The Speed of the Shift: Which UK MSPs Will Survive the AI Acceleration?

What the AI shift actually does to the MSP model

Traditional managed services built their economics on a simple foundation: IT is complex, expensive to staff, and organisations would rather pay a monthly fee than manage it themselves. That fee covered the genuine cost of expertise, the tooling to deliver it at scale, and a margin that compensated for the capital and operational risk the MSP carried.

AI changes the cost basis of expertise delivery faster than any previous technology shift. Tasks that required a skilled engineer - diagnosing infrastructure faults, writing runbooks, interpreting monitoring alerts, planning capacity, drafting change requests - are increasingly automated, AI-assisted, or entirely AI-handled. An MSP that needed twenty engineers to deliver a given service level now needs twelve. One that needed twelve now needs six.

For well-run, low-debt MSPs with modern platforms, this is a margin expansion event. For the rest of the market - businesses carrying significant debt loads from acquisition-funded growth, running legacy tooling that cannot absorb AI-native workflows, with cultures and workforces designed for the old model - it is an existential pressure arriving faster than they can restructure.

iomart: what a genuine collapse looks like

iomart was once a credible UK cloud and hosting business. Listed on AIM, with a portfolio of data centres, a managed hosting customer base built over two decades, and revenues that looked stable from the outside. What the outside view missed was the structural erosion happening underneath: customers migrating workloads to hyperscaler infrastructure, the legacy hosting revenue base shrinking faster than new managed services revenue could replace it, and a cost structure that reflected the old model rather than the new one.

The result was a business that found itself trapped. The revenue lines that funded the cost base were declining. The investment required to build competitive cloud-native managed services capabilities would have required capital the business could not service alongside its existing debt. The data centre assets that once looked like valuable infrastructure increasingly looked like liabilities - requiring capital to maintain and operate, generating declining returns as hosting economics deteriorated.

iomart's trajectory is the most visible illustration of what happens when a UK MSP built on legacy hosting economics meets a market that has moved on. It is not a story about bad management. It is a story about a business model that was viable in 2010, under pressure by 2018, and no longer sustainable by the mid-2020s.

Pulsant and Node4: running hard to stay still

Pulsant and Node4 represent a different variant of the same structural problem: the capex treadmill. Both businesses have invested heavily in UK data centre infrastructure - physical assets that require continuous capital investment to remain competitive, and that generate returns tied to utilisation and customer density.

Data centre infrastructure is not inherently a bad business. At the right scale, with the right customer base and the right capital structure, it generates predictable cash flows and genuine value. The problem for mid-market UK data centre operators is that the hyperscalers have permanently altered the competitive landscape. AWS, Azure and GCP invest tens of billions per year in infrastructure. They achieve energy efficiency, hardware performance and network density that a UK regional operator simply cannot match at any reasonable capital cost.

The response from businesses like Pulsant and Node4 has been to layer managed services on top of the infrastructure - to extract more value per rack, per customer, per square foot. This is rational. It is also a strategy that requires those managed services to be genuinely competitive: modern tooling, AI-native delivery, engineering talent that can operate at the level clients increasingly expect. Building that capability while simultaneously servicing the capex requirements of data centre operations is a difficult combination to fund.

Node4, private equity-backed and acquisitive, is running the consolidation playbook: buy complementary capabilities, extract synergies, build scale. Whether that playbook works depends on whether the acquired capabilities are actually modern, and whether integration happens fast enough to compete with providers who never had the legacy to integrate in the first place.

The squeezed middle: Redcentric, ANS, Nasstar, Claranet

The most interesting part of the UK MSP market is the tier just below the largest players - businesses with revenues in the £50m-£300m range, recognisable brand names, established customer bases, and varying degrees of genuine modernisation.

Redcentric has spent years repositioning from a connectivity-led business towards managed cloud and security services. The challenge is that repositioning takes time and investment, and the market does not wait. A business that is genuinely modern in its delivery today would have been building that capability for the last five years. Redcentric has been building, but the question for any business in this position is always whether the rate of change matches the rate at which the market is moving.

ANS has made genuine investments in cloud-native delivery and Microsoft partnership depth. It has built real capabilities in Azure and Microsoft 365 managed services that some of the larger incumbents struggle to match. The risk for ANS is whether its Microsoft-centric positioning is broad enough as the market diversifies - clients who were all-in on Azure are increasingly running multi-cloud or hybrid architectures, and a deep Microsoft specialist needs to credibly cover the rest of the landscape.

Nasstar occupies an interesting position: a managed services provider with strong vertical depth in legal and professional services, built on a foundation of hosted desktop and application services. The hosted desktop model is under pressure from Microsoft 365 and cloud workspace alternatives, and Nasstar's ability to evolve that customer base towards modern managed services is the central strategic question for the business.

Claranet is probably the most interesting of this group. A pan-European managed services provider with genuine technical depth, strong DevOps and cloud engineering capability, and a culture that attracts engineers rather than just account managers. Claranet's challenge is one of narrative clarity: a business with capabilities spread across security, cloud, networking and managed services can struggle to articulate a clear proposition in a market where buyers have limited attention and are looking for a clear answer to a specific problem.

What unites this tier is a combination of genuine capability and genuine constraint. They are not failed businesses. They are businesses that built for a market that looked different five years ago and are now investing to catch up with a market that has moved faster than anyone predicted. Some will make it. Some will find the combination of debt, legacy infrastructure, workforce costs and competitive pressure too much to overcome while simultaneously funding the transformation they need.

SysGroup: the broken promise of the acquisitive model

SysGroup made a clear strategic bet: build a mid-market MSP through acquisition, targeting the SME and mid-market segment with a broad managed services and connectivity proposition. The pitch to investors and clients was consolidation delivering scale benefits without the impersonal service of the largest providers.

The challenge that SysGroup and businesses like it face is that the acquisition-funded growth model only works if the underlying business being built is genuinely better than the sum of its parts. If each acquisition brings its own legacy tooling, its own service delivery approach, its own culture and its own client expectations, integration is slow and expensive - and the scale benefits materialise later, or not at all.

For clients, the acquisitive MSP model has a consistent failure pattern: the business that sold them on the relationship-based service proposition gets acquired into something larger, the relationship managers change, the pricing changes, the service standards change, and the client finds themselves in a managed service from a business they did not choose. The acquisition playbook optimises for AUM (assets under management, in this case client contracts) rather than for client outcome. Eventually clients notice.

The AI acceleration makes this worse. Acquisitive MSPs need to not only integrate their acquisitions but simultaneously modernise their delivery to remain competitive. Doing both at once, with private equity capital structures that require returns on a defined timeline, is a genuinely difficult ask.

What a different model looks like

Node Digital is not a comparison made lightly or without acknowledgement of our own limitations. We are not the largest provider in this market, and scale matters for some categories of client requirement. But the structural characteristics of how we operate are meaningfully different from the businesses described above.

We do not run data centres. We are not on the capex treadmill that requires every new client win to fund infrastructure that is already committed. When a client needs infrastructure, we deploy on their cloud accounts, or on our own lean platform, or on hardware they own. The economics are radically different.

We do not grow by acquisition. Our capability is built rather than bought, which means it is integrated, coherent and actually modern. When we say we deliver cloud-native managed services, we mean that the tooling, the automation and the engineering are cloud-native - not that we have bolted a cloud-native label onto a managed hosting business.

AI, for us, is a delivery accelerator rather than an existential threat. The tasks that AI automates in managed services - monitoring interpretation, runbook execution, change planning, documentation, first-line triage - are tasks we were already building automation around. An engineering team that was already working with infrastructure-as-code, GitOps delivery and automated observability integrates AI into those workflows naturally. A team running legacy tooling and manual processes has a much harder time.

The AI acceleration will not eliminate managed services. Organisations will always need a partner to operate complex technology environments, navigate vendor relationships, and take accountability for service outcomes. What it will eliminate is the managed services model that charged for complexity and delivered it. The businesses that survive and grow are those that charge for outcomes and deliver them - efficiently, transparently, and with the engineering depth to actually know what they are doing.

The UK MSP market has around five years of significant disruption ahead. Some of the names above will consolidate further. Some will find private equity exits before the structural pressure becomes fully visible in their numbers. Some will genuinely modernise and compete. And some will follow iomart.

We intend to be competing effectively throughout.