HM Revenue and Customs (HMRC), the UK’s tax authority, has embarked on an ambitious £500 million project to exit three Fujitsu-hosted datacentres and migrate its services to a hyperscale cloud environment over a 10-year period, with the deal set to last until 2036. This "Datacentre Exit" (DCE) programme aims to modernise HMRC’s IT infrastructure, shifting from legacy on-premises systems to a cloud platform managed by one of the world’s leading hyperscalers—think Amazon Web Services (AWS), Microsoft Azure, Google Cloud, IBM, or Oracle. The stakes are high: HMRC manages £500 billion in annual tax revenue, and any disruption could have significant economic repercussions.
But given HMRC’s track record with large-scale IT projects, can it realistically deliver this initiative on time, on budget, and within specification? Let’s break it down.
The Scope and Complexity of the Project
The DCE programme is not a simple lift-and-shift operation. HMRC’s current datacentres, operated by Fujitsu, host a sprawling array of systems running on diverse operating software, including P-Unix, IBM AIX, Sun Solaris, Red Hat Enterprise Linux, SUSE Linux Enterprise, Windows, VMware ESXi, Oracle Linux, and NetApp ONTAP, among others. Migrating these to a unified hyperscale cloud environment requires meticulous planning, compatibility assessments, and execution across hundreds of services. The chosen hyperscaler will need to provide a platform that not only supports this migration but also ensures resilience, security, and the flexibility to accommodate future business changes.
The 10-year timeline suggests a phased approach, which could mitigate some risks, but it also introduces long-term dependencies on the selected provider. HMRC has acknowledged that the hyperscaler may subcontract elements of the service, adding another layer of complexity to project management and accountability. The tax agency has also conducted an initial "R-treatment" assessment—evaluating what to rehost, retire, or retain—but this plan will evolve during procurement, indicating that the full scope is not yet locked in. This fluidity could be a strength, allowing adaptability, or a weakness, inviting scope creep and cost overruns.
HMRC’s Historical Performance: A Mixed Bag
To assess HMRC’s capability, we must first examine its history with large IT projects. The agency has a reputation for ambitious transformation efforts, but delivery has often been uneven.
The Aspire Contract (2004-2017): HMRC’s £10 billion IT outsourcing deal with Capgemini and Fujitsu, known as Aspire, was one of Europe’s largest civil IT projects. It successfully collected £500 billion annually but came under fire for its cost—£7.9 billion by 2014, with a 15.8% profit margin for suppliers—and lack of flexibility. The transition away from Aspire, intended to bring skills in-house and diversify suppliers, was deemed "highly risky" by insiders, with concerns over skill retention as key staff opted to stay with Capgemini and Fujitsu rather than transfer to HMRC. This suggests challenges in managing large-scale transitions and retaining expertise—both critical for the DCE programme.
Making Tax Digital (MTD): HMRC’s flagship digitalisation effort has faced delays and criticism. The National Audit Office noted that HMRC underestimated the scale of work and added complexity by attempting to replace legacy systems while introducing digital record-keeping simultaneously. This lack of sequencing contributed to timetable slips and increased costs, raising doubts about HMRC’s ability to manage overlapping IT priorities effectively.
Securing Our Technical Future: Part of this ongoing five-year programme involves moving 600 services from Fujitsu datacentres to cloud or Crown Hosting infrastructure by 2025. A £40 million deal with AWS is already in place, but Brexit-related pauses and shifting priorities have stalled progress, according to a 2022 report. The DCE programme builds on this effort, but its predecessor’s delays suggest that external factors and internal replanning could derail timelines.
Unity ERP Project: Launched in 2023, this £500 million effort to replace SAP ERP across HMRC and two other departments earned a "red" rating from the Infrastructure and Projects Authority, indicating that "successful delivery appears unachievable." This reflects ongoing struggles with large-scale IT procurement and integration.
These examples paint a picture of an organisation capable of managing complex systems but prone to underestimating challenges, losing control of costs, and slipping schedules. The DCE programme’s reliance on a single hyperscaler echoes past dependencies on monolithic suppliers like Fujitsu and Capgemini, which HMRC has struggled to unwind efficiently.
Key Factors Influencing Success
Time: The 10-year horizon provides breathing room, unlike the tighter deadlines of previous projects like MTD or the 2022 Fujitsu contract expiry. However, HMRC’s history of delays—exacerbated by external shocks like Brexit—suggests that even a decade may not guarantee timely completion. The agency’s ability to stick to a phased migration plan and avoid "significant replanning exercises" (as seen in 2022) will be critical.
Budget: At £500 million, the DCE deal is substantial but not astronomical compared to Aspire’s £10 billion price tag. Yet HMRC has a track record of cost overruns, partly due to poor negotiation (Aspire’s high profit margins) and evolving requirements (MTD’s complexity). The hyperscaler market’s dominance by AWS and Microsoft, which hold 80% of UK share, may limit HMRC’s leverage to secure competitive pricing. If the scope expands or subcontractors inflate costs, staying within budget could prove elusive.
Specification: Meeting the technical and security requirements of a hyperscale migration is within HMRC’s grasp, given its experience with AWS and the Customs Declaration Service rollout. However, the diverse legacy estate and evolving R-treatment plan introduce uncertainty. Past projects have faltered when specifications shifted midstream, and the DCE’s dependence on a single hyperscaler to "manage the migration and hosting" risks repeating this pattern if the provider struggles with HMRC’s unique needs.
Mitigating Factors and Potential Strengths
HMRC isn’t starting from scratch. It has bolstered its in-house capabilities, absorbing 565 skilled staff from its dissolved IT provider, RCDTS, in 2023, which could reduce reliance on external expertise. The agency’s existing £40 million AWS deal provides a foundation for hyperscale adoption, and its focus on resilience and security aligns with the DCE’s goals. The 10-year timeframe also allows for course corrections, unlike shorter-term projects that collapsed under pressure.
Moreover, the hyperscaler model offers scalability and expertise that on-premises systems lack. If HMRC selects a provider with a proven track record (e.g., AWS or Azure) and negotiates clear deliverables, it could offload some technical burden. The agency’s stated intent to validate the single-hyperscaler approach during procurement suggests a willingness to adapt, though this flexibility must be balanced against decisive execution.
The Verdict: Cautious Scepticism
HMRC is remotely capable of delivering the £500 million datacentre-exit deal, but significant hurdles remain. Its technical know-how and long-term vision are assets, but historical missteps—underestimating complexity, losing cost control, and succumbing to delays—cast doubt on a smooth outcome. The 10-year timeline offers a buffer, yet external disruptions (like Brexit) and internal replanning could erode it. Staying within budget hinges on strong procurement and scope management, areas where HMRC has faltered before. Meeting specifications is feasible but depends on locking down requirements early and managing the hyperscaler effectively.
Critically, HMRC must learn from past failures: sequence tasks logically, retain skilled staff, and resist over-reliance on a single supplier’s promises. Without these adjustments, the DCE risks joining the litany of government IT projects that promise transformation but deliver frustration. On balance, while success is possible, the odds lean toward a familiar story—late, over budget, and partially delivered. Taxpayers, as always, will be watching.
.jpg)