What Is Actually Negotiable in an Enterprise Cloud and SaaS Renewal in 2026

Fewer than 1% of cloud customers switch provider in any given year, according to the UK Competition and Markets Authority’s final decision on the cloud services market investigation, published on 31 July 2025. Enterprise buyers read that number as proof they have no leverage. It is closer to the opposite. Six things move in a large cloud or software renewal even when the platform itself does not: the commitment structure, the discount tier and its duration, exit and transition terms, licensing portability, data residency and assurance commitments, and price protection over the contract term. What rarely moves is the headline unit price of commodity compute or storage, which is the one thing most buyers spend their preparation time on.

The gap between those two lists is where enterprise money is lost. This article sets out what each of the six levers is worth, what vendors say about it, and what the current regulatory position in Europe and the UK does to the balance of power.

The market the numbers describe

UK customers spent £10.5 billion on cloud services in 2024, with the market growing at close to 30% a year since 2020. Microsoft and AWS each held between 30% and 40% of the infrastructure market; Google held between 5% and 10%. The CMA concluded that both leaders have generated sustained returns substantially above their cost of capital for years, and estimated that if prices sat on average 5% above competitive levels, customers would pay around £500 million a year more than they should.

The authority identified three adverse effects on competition: concentration with high entry barriers, technical and commercial barriers that lock customers in, and Microsoft licensing practices that make its software more expensive to run on rival clouds.

What happened next matters for buyers planning a renewal. On 31 March 2026 the CMA declined to designate Microsoft and AWS with Strategic Market Status, accepting voluntary commitments on cloud egress fees and interoperability instead and opening a separate investigation into Microsoft’s corporate software services. Those commitments cover egress fees and interoperability, and they are not binding. Computer Weekly’s account of the CMA antitrust investigation into AWS and Microsoft traces how three years of investigation ended without designation.

The full case file sits on the CMA’s cloud services market investigation page. The practical conclusion is unglamorous. Regulators have described the problem accurately and are not going to solve it inside your contract term. The correction happens at the negotiating table or it does not happen.

That makes the renewal a capability question rather than a procurement one, and it is worth establishing where a buying organisation stands before the first vendor meeting rather than after it. The Gap Partnership, a global negotiation consultancy that works with commercial and procurement teams on live deals as well as on capability, publishes a negotiation maturity assessment covering people, process and organisational factors. It is a reasonable way to test whether the discipline described in the rest of this article already exists internally.

How the six levers compare

The SaaS layer behaves differently

Cloud infrastructure and enterprise SaaS get bundled together in procurement conversations and behave nothing alike. Infrastructure pricing is broadly transparent and broadly commoditised. Application pricing is neither.

Zylo’s 2026 SaaS Management Index reports that 79% of IT leaders saw price increases at renewal in the previous twelve months, and 78% encountered unexpected charges tied to consumption or to newly added AI features. The same research puts average savings achieved at renewal at 16.8%, and the average enterprise portfolio at 305 applications. The methodology behind those figures is not published in detail and the source is a vendor with an interest in the answer, so treat them as directional. The direction is not in dispute: prices rose, the increases arrived through consumption and AI line items rather than through the headline rate, and buyers who negotiated recovered a double-digit percentage.

AI feature pricing is the newest and least disciplined part of the bill. Vendors are adding AI capability to existing products and charging for it either as a per-seat uplift or as consumption, often mid-term. A renewal negotiated in 2026 that does not define what happens when the vendor ships an AI feature into a module you already licence will produce an unbudgeted increase inside eighteen months.

Why switching cost is a position rather than a fact

The single most common statement in an enterprise renewal is that switching is impossible. It is usually said by the buyer, about themselves, in front of the vendor.

Switching cost is real. It is also composed of parts that behave differently. Data egress is a fee. Retraining is a project. Application refactoring is engineering effort. Contractual exit penalties are terms someone agreed to. Only the third of those is a genuine technical constraint, and even that is a number rather than a wall.

European law is now moving one of the components directly. Article 29 of the EU Data Act removes switching charges entirely from 12 January 2027, and has already restricted them to cost-justified amounts during the transitional period. The Debate’s analysis of how Europe’s switching rules are rewriting cloud contracts makes the commercially important point: a fee that disappears by operation of law on a known date is a concession with an expiry, and a buyer who pays for it in the meantime is paying for something they are about to be given. The same piece traces how the Commission’s Cloud and AI Development Act, published on 3 June 2026, is doing something similar to sovereignty and residency terms that are currently sold as premium tiers.

For a buyer, the sequence is straightforward. Establish what each component of the switching cost actually is, in a number. Identify which components are being reduced by regulation on a known timetable. Ask for those now, in writing, while the vendor still has something to give.

The telecoms and financial services variants

Two categories of buyer routinely conclude they have no position at all, and both are wrong for the same reason.

A telecoms operator part-way through a network deployment cannot credibly threaten to change vendor. The equipment is installed, the integration work is done and the deployment schedule is public. What the operator can still negotiate is everything that is not the decision to stay: the scope of future phases, spares and support pricing, software release commitments, the treatment of interoperability with other vendors’ equipment, and the price of the next tranche of capacity. The leverage moved from the choice of supplier to the terms of continuation, and the terms of continuation are worth more over a ten-year deployment than the original selection was.

A financial institution dependent on a primary market data vendor is in the same shape. It cannot switch. It can still negotiate what it pays for, which users are licensed, how redistribution and derived data are treated, how increases are capped across the term, and what resilience and exit-planning commitments the vendor provides. That last item is not a favour. A regulated institution has supervisory obligations around concentration risk and operational resilience, and a vendor that wants to keep a regulated account has to help the customer meet them. Obligation, used properly, is leverage.

When internal teams are enough, and when they are not

Most renewals should be run internally. Procurement and legal teams in large organisations negotiate constantly and are good at it.

Three conditions change that. The first is asymmetry of practice. Hyperscaler and enterprise software sales teams run hundreds of these negotiations a year against buyers who run one every three to five years, and the difference in repetitions is usually worth more to the vendor than any analytical advantage it holds. The second is value at stake, where a one percent movement on a nine-figure agreement exceeds the cost of any support several times over. The third, and the most common, is internal disagreement: when procurement, legal, compliance, the technology function and the business owner cannot agree a single position, the vendor negotiates with whichever of them is most anxious.

That third condition is what specialist negotiation support actually addresses in a complex technology deal. It is not legal advice and it is not benchmarking. It is getting one organisation to arrive with one position, one set of concessions in priority order, and one agreed point at which it stops.

The test is simple enough to run without anyone’s help. Ask the five internal parties, separately, what the organisation will not agree to. If the answers differ, the negotiation has already started and the vendor is winning it.

Frequently asked questions

What leverage does an enterprise buyer actually have in a SaaS renewal when switching costs are prohibitively high?

Considerably more than the switching cost implies, because leverage sits in the terms rather than in the threat to leave. Commitment length, discount duration, price protection caps, user and consumption definitions, AI feature pricing, service credits and exit assistance are all negotiable in a large agreement, and none of them requires a credible alternative supplier. In the European Union, switching charges themselves are being removed from 12 January 2027 under Article 29 of the Data Act, which turns a vendor’s exit fee into a concession with a known expiry date. A buyer who cannot leave can still refuse to sign a long commitment.

What is genuinely negotiable in an enterprise cloud services agreement with a major hyperscale provider?

Commitment structure and ramp, discount tiers and how long they survive, egress and transition terms, service credits, data residency and audit rights, and licensing portability all move regularly in large agreements. The CMA’s 2025 investigation found that the two largest providers have sustained returns substantially above their cost of capital, which tells a buyer there is margin in the relationship to negotiate against. What is genuinely hard to shift is the unit price of commodity compute and storage, where the provider is competing on published rates. Buyers who concentrate on unit price and accept the surrounding terms as standard give away most of the available value.

What leverage does a telecoms operator have in a vendor renegotiation when mid-deployment switching is not realistic?

Its leverage transfers from supplier selection to the terms of continuation. Future deployment phases, capacity tranches, spares and support pricing, software release and roadmap commitments, and interoperability obligations with other vendors’ equipment are all still open, and over a multi-year deployment they are worth more than the original selection decision. The operator’s position strengthens where the vendor’s own revenue forecast depends on the later phases, which it usually does. The error is treating the deployment commitment as though it settled every remaining commercial question.

When does a cloud or software negotiation need specialist external expertise rather than the internal team?

When three things are true together: the counterparty negotiates these agreements far more often than the buyer does, the value at stake means a small percentage movement outweighs the cost of support, and the internal parties cannot reach a single position without a neutral party in the room. External support is least useful where the organisation already has a repeatable preparation process and a genuine alternative supplier. It is most useful in regulated environments where legal, compliance, procurement and the business each hold a veto and none of them holds the commercial objective.

What does specialist negotiation support add to a complex technology contract negotiation in a regulated environment?

It supplies the two things internal teams structurally struggle to provide for themselves: an agreed single position across functions that each have a veto, and practice at holding that position when a well-drilled counterparty applies pressure. Internal legal teams are expert in risk and internal procurement is expert in price, but neither typically owns the sequencing of concessions or the decision about when to stop. In regulated sectors the additional contribution is translating supervisory obligations, on resilience, concentration risk and exit planning, into contract clauses the vendor has to meet, so that compliance requirements become negotiating currency rather than late-stage objections.