Who Controls AI Platforms and Why It Matters

Introduction

The promise of an AI-enabled human bloom depends not only on how powerful AI becomes, but on who controls it. If advanced AI helps accelerate medicine, science, education, engineering and economic production, then access to AI systems may become as important as access to electricity, communications networks or financial infrastructure. A future of abundant intelligence could expand human flourishing on a vast scale. But if the key platforms remain concentrated in a handful of companies, many of the gains may be shaped by private decisions about pricing, access, standards and acceptable use rather than by democratic institutions.

This tension is already visible. A small number of firms dominate the cloud infrastructure, specialised chips and foundation model ecosystems on which advanced AI depends. Governments and competition authorities increasingly view this not as a normal software market, but as a strategic infrastructure question. The debate is not simply whether concentration slows competition. It is whether a technology that could influence scientific discovery, healthcare, education, labour markets and public administration should be governed primarily through a small set of commercial gatekeepers.[OECD]oecd.orgcomponent 5AI supply chain: Competition in artificial intelligence…14 Nov 2025 — 4 The most recently reported market share estimates for the…[GOV.UK]GOV.UKcloud services market investigationservices market investigation28 January 2025: The CMA has published its provisional decision in its market investigation into the supply…Published: January 2025

Why AI tends towards concentration

Many digital technologies become concentrated because scale creates advantages. AI may intensify this pattern.

The largest frontier models require enormous quantities of computing power, specialised chips, engineering talent, training data and electricity. Building and operating these systems demands billions of pounds in capital expenditure. The firms best positioned to make those investments are typically large cloud providers and technology companies with existing global infrastructure.[GOV.UK]GOV.UKai foundation models initial reviewFoundation Models: initial review4 May 2023 — This initial review will help create an early understanding of the market for foundation mo…Published: May 2023

Several reinforcing mechanisms push power towards a relatively small group of actors:

• Compute concentration: Training and serving advanced models requires vast data centres and access to scarce AI accelerators.
• Capital concentration: Frontier development often requires funding levels unavailable to most universities, startups or public institutions.
• Data advantages: Large firms often possess extensive proprietary datasets and customer relationships.
• Distribution advantages: Companies that already control operating systems, productivity software, cloud platforms or app ecosystems can rapidly deploy AI products to hundreds of millions of users.
• Feedback loops: Larger user bases generate more revenue, data and product feedback, which can improve future systems.

The result is not necessarily a monopoly, but a market structure in which a few firms possess disproportionate influence over the direction of AI development and deployment.

Market concentration in cloud infrastructure

Cloud computing has become the physical backbone of modern AI. Even organisations building their own models often depend on rented infrastructure from major cloud providers.

OECD analysis notes that cloud markets are highly concentrated, with the largest providers collectively holding the majority of global market share. Multiple studies place Amazon Web Services (AWS), Microsoft Azure and Google Cloud as the dominant providers worldwide, together accounting for more than 60% of global cloud infrastructure spending.[OECD]oecd.orgcomponent 6Global share of largest provider reported over 80%. Global share of largest 3 players reported as having over 60% share…Read more…[OECD]oecd.orgpartnerships involving generative AI companies and cloud providers.Read morePotential competition policy responses in AI infrastructureNov 14, 2025 — The US Federal Trade Commission in January 2025 published s…Published: January 2025

This matters because cloud platforms are no longer merely storage and computing businesses. They increasingly function as AI platforms, offering:

A university, hospital trust, pharmaceutical company or government department may therefore depend on the same providers for both general computing infrastructure and advanced AI services.

Competition authorities worry that this creates bottlenecks. If a small number of firms become unavoidable intermediaries for advanced AI, they gain substantial influence over pricing, technical standards and market access. The UK’s Competition and Markets Authority (CMA) has repeatedly highlighted the strategic importance of cloud infrastructure to AI competition and innovation.[GOV.UK]GOV.UKai foundation models initial reportFoundation Models: Initial report18 Sept 2023 — The CMA started its initial review into AI Foundation Models in May 2023, to help create…Published: May 2023[GOV.UK]GOV.UKai foundation models update paperFoundation Models: Update paper11 Apr 2024 — Update paper as part of the CMA's AI Foundation Models: initial review, following initial re…

Vertical integration and the rise of AI ecosystems

The concentration question is not only about market share. It is also about vertical integration.

Many of the largest AI firms operate across multiple layers of the stack at once:

LayerExamples of controlInfrastructureData centres, cloud platforms, networkingComputeAI chips, chip design, procurement relationshipsModelsFoundation models and AI assistantsApplicationsWorkplace software, coding tools, search, consumer productsDistributionEnterprise contracts, operating systems, app ecosystems

A company that controls several layers simultaneously can create efficiencies. Customers often benefit from integrated systems that work together smoothly.

However, vertical integration can also increase dependency. A business that trains models on one cloud platform, stores data there, integrates AI into workplace software from the same provider and relies on its security infrastructure may find switching increasingly difficult.

The concern is not merely that large firms become successful. It is that the costs of leaving their ecosystems become so high that competitors struggle to emerge even if they build technically superior products.

This is one reason regulators have examined partnerships between major cloud providers and leading AI developers. The CMA, FTC and other authorities have investigated whether investments and commercial agreements could strengthen existing infrastructure dominance or create barriers for rivals.[Federal Trade Commission]ftc.govFederal Trade CommissionFTC Issues Staff Report on AI Partnerships & Investments…The report details key aspects regarding the structur…[Federal Trade Commission]ftc.govFederal Trade CommissionFTC Issues Staff Report on AI Partnerships & Investments…The report details key aspects regarding the structur…[CIO Dive]ciodive.comregulators, who have taken a sweeping aim at AI startup deals.Read moreCIO DiveUK regulators add Amazon, Anthropic partnership to…Aug 8, 2024 — The probe into Anthropic and Amazon's partnership is the late…

Why regulators worry about AI partnerships

Many of the most important AI companies are connected through investment, cloud supply agreements and long-term infrastructure contracts.

Microsoft’s relationship with OpenAI became the most prominent example, but similar arrangements have emerged elsewhere, including Amazon’s partnership with Anthropic and Google’s investments in frontier AI firms.[GOV.UK]GOV.UKto GOV.UKServices and information, benefits, includes eligibility, appeals, tax credits and Universal Credit, births, deaths, marriages a…

Supporters of these partnerships argue that they solve a genuine financing problem. Frontier AI development requires infrastructure that startups could not otherwise afford. Large cloud providers can supply capital, chips and engineering support that accelerate innovation.

Critics argue that the same arrangements may entrench existing power structures.

In 2025, the US Federal Trade Commission published findings from its inquiry into AI partnerships and investments. The report identified several areas of concern, including control over critical inputs such as compute resources, increased switching costs for AI developers, and access by cloud providers to commercially sensitive information. The FTC also explored whether some partnership structures could provide strategic influence without triggering traditional merger review processes.[Federal Trade Commission]ftc.govFederal Trade CommissionFTC Issues Staff Report on AI Partnerships & Investments…The report details key aspects regarding the structur…[Federal Trade Commission]ftc.govFederal Trade CommissionFTC Issues Staff Report on AI Partnerships & Investments…The report details key aspects regarding the structur…[Federal Trade Commission]ftc.govFederal Trade CommissionFTC Issues Staff Report on AI Partnerships & Investments…The report details key aspects regarding the structur…

The broader governance question is whether frontier AI remains an open field with multiple competing centres of innovation, or whether it gradually consolidates around a few interconnected infrastructure providers.

What concentration could mean for an AI-enabled future

For advocates of AI abundance, concentration presents a paradox.

On one hand, large firms may be essential for building the infrastructure needed to support transformative AI. Massive investments in chips, data centres, energy systems and global deployment networks are often beyond the reach of governments, universities and smaller companies.

On the other hand, a future in which intelligence becomes abundant may look very different depending on ownership structures.

Consider several possibilities:

The optimistic AI bloom vision usually assumes broad diffusion of intelligence. Commercial concentration raises the possibility that intelligence becomes abundant in technical terms while remaining institutionally controlled.

The argument that concentration may not last

The concentration story is not uncontested.

Some economists and technology analysts argue that AI markets remain highly dynamic. Today’s leaders face competition from open-source models, specialised AI firms, new cloud providers and emerging infrastructure companies.

Recent years have already seen significant challenges to incumbents. Open-weight models have reduced some barriers to experimentation. Specialised AI cloud providers have emerged. Large firms themselves compete intensely against one another rather than functioning as a single bloc.[Business Insider]businessinsider.comTraditionally, AWS was a go-to platform for startups due to its scalable compute and storage offerings. However, the rise of generative A…

There are also historical reasons for caution before assuming permanent dominance. Technology markets often appear locked up until new technical shifts create opportunities for challengers. Mainframe computing, personal computing, web browsers, social media and smartphones all experienced periods of dramatic change in competitive leadership.

The strongest governance arguments therefore do not necessarily assume inevitable monopoly. Instead, they focus on preserving contestability: ensuring that future competitors have a realistic chance to emerge.

Competition and interoperability policies

The policy debate increasingly centres on how to encourage innovation without allowing infrastructure power to become self-reinforcing.

Several approaches have attracted attention.

Interoperability and portability

One goal is making it easier for organisations to move between providers.

Policies can include:

Advocates argue that lower switching costs strengthen competition while preserving incentives to innovate.

Scrutiny of partnerships and acquisitions

Competition authorities increasingly examine investments, exclusive agreements and hiring arrangements involving major AI companies.

The concern is not only traditional mergers but also forms of influence that may shape market structure without formal ownership changes. Recent CMA and FTC investigations have focused heavily on this issue.[Federal Trade Commission]ftc.govFederal Trade CommissionFTC Issues Staff Report on AI Partnerships & Investments…The report details key aspects regarding the structur…[Federal Trade Commission]ftc.govFederal Trade CommissionFTC Issues Staff Report on AI Partnerships & Investments…The report details key aspects regarding the structur…[Federal Trade Commission]ftc.govFederal Trade CommissionFTC Issues Staff Report on AI Partnerships & Investments…The report details key aspects regarding the structur…

Public and academic compute

Some researchers argue that governments should support shared computing infrastructure for universities, public-interest research and smaller firms.

The logic resembles earlier investments in scientific infrastructure. If access to advanced computing becomes a prerequisite for participation in frontier research, public alternatives may help prevent capability from becoming exclusively commercial.

Open models and open standards

Open-source and open-weight AI systems are often presented as a counterbalance to concentration.

The advantages include:

However, open systems also create governance challenges involving misuse, safety and accountability. The trade-off is not straightforward.

The deeper question: who governs intelligence?

The long-term significance of AI concentration goes beyond economics.

Previous industrial revolutions changed access to energy, manufacturing capacity and transportation. Advanced AI may change access to intelligence itself: the ability to generate knowledge, solve problems, design technologies and coordinate complex systems.

If that happens, control over AI platforms could become a form of structural power comparable to control over financial systems, communications networks or energy infrastructure.

This does not mean private firms should be excluded from building advanced AI. Many of the breakthroughs that could contribute to a flourishing future may emerge from commercial research laboratories. But the governance challenge becomes more serious as AI capabilities become more central to medicine, science, education and public decision-making.

The central question is therefore not whether successful AI companies should exist. It is whether societies allow the institutions that shape intelligence production to become so concentrated that public choices increasingly depend on private platform governance. A world in which AI helps humanity bloom may require not only more intelligence, but more pluralism in who can access, develop and direct it.[OECD]one.oecd.orgmore…[GOV.UK]GOV.UKcloud services market investigationservices market investigation28 January 2025: The CMA has published its provisional decision in its market investigation into the supply…Published: January 2025[OECD]oecd.orgcomponent 5AI supply chain: Competition in artificial intelligence…14 Nov 2025 — 4 The most recently reported market share estimates for the…

Amazon book picks

Further Reading

Books and field guides related to Who Controls AI Platforms and Why It Matters. Use these as the next step if you want deeper reading beyond the article.

Book

Prediction Machines

By Ajay Agrawal, Joshua Gans et al.

Explains the economics of AI platforms and adoption.

See on Amazon

Book

Atlas of AI

By Kate Crawford

Explores infrastructure, power and ownership in AI systems.

See on Amazon

Book

The Coming Wave

By Mustafa Suleyman

Directly addresses concentration and control of advanced technologies.

See on Amazon

Book

The Master Switch

By Tim Wu

Provides historical context on concentration in network industries.

See on Amazon

Browse more on Amazon:Prediction MachinesAtlas of AIThe Coming Wave

As an Amazon Associate I earn from qualifying purchases.

eBay marketplace picks

Marketplace Samples

Example marketplace items related to this page. Use the search link to explore similar finds on eBay.

Shop location

UsingUSA USA USAUKCanadaAustraliaIreland

Example eBay listing

Miniature rack for palm-sized network devices Nitto Kogyo Server Rack New Japan

SearcheBay.co.uk: server rack model

Browse similar oneBay.co.uk

Example eBay listing

Dell PowerEdge Server Rack-mount Model with 2x 300gb Sas

SearcheBay.co.uk: server rack model

Browse similar oneBay.co.uk

Example eBay listing

Dell PowerEdge R730xd 16LFF+2SFF Server /IndigoVision NVR-AS 4000/32GB RAM [16D]

SearcheBay.co.uk: server rack model

Browse similar oneBay.co.uk

Example eBay listing

Rack Server w/ 4x TESLA K10 GPU ACCELERATOR *12288 Cores* PhotoScan 3D Modelling

SearcheBay.co.uk: server rack model

Browse similar oneBay.co.uk

Browse more oneBay.co.uk

Example items shown for inspiration; availability and pricing can change. Branchoria may earn a commission if you purchase through outbound eBay links.

Endnotes

1. Source: oecd.org
Title: component 5
Link:https://www.oecd.org/en/publications/competition-in-artificial-intelligence-infrastructure_623d1874-en/full-report/component-5.html

2. Source: assets.publishing.service.gov.uk
Title: UKA I Foundation Models
Link:https://assets.publishing.service.gov.uk/media/661e5a4c7469198185bd3d62/AI_Foundation_Models_technical_update_report.pdf

Published: September 2023

3. Source: ftc.gov
Link:https://www.ftc.gov/news-events/news/press-releases/2025/01/ftc-issues-staff-report-ai-partnerships-investments-study

4. Source: assets.publishing.service.gov.uk
Title: Full report
Link:https://assets.publishing.service.gov.uk/media/65081d3aa41cc300145612c0/Full_report_.pdf

5. Source: oecd.org
Title: component 6
Link:https://www.oecd.org/en/publications/competition-in-artificial-intelligence-infrastructure_623d1874-en/full-report/component-6.html

6. Source: GOV.UK
Title: cloud services market investigation
Link:https://www.gov.uk/cma-cases/cloud-services-market-investigation

Published: January 2025

7. Source: GOV.UK
Title: ai foundation models initial review
Link:https://www.gov.uk/cma-cases/ai-foundation-models-initial-review

Published: May 2023

8. Source: ftc.gov
Title: launches inquiry generative ai investments partnerships
Link:https://www.ftc.gov/news-events/news/press-releases/2024/01/ftc-launches-inquiry-generative-ai-investments-partnerships

9. Source: ftc.gov
Title: staff report ai partnerships investments 6b study
Link:https://www.ftc.gov/reports/ftc-staff-report-ai-partnerships-investments-6b-study

Published: January 2025

10. Source: ftc.gov
Link:https://www.ftc.gov/system/files/ftc_gov/pdf/p246201_aipartnerships6breport_redacted_0.pdf

11. Source: ftc.gov
Title: behind ftcs 6b report large ai partnerships investments
Link:https://www.ftc.gov/policy/advocacy-research/tech-at-ftc/2025/01/behind-ftcs-6b-report-large-ai-partnerships-investments

12. Source: oecd.org
Title: partnerships involving generative AI companies and cloud providers.Read more
Link:https://www.oecd.org/en/publications/competition-in-artificial-intelligence-infrastructure_623d1874-en/full-report/component-7.html

Published: January 2025

13. Source: one.oecd.org
Link:https://one.oecd.org/document/DAF/COMP%282025%298/en/pdf

14. Source: GOV.UK
Title: ai foundation models initial report
Link:https://www.gov.uk/government/publications/ai-foundation-models-initial-report

Published: May 2023

15. Source: GOV.UK
Title: ai foundation models update paper
Link:https://www.gov.uk/government/publications/ai-foundation-models-update-paper

16. Source: GOV.UK
Link:https://www.gov.uk/

17. Source: assets.publishing.service.gov.uk
Link:https://assets.publishing.service.gov.uk/media/661941a6c1d297c6ad1dfeed/Update_Paper__1_.pdf

Published: September 2023

18. Source: blog.promise.legal
Title: ftc report on cloud providers and ai partnerships
Link:https://blog.promise.legal/ftc-report-on-cloud-providers-and-ai-partnerships/

19. Source: ciodive.com
Title: regulators, who have taken a sweeping aim at AI startup deals.Read more
Link:https://www.ciodive.com/news/uk-regulatory-scrutiny-anthropic-amazon-competition-market/723777/

20. Source: businessinsider.com
Link:https://www.businessinsider.com/amazon-ai-startups-delaying-aws-spending

21. Source: ciodive.com
Title: ftc AI cloud inquiry
Link:https://www.ciodive.com/news/ftc-AI-cloud-inquiry/705671/

22. Source: awards.concurrences.com
Link:https://awards.concurrences.com/IMG/pdf/bradley_weber_taylor_levesque_cloud_and_competition_policy-_part_viii_-_cloud_service-ai_partnerships-_the_ftc_s_section_6_b_report_and_its_antitrust_implications_in_the_trump_2.0_administration_1_august_2025_concurrences_no_8-2025.pdf?149193%2Fcc0306ddb4ee1f75142844f9ece7723833c7aa2cb5a11f9042244118d9b420e7=

Additional References

23. Source: linkedin.com
Link:https://www.linkedin.com/posts/oecd-competition-law-and-policy_cloud-cloudcomputing-competition-activity-7331691640044060672-HjNY

24. Source: visitbritain.com
Link:https://www.visitbritain.com/en

25. Source: nortonrosefulbright.com
Link:https://www.nortonrosefulbright.com/en-gb/knowledge/publications/8ca8f277/the-uk-cmas-ai-foundation-models-initial-report

26. Source: nortonrosefulbright.com
Link:https://www.nortonrosefulbright.com/en/knowledge/publications/8ca8f277/the-uk-cmas-ai-foundation-models-initial-report

27. Source: aifray.com
Link:https://aifray.com/ftc-releases-staff-report-on-ai-partnerships-investments-collection-of-information-and-innuendo-with-unclear-contours/

28. Source: linkedin.com
Link:https://www.linkedin.com/posts/israeldaniel_competition-in-artificial-intelligence-infrastructure-activity-7416848936986128386-goye

29. Source: clearygottlieb.com
Link:https://www.clearygottlieb.com/-/media/files/alert-memos-2023/cma-publishes-initial-report-on-ai-foundation-models-and-guiding-principles-for-firms.pdf

30. Source: euronews.com
Title: uk regulator says microsoft and amazons cloud dominance hurts competition
Link:https://www.euronews.com/next/2025/07/31/uk-regulator-says-microsoft-and-amazons-cloud-dominance-hurts-competition

31. Source: businessatoecd.org
Title: Artificial Intelligence, Data, and Competition
Link:https://www.businessatoecd.org/hubfs/Artificial%20Intelligence%2C%20Data%2C%20and%20Competition.pdf?hsLang=en

32. Source: inquisitiveminds.bristows.com
Title: google anthropic the latest chapter in the cmas scrutiny of generative ai
Link:https://inquisitiveminds.bristows.com/post/102jnrs/google-anthropic-the-latest-chapter-in-the-cmas-scrutiny-of-generative-ai