Screened by web search on the date shown; Google Scholar, SSRN, paywalled journals and the CMA's own working papers were not searched exhaustively, and the full Jin paper was not available. Absence of a hit is not proof that the work does not exist. Literature check of 2026-10-13: Read the CMA March 2026 decision only through press coverage, not the CMA's own documents. Did not access Tussell or other paid procurement analytics. No Contracts Finder data were inspected. Journal searches returned older lock-in literature only. A TechMarketView review of G-Cloud spending (10 Sep 2026) is paywalled and was not read.
- Before building a project on it, run your own search in a scholarly database such as Scopus, Web of Science, Google Scholar or SSRN.
Nearest work found
- Biglaiser, Crémer, de Cornière & Mantovani, "Should Egress Fees Be Eliminated? An Analysis of Cloud Services and Beyond," CRESSE conference paper (Jan 2025)
A two-period duopoly model of when banning egress fees helps customers; theoretical, no data on actual switching. - Biglaiser, Crémer & Mantovani, "The Economics of the Cloud," TSE working paper (Mar 2024)
Review of cloud economics stating that the authors know of only one paper (Jin, Peng and Wang, 2023) that measures switching costs in the cloud; does not measure switching itself. - Jin, "Sticky Consumers and Cloud Welfare," seminar paper, ACSS-PSL Digital Regulation seminar (Feb–Mar 2026) Judged from an abstract, summary or search listing only
Uses proprietary customer-level data to estimate inertia when moving between products of the same provider, as a lower bound, and argues a fee ban alone would not end stickiness; does not observe between-provider switching or the 2024 waivers. Seminar abstract read only. - Lammin, "Public Sector Cloud Services Market: present and future trends," Tussell (9 Sep 2022)
Describes about £4.4bn of UK public cloud-related awards in 2016–2021 with AWS the largest supplier by value, and many contracts due for renewal in 2023–2025; no analysis of switching. - UK Competition and Markets Authority, "Cloud services market investigation, Appendix N: Egress fees and free switching programmes," final decision appendix (2025)
Sets out the terms of the Google, AWS and Microsoft free-exit programmes and finds they leave out ongoing multi-cloud use and some partial switches; describes uptake as low and treats that as inconclusive about whether egress fees are a barrier. Reports programme terms and uptake, not whether customers moved between providers. - The Register, "Microsoft to face CMA scrutiny over cloud software licensing," news article (31 Mar 2026; author not recorded)
News report of the CMA's 31 March 2026 decisions: no SMS designation of Microsoft and AWS in cloud for now; an SMS investigation into Microsoft's business software licensing (Windows Server, SQL Server pricing on rival clouds); Microsoft commitments extending free egress to 180 days and counting exit from an individual Azure service as a switch; a six-month CMA review. Reports no measured switching. Relevant as a second policy event (2026) after the 2024 waivers, and supports the Microsoft-stack comparison in the design. - Yeandel, "Parliament, seven years, no number" (part one of a series on UK cloud sovereignty in parliamentary records), yeandel.co.uk blog (27 Jul 2026)
Non-scholarly analysis of 2026 parliamentary debates; quotes an MP that dependence data on AWS and Microsoft is "not tracked across Government" and finds no switching measurement. Supports the gap; it is commentary, not research evidence. - Biglaiser, Cremer, de Corniere & Mantovani, "Should Egress Fees Be Eliminated? An Analysis of Cloud Services and Beyond" (forward search)
Forward search found no later empirical work citing or extending it; only the existing conference version and seminar listings appeared. Listed here only to record the forward search; already in nearest_work.
Literature check
Stands as first published. No study measuring between-provider cloud switching in UK public procurement, or anywhere after the 2024 egress-fee waivers, was found. New 2026 developments are policy events, not evidence: on 31 March 2026 the CMA accepted Microsoft and AWS commitments (including a 180-day free-egress window and partial-exit switching for Azure) and opened an SMS investigation into Microsoft licensing, and a July 2026 review of parliamentary records finds government does not track its dependence on AWS and Microsoft. The opportunity stands; the 2026 commitments add a second event the design can exploit.
Searched: web (extended); arxiv.org; ssrn.com; osf.io; nber.org; dl.acm.org; sciencedirect.com; academic.oup.com; journals.sagepub.com; tandfonline.com; link.springer.com; theregister.com; computerweekly.com. Checked by Claude (SocioLens Project Lead) with AI research agents, by web search; not a person and not a domain expert.
The screen before this check
The search found no study that does the core design (first-pass screen of 12 October 2026).
No study measuring actual switching between cloud providers after the 2024 egress-fee waivers or the CMA investigation was found. Economic work on egress fees is theoretical (Biglaiser, Crémer, de Cornière & Mantovani); the one empirical inertia study found (Jin) measures switching between products of the same provider using proprietary data, by design not between providers. A UK procurement analysis (Tussell, 2022) describes public cloud award values but not switching. The CMA's final decision describes uptake of the providers' free-exit programmes as low, but does not measure between-provider switching. Between-provider switching in public procurement before and after remedies remains unmeasured as far as this screen could find.
Search queries used
cloud switching egress fees removal effect empirical study customers switching multi-cloud after 2024 egress waiverpublic sector cloud procurement concentration hyperscalers analysis contract data G-Cloud OR TED OR Contracts Finder studyEU Data Act switching charges cloud empirical evaluation lock-in measurement firmsJin Peng Wang 2023 switching costs cloud computing IaaS user inertia estimationegress waiver uptake cloud switching evidence after 2024 study customers switched providers measured
Literature check, 13 October 2026:
G-Cloud public sector hyperscaler spend AWS Microsoft lock-in analysis 2025 2026 contract renewals switchingcloud vendor lock-in switching costs empirical evidence public procurement (sciencedirect.com, tandfonline.com, journals.sagepub.com, academic.oup.com, cambridge.org, dl.acm.org, link.springer.com)cloud switching egress fees Microsoft licensing empirical (ssrn.com, arxiv.org, osf.io, nber.org)Tussell OR Open Contracting cloud contracts UK government AWS dependency report 2026 incumbencyCMA cloud market investigation remedies 2026 switching data customers multi-cloud evidence after free exit programmes"Should Egress Fees Be Eliminated" Biglaiser Cremer 2026
Provenance
RO-LD11-03 v3. Drafted by Claude (SocioLens Project Lead), AI-drafted from the LD-11 debate on 2026-10-12. Checked by an independent AI reviewer with fresh context (not a person), 12 October 2026: every nearest-work citation checked against its source page; corrections applied. Not yet reviewed by a domain expert. This is a research hypothesis, not a finding.
Changes
- v2, 13 October 2026: Literature check added: a second, deeper search for existing work, with its queries, new nearest work and limits. The question and verdict stand. See v1 as published.
- v3, 13 October 2026: Tightened after a test reading of the opportunity pages on 13 October 2026 (an AI stand-in run of the founder test, not an outside reader) found that the page did not show the procurement data exist in a usable form, the first thing a supervisor would ask. The data section now says what the open records contain (machine-readable notices from Contracts Finder and Find a Tender; G-Cloud sales with supplier and customer names, currency to be checked) and what they lack (the hosting provider behind a reseller), and the design now starts with a sample check of how often the provider can be identified, with freedom-of-information requests as the fallback. The question, the literature check and the verdict are unchanged. See v2 as published.