A moment of institutional silence often tells more than words. In late March, the UK's tax authority completed a procurement process that had been shaped by forces largely invisible to public scrutiny, and the result spoke volumes: no competition at all.
His Majesty's Revenue and Customs (HMRC) awarded Amazon Web Services a contract worth £472.8 million including VAT to migrate services from three ageing Fujitsu datacentres to the cloud. The deal, spanning seven to ten years, went to a single bidder. According to HMRC's own contract award notice published on 23 March, AWS was "the only remaining bidder" after a process that began with initial interest from at least five providers. By the time formal tenders closed, the field had narrowed to one.
What makes this noteworthy is not the outcome alone, but the machinery that produced it. According to reporting, HMRC unofficially shortlisted AWS, Google, and IBM in October 2025. Yet by the final stage, only AWS remained willing to bid. Google and IBM withdrew from the process. Microsoft, the only other cloud provider at AWS's scale, had never been shortlisted in the first place.
Insiders familiar with the bidding described a process weighted toward a predetermined outcome. HMRC anticipated that the appointment would be limited to a single Hyperscaler able to migrate services and infrastructure from Fujitsu-run datacentres that use about a dozen operating systems including HP's Unix, IBM's AIX, and Sun's Solaris to UK-based cloud hosting. The tender criteria, published in June 2025, dedicated 70 per cent of the contract decision to quality, with only 20 per cent on price and 10 per cent on social value. In other words, cost offered little leverage for negotiation.
One source told The Register that the tender had been structured so that "only AWS or Microsoft could realistically win it," owing to their "ability to execute" and "proven history of working with departments like this." When IBM decided not to bid further and Google followed suit, the government's negotiating position evaporated entirely.
The timing is uncomfortable. AWS is the only supplier left in the running for the contract also comes hot on the heels of the Treasury Select Committee querying HMRC's reliance on AWS's US-based infrastructure, after the agency's operations were disrupted on 20 October 2025 by Amazon suffering a 14-hour outage in its US-based datacentre region. Yet that disruption did not dissuade the authority from deepening its dependency on the same provider for the next decade.
This episode reflects a deeper structural problem in the UK's cloud services market. In the UK, AWS and Microsoft together account for a combined 70–80% of customer spend in the public cloud market. The UK's Competition and Markets Authority (CMA) concluded its cloud services investigation in July 2025 with findings that ought to trouble policymakers. With the two US hyperscalers accounting for up to 90 per cent of the country's cloud services market, the CMA inquiry in July advised designating the pair with strategic market status for cloud computing, meaning they'd face stricter rules as tech firms with "substantial and entrenched market power" and significant influence over specific digital activities.
The CMA identified barriers to competition that reinforce concentration: switching costs that lock customers into initial platform choices, licensing practices that disadvantage rivals, and technical hurdles that complicate multi-cloud strategies. UK customers are paying around £500 million more per year for these services than they would in more competitive markets.
Yet regulatory action remains stalled. A year after that finding, and almost two-and-a-half years after the investigation first began, the industry is still waiting for the regulator to take action. Meeks left in late January, the CMA's inquiry chair departing over the slow pace of reform. The CMA board is expected to consider this recommendation in early 2026 regarding strategic market status designation. In the meantime, the government continues to lock itself into long-term contracts with the very providers the regulator has found to be exercising monopolistic power.
HMRC defended the contract award, noting it followed government procurement rules and that it had achieved value for money. The authority added that it publishes contract details for transparency. Yet transparency after the fact does not remedy competitive failure during the procurement itself. When a government department responsible for collecting the nation's taxes enters a billion-pound cloud services arrangement as a sole-bidder process, the underlying questions persist: Did the tender structure inadvertently exclude viable alternatives? Were the technical requirements genuinely necessary, or did they simply favour the incumbent? And why did the CMA's warnings about market concentration not shape procurement policy?
These questions cut to the heart of fiscal responsibility and institutional accountability. The government cannot, on one hand, task a regulator with investigating anti-competitive conduct and, on the other hand, use procurement practices that entrench the market power the regulator has identified as harmful. The HMRC contract represents the cost of that contradiction in real terms: a guarantee of continued reliance on a single supplier in a sector where the UK government and public sector's dependence on US hyperscalers is consistent with their overall position in cloud services markets.
For observers concerned with both sound spending and competitive markets, the lesson is clear. Procurement reform must precede or accompany regulatory action. Government agencies cannot wait for the CMA to finish its deliberations if they continue to shape tenders in ways that pre-select vendors. The architecture of competition is built not just through regulation, but through the choices made at the point of purchase. When those choices narrow to one, the market has failed, regardless of what any contract notice says about fairness.