Government Cloud Procurement Competition Issues

1. Introduction

Government cloud procurement concerns the acquisition of cloud infrastructure, platform services, software, storage, cybersecurity, data-management and AI-computing capacity by public authorities. Unlike ordinary IT procurement, cloud procurement can create long-term dependence on a small number of hyperscale providers.

The competition-law problem therefore operates at two levels:

1. Procurement competition — whether the government tender is designed and conducted so that qualified suppliers can compete fairly; and
2. Market competition — whether the resulting contract, technical architecture or supplier relationship reinforces dominance, foreclosure, switching costs or exclusionary conduct.

The issue has become particularly important because cloud services increasingly constitute critical infrastructure. The UK's CMA cloud investigation, concluded in July 2025, found significant market-power concerns involving Amazon and Microsoft and identified issues including egress fees, interoperability barriers and software licensing.

2. Meaning of Government Cloud Procurement

Government cloud procurement includes contracts for:

• Infrastructure-as-a-Service (IaaS);
• Platform-as-a-Service (PaaS);
• Software-as-a-Service (SaaS);
• sovereign or government cloud;
• cloud storage;
• government databases;
• cybersecurity infrastructure;
• disaster recovery;
• AI and machine-learning compute;
• GPU/accelerator capacity;
• identity and access management;
• cloud migration;
• managed cloud services;
• data analytics;
• cloud-based public-service applications; and
• multi-cloud or hybrid-cloud arrangements.

A procurement authority may conduct:

Tender → qualification → technical evaluation → price evaluation → framework/contract → migration → long-term consumption → renewal.

The competition problem is that the procurement decision at the beginning may determine the competitive environment for many years afterward.

3. Principal Competition-Law Issues

A. Specification designed around one supplier

A government may unintentionally or deliberately formulate technical specifications that closely correspond to one incumbent's architecture.

Examples include:

• proprietary APIs;
• supplier-specific authentication;
• proprietary database formats;
• particular virtualization technology;
• proprietary AI accelerators;
• mandatory compatibility with an incumbent's software;
• requirements based on a particular certification architecture.

Such requirements can effectively produce a de facto single-source procurement.

The competition-law concern is not that a government can never prescribe demanding technical specifications. Rather, the question is whether the requirements are objectively justified, proportionate and genuinely connected to the procurement objective.

4. Incumbent Advantage

An incumbent cloud provider may possess:

• knowledge of government workloads;
• existing security accreditation;
• existing infrastructure;
• access to historical usage data;
• existing government software relationships;
• migration knowledge;
• established identity systems;
• existing enterprise agreements.

This creates a data-and-infrastructure advantage over potential entrants.

Consequently, even when a new tender is formally open, the incumbent may enjoy substantial practical advantages.

5. Bundling and Tying

A cloud supplier may bundle cloud infrastructure with:

• productivity software;
• identity management;
• cybersecurity;
• databases;
• analytics;
• AI models;
• developer tools;
• collaboration software.

For example, a government department might obtain a discounted cloud package only if it also purchases a supplier's proprietary software ecosystem.

This can raise competition concerns where the supplier possesses substantial market power in one product and uses that position to strengthen another market.

The issue is particularly important in the Microsoft-cloud context. The UK's 2026 regulatory programme expressly identified business-software licensing as an important remaining competition issue affecting cloud competition.

6. Data-Egress Fees and Switching Costs

One of the most important cloud competition problems is the cost of moving data away from a provider.

A government may technically have the right to change providers but face enormous practical costs because of:

• data-transfer charges;
• application redesign;
• API incompatibility;
• proprietary databases;
• retraining;
• security re-certification;
• migration downtime;
• contractual restrictions.

Therefore:

Formal freedom to switch ≠ economically realistic ability to switch.

The UK's CMA specifically investigated egress fees and technical barriers to switching and multi-cloud use.

7. Interoperability

Interoperability is critical to maintaining procurement competition.

A government should ideally be able to:

Cloud A → Cloud B → Cloud C

without rebuilding its entire digital infrastructure.

Restrictions on:

• APIs;
• identity systems;
• container orchestration;
• databases;
• encryption systems;
• networking;
• monitoring tools

may increase switching costs.

Interoperability therefore becomes a competition remedy as well as a technical requirement.

The CMA's cloud investigation identified interoperability restrictions as one of the features affecting competition.

8. Framework Agreements and Long-Term Contracts

Government cloud contracts may last several years and involve enormous aggregate spending.

Long contracts can generate:

Advantages

• predictable pricing;
• security investment;
• infrastructure stability;
• lower transaction costs.

Competition risks

• foreclosure of competing providers;
• inability to accommodate technological change;
• excessive minimum-spend commitments;
• automatic renewal;
• volume-based discounts that disadvantage smaller providers;
• reduced opportunities for new entrants.

Thus, contract duration must be balanced against technological and competitive change.

9. Volume Discounts and Committed-Spend Agreements

Large governments have significant bargaining power because they can aggregate demand.

However, volume discounts may have different competition effects.

A legitimate discount may reflect genuine economies of scale.

But an exclusionary arrangement may arise where:

"The government receives the largest discount only if almost all cloud expenditure is committed to one provider."

Such arrangements can make competing providers economically non-viable.

The CMA specifically examined committed-spend agreements in its cloud investigation.

10. Multi-Cloud Procurement

A competition-oriented government procurement strategy may permit:

• AWS + Azure;
• Azure + Google Cloud;
• private cloud + public cloud;
• sovereign cloud + commercial cloud.

Multi-cloud procurement can reduce:

• supplier dependence;
• switching costs;
• operational concentration;
• bargaining asymmetry.

However, excessive fragmentation can increase:

• cybersecurity complexity;
• administrative costs;
• integration costs;
• duplicated expenditure.

Therefore, competition law does not automatically require multi-cloud procurement. The question is whether the procurement structure preserves credible alternatives without compromising legitimate security and efficiency objectives.

11. Sovereignty and Security Requirements

Governments may legitimately require:

• domestic data residency;
• national-security controls;
• government-only facilities;
• security clearances;
• sovereign encryption;
• domestic personnel;
• restricted foreign ownership.

However, such requirements can substantially reduce the supplier pool.

The competition question becomes:

Is the restriction necessary and proportionate to the government's legitimate security objective?

A government should therefore distinguish between:

genuine security requirement

and

unnecessarily restrictive procurement criterion.

12. Small and Emerging Cloud Providers

Cloud procurement can produce a structural disadvantage for smaller providers because governments frequently demand:

• enormous capacity;
• global redundancy;
• 24/7 support;
• extensive certifications;
• high insurance limits;
• substantial financial guarantees;
• nationwide or international availability.

These requirements may be justified for critical infrastructure.

But if unnecessarily high thresholds exclude smaller suppliers, procurement can become structurally concentrated.

A competition-sensitive tender can instead use:

• lots;
• subcontracting;
• modular procurement;
• regional procurement;
• interoperability requirements;
• technology-neutral specifications.

13. Cloud Market Concentration

The cloud market is characterized by significant economies of scale.

Large providers can spread infrastructure costs over enormous customer bases.

This can produce a feedback loop:

Scale → lower cost → more customers → greater investment → better infrastructure → greater scale.

Government procurement can amplify this effect.

A very large government contract may therefore do more than satisfy government demand: it can strengthen the incumbent's position across the wider commercial market.

The CMA's 2025 investigation concluded that Amazon and Microsoft had significant market power in UK cloud infrastructure and recommended consideration of strategic-market-status investigations.

14. Government as a Strategic Buyer

Government is not merely an ordinary customer.

It can be one of the largest purchasers of:

• compute;
• storage;
• cybersecurity;
• AI infrastructure;
• cloud services.

Consequently, procurement rules can influence market structure.

A government can use procurement to encourage:

• open standards;
• portability;
• interoperability;
• competitive neutrality;
• domestic innovation;
• smaller suppliers;
• resilient supply chains.

The CMA's September 2026 work expressly identified cloud services and public procurement as areas where procurement can be used strategically to improve market outcomes.

15. Relevant Case Laws

There is not yet a large body of reported judicial decisions specifically concerning government procurement of hyperscale cloud services. Accordingly, the following cases combine cloud-specific procurement litigation with leading public-procurement and competition authorities that establish principles directly applicable to government cloud tenders.

Case 1 — Oracle America, Inc. v. United States

United States Court of Federal Claims / Federal Circuit — JEDI cloud procurement

This litigation concerned the US Department of Defense's Joint Enterprise Defense Infrastructure (JEDI) cloud procurement.

The dispute involved Oracle's challenge to the procurement process and alleged conflicts of interest.

Importance

The litigation demonstrates that cloud procurement can generate competition disputes involving:

• procurement neutrality;
• conflicts of interest;
• evaluation methodology;
• treatment of bidders;
• agency discretion;
• procurement integrity.

It is particularly important because the contract was not an ordinary IT purchase: it involved strategic military cloud infrastructure.

Competition principle

A government cloud tender must maintain procedural integrity and meaningful competitive opportunity, particularly where the procurement may create enormous market advantages for the successful supplier.

16. Case 2 — Oracle America, Inc. v. United States, 975 F.3d 1279 (Fed. Cir. 2020)

This Federal Circuit litigation arose from Oracle's challenge to the Department of Defense's JEDI procurement.

The Federal Circuit addressed issues concerning the procurement process and the Court of Federal Claims' handling of Oracle's objections.

Relevance to cloud competition

The case demonstrates that:

• government cloud procurement can involve exceptionally complex evaluation structures;
• technical requirements can have substantial competitive consequences;
• judicial review focuses on whether the procurement agency acted within lawful bounds;
• allegations concerning procurement irregularity can become central to cloud-market competition.

Principle

Strategic cloud procurement must satisfy procurement-law requirements even when the government invokes complex technological or national-security considerations.

17. Case 3 — Palantir USG, Inc. v. United States, 904 F.3d 980 (Fed. Cir. 2018)

The case concerned the US Army's procurement of a Distributed Common Ground System.

Palantir challenged the procurement methodology, arguing that the government had failed to properly consider commercially available alternatives.

Importance for cloud procurement

The case is particularly useful for modern cloud procurement because it supports the principle that government agencies should not automatically construct highly customized procurement solutions where commercially available technologies can satisfy the requirement.

This is relevant to:

• cloud platforms;
• commercial AI;
• cybersecurity;
• data platforms;
• cloud-native infrastructure.

Competition principle

Over-customization can create unnecessary barriers to entry.

A procurement authority should ask:

Can multiple commercial suppliers satisfy the requirement using interoperable technologies?

18. Case 4 — Concordia Bus Finland Oy Ab v. Helsingin Kaupunki, C-513/99

The CJEU examined the use of environmental criteria in public procurement.

The Court accepted that contracting authorities may consider criteria beyond the lowest price, provided that the criteria are connected with the subject matter and comply with procurement principles.

Cloud relevance

Government cloud procurement may legitimately consider:

• energy efficiency;
• carbon footprint;
• resilience;
• security;
• sustainability;
• data-centre efficiency.

But criteria must not be constructed in a way that arbitrarily favours one provider.

Principle

Non-price procurement criteria are legitimate, but they must remain objectively connected with the procurement subject and compatible with equal treatment.

19. Case 5 — Telaustria Verlags GmbH v. Telekom Austria AG, C-324/98

The CJEU established important principles concerning transparency and equal treatment in public procurement.

Cloud relevance

Government cloud procurement frequently involves:

• complex technical specifications;
• confidential security requirements;
• evaluation algorithms;
• supplier negotiations;
• framework agreements.

These circumstances make transparency particularly important.

A procurement authority should provide sufficiently clear information regarding:

• evaluation criteria;
• technical requirements;
• weighting;
• contractual requirements;
• supplier qualification.

Principle

Government purchasing power cannot be exercised through opaque procedures that distort equal competitive opportunities.

20. Case 6 — Pressetext Nachrichtenagentur GmbH v. Republik Österreich, C-454/06

This case concerned substantial modifications to a public contract.

The CJEU developed the principle that a sufficiently material modification to an existing public contract may require a new procurement procedure.

Cloud relevance

This is extremely important for long-term cloud contracts.

Suppose a government initially awards a cloud contract for:

£100 million of infrastructure services.

Several years later, the government substantially expands the arrangement to include:

• AI compute;
• government-wide identity services;
• cybersecurity;
• new departments;
• additional data centres;
• major additional expenditure.

If the modification materially changes the economic scope of the contract, the government cannot necessarily treat the expansion as an ordinary contractual amendment.

Principle

Material expansion of a cloud contract may create an obligation to reopen competition.

21. Case 7 — Manova A/S v. Region Syddanmark, C-336/12

The CJEU considered the treatment of missing information during a public procurement procedure.

The case is relevant to the principle that procurement authorities must balance:

• procedural flexibility;
• equal treatment;
• transparency;
• avoidance of unfair advantages.

Cloud relevance

Cloud tenders often involve complicated documentation concerning:

• certifications;
• security architecture;
• financial capacity;
• technical capability;
• subcontractors.

An authority may permit clarification in appropriate circumstances, but cannot use clarification to give one supplier an opportunity to substantially improve its competitive position after the deadline.

Principle

Clarification cannot become an opportunity for discriminatory renegotiation.

22. Case 8 — Wall AG v. Stadt Frankfurt am Main, C-91/08

The CJEU addressed the modification and transfer of public-service arrangements.

Cloud relevance

The case is useful for long-term government cloud arrangements because governments sometimes:

• transfer contracts;
• change providers;
• introduce subcontractors;
• restructure cloud service arrangements;
• expand the contractual scope.

Where such changes materially alter the competitive position created by the original tender, procurement principles may require renewed competition.

Principle

Post-award restructuring cannot be used to circumvent competitive procurement requirements.

23. How These Cases Apply to Cloud Procurement

24. Competition-Law Theories Potentially Engaged

Article 101 / Chapter I-type concerns

Potential concerns include:

• collusion among cloud providers;
• bid rigging;
• allocation of government contracts;
• exchange of sensitive bidding information;
• coordinated pricing.

Public procurement is particularly vulnerable to bid-rigging because tenders provide predictable purchasing opportunities.

Article 102 / Chapter II-type concerns

A dominant cloud provider could potentially face concerns involving:

• discriminatory access;
• tying;
• bundling;
• excessive switching costs;
• refusal to provide interoperability;
• exclusionary rebates;
• discriminatory licensing;
• exploitative contractual conditions.

The precise legal analysis depends upon dominance, market definition, conduct and effects.

25. Procurement-Law and Competition-Law Interaction

The two regimes should not be treated as identical.

Procurement law asks:

Was the government tender conducted fairly and lawfully?

Competition law asks:

Does the conduct or structure restrict competition in the relevant market?

Digital-market regulation asks:

Does a structurally powerful cloud provider require additional ex ante obligations?

Modern cloud markets increasingly require all three perspectives.

The EU is currently examining cloud services under the Digital Markets Act, including interoperability, financial conditions and contractual conditions; in June 2026 the Commission announced a preliminary view concerning AWS and Azure as potential cloud gatekeepers.

26. Government Procurement as a Competition Remedy

Government procurement itself can be designed to preserve competition.

A competition-sensitive cloud tender can require:

1. Data portability

Government data must be exportable in usable formats.

2. API interoperability

Providers must support open and documented interfaces.

3. Exit assistance

The supplier must assist migration at contract termination.

4. Transparent egress pricing

Data-transfer charges should be predictable and proportionate.

5. Multi-cloud compatibility

Where technically justified, systems should operate across multiple providers.

6. Modular contracting

Large contracts can be divided into technologically coherent lots.

7. Periodic re-tendering

Government should periodically test whether market conditions have changed.

8. Competitive neutrality

Technical requirements should not unnecessarily replicate one supplier's proprietary architecture.

27. Special Problem: AI + Government Cloud

The competition problem is becoming more complicated because cloud infrastructure increasingly includes AI.

A government may procure:

Cloud → GPUs → foundation models → AI applications → data → government services.

This can create vertical integration.

A cloud provider that controls both:

• computing infrastructure, and
• AI models

may have incentives or capabilities to favour its own AI ecosystem.

The CMA has expressly considered the relationship between foundation models and cloud competition.

Thus, future procurement rules may need to consider AI portability and model portability, not merely data portability.

28. Government Cloud Procurement and Market Foreclosure

The most significant structural concern can be expressed as:

Government demand

↓

Large cloud contract

↓

Higher provider scale

↓

Greater infrastructure investment

↓

More ecosystem integration

↓

Higher switching costs

↓

Reduced competitive pressure

↓

Further government dependence

This creates a potential procurement–market-power feedback loop.

Therefore, government procurement decisions can have effects extending far beyond the immediate contract.

29. Competition Risks in Different Procurement Models

Single-cloud procurement

Risk: dependency and foreclosure.

Multi-cloud procurement

Risk: complexity and higher coordination costs.

Sovereign cloud procurement

Risk: reduced supplier pool.

Framework procurement

Risk: incumbency and framework concentration.

Long-term committed-spend contracts

Risk: exclusionary volume commitments.

Modular procurement

Risk: fragmentation and integration costs.

Open-market procurement

Risk: security and assurance complexity.

The appropriate model depends upon the government's technical, security and operational requirements.

30. Recommended Competition-Law Checklist

Before awarding a major cloud contract, a public authority should ask:

1. Are the technical specifications supplier-neutral?
2. Can more than one credible provider satisfy them?
3. Are proprietary technologies genuinely necessary?
4. Can data be exported?
5. What are the egress costs?
6. Can applications operate across clouds?
7. Are APIs sufficiently interoperable?
8. Are volume discounts potentially exclusionary?
9. Does the contract create excessive minimum-spend obligations?
10. Is the contract excessively long?
11. Are renewal provisions competitive?
12. Can the government change providers?
13. Does the contract permit substantial future expansion without re-tendering?
14. Could bundling with other software foreclose competitors?
15. Are security requirements proportionate?
16. Can SMEs participate?
17. Is there a credible exit plan?
18. Could the procurement reinforce an already dominant provider?
19. Are AI and cloud services being bundled?
20. Will the procurement preserve competition over the entire contract lifecycle?

31. Key Legal Principle

The central principle can be summarized as:

Government cloud procurement should not merely obtain the lowest immediate price; it should preserve the conditions necessary for continued competition throughout the life of the digital infrastructure.

This is particularly important because the economics of cloud computing mean that switching costs, interoperability, data portability, licensing and ecosystem effects may matter more than the initial tender price.

The UK's current regulatory approach illustrates this shift: following its cloud market investigation, the CMA has continued work on egress fees, interoperability, cloud procurement and software licensing rather than treating the original procurement event as the end of the competition analysis.

32. Conclusion

Government cloud procurement sits at the intersection of public procurement law, competition law, digital-market regulation, cybersecurity and national infrastructure policy.

The principal risks are:

• supplier-specific specifications;
• incumbent advantage;
• bid rigging;
• bundling and tying;
• proprietary ecosystems;
• egress fees;
• interoperability restrictions;
• excessive switching costs;
• committed-spend arrangements;
• long-term foreclosure;
• discriminatory licensing;
• excessive contract modifications;
• concentration of government data and workloads; and
• reinforcement of existing cloud market power.

The cases of Oracle/JEDI, Palantir, Concordia Bus, Telaustria, Pressetext, Manova and Wall collectively demonstrate that competitive procurement requires more than a formally open tender. The procurement must also preserve equal treatment, transparency, proportionality, competitive neutrality and continuing contestability.

In modern cloud markets, therefore, the decisive question is not simply:

“Which cloud provider wins the government contract?”

It is also:

“Does the procurement architecture preserve the government's ability to choose, switch, negotiate and re-tender in the future?”