A decade after the UK Government set out to dismantle its reliance on a “tiny oligopoly” of major IT outsourcers, one senior digital leader says the public sector is facing a new era of dependency – this time centred on cloud hyperscalers and internal capability gaps rather than monolithic contracts.
Speaking at the recent Think AI in Government event, Chad Bond, Strategy and Innovation Director at Zaizi, said the reforms achieved their primary aim, but also created structural weaknesses now hampering government transformation and AI adoption.
“Government absolutely succeeded in breaking the hold of the old IT oligopoly,” Bond told Think Digital Partners. “But in doing so, we created a new set of dependencies – one on hyperscale cloud providers, and another on internal capability that was never built up to match the new model.”
Breaking Up the Outsourcers
The 2011 Cabinet Office ICT Strategy was one of the most consequential policy shifts in UK public sector technology. Convinced that multi-billion-pound systems integrator contracts had led to spiralling costs and poor delivery, ministers launched a three-part strategy:
- Open Standards to reduce reliance on proprietary software.
- Contract Size Limits with a “presumption against” IT deals over £100 million.
- Disaggregation of Large Expiring IT Contracts (LECs) into smaller components.
This was reinforced by a Cloud First policy, which Bond described as “a procurement decision as much as a technology one.”
Digital Marketplace: A Game-Changer for SMEs
One of the clearest successes of this reform wave was the creation of the Digital Marketplace and the G-Cloud framework. These platforms dramatically simplified procurement and opened government opportunities to smaller firms.
In 2023/24, SMEs secured 44 percent of G-Cloud spend – £1.32 billion, far above SME participation rates elsewhere in government.
Bond recalled the impact: “For SMEs like Zaizi, the Digital Marketplace was transformative. It broke down barriers and gave agile teams real access to government for the first time.”
The model was widely praised for weakening the grip of major suppliers and empowering a more diverse, competitive ecosystem.
But the success of disaggregation came with unintended consequences that are now increasingly visible.
1. The Internal Capability Crisis
With prime contractors removed, departments became their own systems integrators, responsible for coordinating dozens of suppliers and maintaining architectural coherence.
Bond warned that this shift revealed a deep talent deficit: “We solved the supplier problem, but in doing so we created a huge internal capability problem. Government didn’t build the technical or commercial capacity needed to run a complex multi-supplier ecosystem.”
Engineering, architecture and commercial skills remain among the most acute shortages across the civil service.
If you liked this content…
2. The Cloud Oligopoly
The move to cloud hosting also concentrated spending among a small number of hyperscale providers. The public sector now invests more than £1 billion annually in cloud services.
“What we’ve done is trade one form of lock-in for another,” Bond said. “Integrator lock-in became platform lock-in. It’s harder to see, but just as risky.”
Departments report growing difficulty moving workloads or negotiating competitively, raising concerns about long-term flexibility.
Moving Beyond the Tetris Mindset
Bond argued that the government needs a new mental model for thinking about legacy systems. The old “Tetris” metaphor of technical debt piling up until the board fills no longer captures the complexity and interdependence of the estate.
He suggested the game Factorio as a more accurate analogy. “Government legacy isn’t about blocks stacking up – it’s an entire production chain,” Bond explained. “You’ve got spaghetti factories, main buses, and pollution that attracts attackers. And the whole thing has to keep running while you refactor it.”
NCA Shows the Power of an SME-Driven AI Model
At the Think AI event, the National Crime Agency offered a case study in the value of SME-led delivery. Moving away from large consulting firms it found lacking in AI expertise, the NCA rebuilt its supply chain to focus on smaller, specialist partners.
The results were dramatic: development cycles that once took 8–16 months now take two to three weeks.
According to Bond, the NCA example shows what’s possible when capability and supplier agility align. “It proves the model works – when departments have the skills to manage it.”
The Next Chapter: Capability, Cloud Leverage and Cross-Government Platforms
Bond outlined five priorities for the next wave of transformation:
- Adopt systems-thinking to map and address cross-department legacy interdependencies.
- Centrally fund “pollution control”, creating secure APIs and data layers to shield new services from toxic legacy.
- Re-factor core platforms, prioritising digital identity, authoritative registers and payments.
- Invest aggressively in internal expertise, especially engineering and commercial roles.
- Leverage collective buying power to negotiate with hyperscalers as a single customer.
He emphasised that the SME ecosystem is now firmly established – but the government must modernise the rules around it. “The Marketplace solved the last decade’s problems,” he said. “Now it needs refactoring for the AI economy.”
A related briefing, Smarter Partnerships: Why SMEs Are Key to Government’s AI Success, explores how the new Procurement Act could accelerate this shift.
