The government's first Cloud Challenge Book asks industry to fix public-sector legacy IT. Its own roadmap promised the data to measure that by this autumn, and still has not delivered it.
Estimated reading time: 6 minutes
In short. The Government Digital Service and the Government Commercial Agency published the first Cloud Challenge Book on 7 July 2026, inviting industry to help solve five national "challenges," the first of which is a public sector where "28% of our estate is legacy," progress that "has taken 13 years (so far)" [1], page 8. The government's own roadmap for modern digital government committed, in a change log entry dated 28 May 2026, to publishing "a clearer, fuller view of how and where legacy technology is being used today" by Autumn 2026 [4]. On 29 October 2026, with that season nearly two-thirds gone and no further update recorded, GDS added a Clarifications document to the Cloud Challenge Book that still cannot settle what counts as legacy, writing only that the definition "is broader than ageing or unsupported technology" [2], page 1. The same Clarifications ask suppliers to plan five-to-ten-year investments against a procurement that is itself only "expected to launch in early 2027, subject to... ongoing design and approvals" [2], page 3. Two documented mistakes: four stars.
The Cloud Challenge Book is modelled on the Ministry of Defence's Challenge Set programme: rather than specifying a solution, GDS and the Government Commercial Agency set out five national problems and invite suppliers to propose how to solve them [1], page 5. First published on 7 July 2026, the document sat unchanged until GDS added a three-page Clarifications document on 29 October 2026, responding to supplier feedback gathered "over summer 2026" [3] [2], page 1. Read together, the two documents test two of the government's own commitments: a data promise that falls due this season, and a request for multi-year industry investment made with almost nothing pinned down on the government's side. This site found a similar pattern in the roadmap's predecessor document last month: a stated target substituted with a smaller, friendlier number, and a successor plan that missed its own deadline [5].
The claims, tested
| The document's own words | What the evidence shows | Verdict |
|---|---|---|
| Challenge 1 asks industry to help achieve "comprehensive and automated visibility" of the public sector's legacy estate, because "28% of our estate is legacy" [1], pages 8 to 9 | The government's own roadmap committed to delivering "a clearer, fuller view of how and where legacy technology is being used today" by Autumn 2026, after already slipping the date once from June [4]. As of 29 October, that milestone is still listed under "planning to do next," not delivered, and no change log entry since 28 May records progress [4] | The data the roadmap promised to have in hand by now, which would make Challenge 1 answerable, has not been shown as delivered |
| Challenge 5 opens by quoting the Blueprint for modern digital government on using the public sector's scale "as the country's largest digital buyer" [1], page 21, and asks suppliers to plan around a "horizon of five to ten years" [2], page 3 | The same Clarifications state the anticipated procurement, the National Cloud Infrastructure Programme, is "expected to launch in early 2027, subject to the programme's ongoing design and approvals," and that "the term, structure and nature of any future contractual commitments have not yet been decided" [2], page 3 | Suppliers are asked for a decade-long planning horizon against a procurement whose launch date, and whose contract terms, government has not committed to |
The mistakes, counted
The legacy-visibility data promised by this autumn has not been shown as delivered (1). Challenge 1 of the Cloud Challenge Book frames "comprehensive and automated visibility" of the public sector's technology estate as something industry should help build, citing a legacy rate of 28% and funding that remains "insufficient for the size of the challenge" [1], page 8. But the government had already told itself, in the published roadmap that Challenge 1 footnotes, that better data was the precondition for addressing this: a change log entry dated 28 May 2026 explains that the original June 2026 date slipped to Autumn 2026 because "the team responsible for delivering on this new approach to modernising and fixing Legacy IT was not fully in place" at the time the roadmap launched [4]. Autumn, by the meteorological calendar the roadmap itself uses elsewhere, runs to the end of November; 29 October is roughly two-thirds through that window, and the roadmap page carries no update after May. Rather than publish the promised fuller picture, the same-day Clarifications instead broaden the definition of the problem: asked "what do we mean by legacy," GDS writes that it "is broader than ageing or unsupported technology," extending to any case where "public sector organisations lack the capability or capacity needed to maintain or improve a digital service," and concedes that "upskilling our existing workforce can help address this but is not sufficient to address the challenge at the scale required" [2], page 1. A widened definition is not the clearer measurement that was due.
Suppliers are asked to plan a decade ahead of a procurement government has not committed to (2). Challenge 5 positions government as uniquely able to use its scale as "the country's largest digital buyer" [1], page 21, and the Clarifications ask suppliers "to provide feedback on the basis of a horizon of five to ten years, and to set out how their response and the value they offer would differ across that range" [2], page 3. In the same answer, GDS states that the route most likely to turn any of this into a contract, the National Cloud Infrastructure Programme procurement, is merely "expected to launch in early 2027, subject to the programme's ongoing design and approvals," and that "the term, structure and nature of any future contractual commitments have not yet been decided" [2], page 3. The document explicitly calls its own horizon "a planning assumption for the purposes of engagement," not a commitment. Industry is being asked to do the long-range forecasting that the document says government itself has not yet done.
Credit where due
GDS did not have to publish the Clarifications at all; the original Book had no obligation to answer follow-up questions in public, and doing so, with dated, specific answers rather than vague reassurance, is more transparent than many consultations on this site's desk. Several of the answers are genuinely informative rather than defensive: the disclosure that the Department for Digital, Culture, Media and Sport "has formally acquired a mandate for SECRET Compute and AI infrastructure on behalf of government, Critical National Infrastructure, academia and industry" is new, specific information that was not in the original Book [2], page 2, and the admission that government has not yet decided its own contract terms is a franker answer than simply repeating the five-to-ten-year ask without qualification. The challenge-based format itself, naming problems rather than prescribing suppliers' answers for them, is also a reasonable response to the fragmented procurement Challenge 5 itself describes [1], page 21.
Verdict
Four stars, from two documented mistakes. Neither finding is about whether the underlying ambition, modernising legacy IT and buying cloud infrastructure at national scale, is sound; both plausibly are. What the document does not show is the one thing it needed to show by now: that the data-gathering it promised for this exact season has happened, and that the multi-year commitment it is asking industry to plan around is matched by an equivalent commitment on its own side.
The star score counts two documented mistakes. First, the government's own roadmap committed, in a change log dated 28 May 2026, to delivering fuller data on public-sector legacy technology use by Autumn 2026; as of 29 October 2026, around two-thirds through that season, the roadmap's own page still lists this as planned rather than delivered, and the same-day Clarifications on the Cloud Challenge Book broaden the definition of "legacy" rather than supply the promised data. Second, the Clarifications ask suppliers to plan investment on a five-to-ten-year horizon while stating that the anticipated procurement is only expected to launch in early 2027 "subject to... ongoing design and approvals" and that contract terms "have not yet been decided." Two falls in the 1 to 3 band: four stars; the bands are on the ratings page. This piece makes no finding about the conduct of any named individual; the Cloud Challenge Book, its Clarifications and the digital government roadmap are institutional documents setting out government's own stated plans. Checked directly against the Cloud Challenge Book, its Clarifications and the roadmap's own change log, all fetched 29 October 2026.
Sources
- Cloud Challenge Book 2026 (PDF), Government Digital Service and Government Commercial Agency, published 7 July 2026
- Clarifications on the Cloud Challenge Book 2026 (PDF), Government Digital Service, published 29 October 2026
- Cloud Challenge Book 2026, GOV.UK, published 7 July 2026, last updated 29 October 2026
- Understanding our digital systems to improve them, A roadmap for modern digital government, Government Digital Service, change log last updated 28 May 2026
- A government digital strategy set out to get 50 of its top 75 services to a "great" standard by 2025. A minister's own written answer puts the final score at 29, Tracked Changes, 17 September 2026