Institute for Public Policy Research
BOTTLENECK BRITAIN HOW DIGITAL COMPETITION COULD POWER UK GROWTH AND SOVEREIGNTY
Roa Powell August 2026
ABOUT IPPR IPPR, the Institute for Public Policy Research, is an independent charity working towards a fairer, greener, and more prosperous society. We are researchers, communicators, and policy experts creating tangible progressive change, and turning bold ideas into common sense realities. Working across the UK, IPPR, IPPR North, and IPPR Scotland are deeply connected to the people of our nations and regions, and the issues our communities face. We have helped shape national conversations and progressive policy change for more than 30 years. From making the early case for the minimum wage and tackling regional inequality, to proposing a windfall tax on energy companies, IPPR’s research and policy work has put forward practical solutions for the crises facing society. IPPR 4th floor, 8 Storey's Gate London SW1P 3AY E: info@ippr.org www.ippr.org Registered charity no: 800065 (England and Wales), SC046557 (Scotland) This paper was first published in August 2026. © IPPR 2026 The contents and opinions expressed in this paper are those of the authors only.
The progressive policy think tank
CONTENTS
Summary...........................................................................................................................5 1. Introduction.................................................................................................................8 2. The UK’s big tech rules have not lived up to their promise ...........................9 Political backing has wavered.................................................................................9 Enforcement has won narrow victories..............................................................10 AI is leaving us behind............................................................................................ 11 3. UK businesses are paying the price ...................................................................13 4. Growth and sovereignty: Weighing the trade-offs..........................................15 Good growth and inward investment.................................................................. 15 The UK’s sovereignty and geopolitical reality...................................................18 5. Recommendations...................................................................................................22 Issue a new strategic steer for the CMA..............................................................22 Prioritise the remedies that boost growth.........................................................22 Introduce reporting requirements that sustain the CMA’s mandate...........23 Align broader tech policy priorities to ensure impact....................................23 Maintain a proportionate enforcement framework ........................................24 References.....................................................................................................................25 Appendix: Methods..................................................................................................... 28
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
3
ABOUT THE AUTHORS
Roa Powell is a senior research fellow at IPPR.
ACKNOWLEDGEMENTS
We would like to thank Tom Smith, Philip Marsden, Diane Coyle, George Dibb, Taylor Gese and Max von Thun for taking the time to review and comment on this work and we are grateful to everyone who participated in an anonymous interview to support this project. We are grateful to James Crouch and Elliot Richley for their work on the polling for this report. We are grateful to officials at the Competition and Markets Authority who engaged with this research. And we would like to thank Marcus Bokkerink, Kip Meek and Mike Walker for sharing reflections on their time working on DMCCA implementation within the CMA. This work was kindly funded by the News Media Association and Reset Tech, and we are grateful for input from Dylan Sparks and Theo Bamber. We are especially grateful for Sebastian Cuttill from the News Media Association for his support and expertise throughout this project. We would like to thank Rowena Mayhew and Richard Maclean for their design and editorial support. And Carsten Jung, Laura Chappell and Pranesh Narayan for their advice on the research.
Download This document is available to download as a free PDF and in other formats at: https://www.ippr.org/articles/bottleneck-britain Citation If you are using this document in your own writing, our preferred citation is: Powell R (2026) Bottleneck Britain: How digital competition could power UK growth and sovereignty, IPPR. https://www.ippr.org/articles/bottleneck-britain Permission to share This document is published under a creative commons licence: Attribution-NonCommercial-NoDerivs 2.0 UK http://creativecommons.org/licenses/by-nc-nd/2.0/uk/ For commercial use, please contact info@ippr.org
4
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
SUMMARY Artificial intelligence (AI) could be the UK’s biggest economic opportunity in a generation, but only if its gains are spread across the economy rather than pooled within a handful of companies. The UK has one of the world’s most proportionate competition regimes to help make this happen. It is designed to open up markets, not slow technology down. But political will behind it has collapsed, leaving UK businesses to pay the price. A reset is long overdue.
MARKET CONCENTRATION HARMS UK BUSINESS
A handful of companies hold extraordinary control over the digital infrastructure we all rely on: • Google handles more than 90 per cent of UK internet searches. • Microsoft and Amazon Web Services (AWS) each control 30–40 per cent of UK customer spend in the cloud, with Google as the third provider. • Microsoft business software is now under investigation for its market power. • Apple and Google have a tight grip over all UK app distribution. These companies have earned much of their success and deserve to profit. But their success should not come at the expense of consumers paying higher prices or challenger firms being unable to scale their business. A new online survey – conducted by the market research agency Opinium and designed by IPPR – of more than 400 UK businesses, screened to exclude those with no digital dependency,1 shows that 79 per cent are concerned about dominant platforms limiting competition. These businesses rank technology giants’ market power as the second biggest constraint on their growth, above access to finance and talent. Half of these businesses would cease to function within days if their main technology provider cut access. In addition, 62 per cent say it would be costly or disruptive to switch to a different provider. If competition remedies were introduced, 72 per cent say they would lower their prices or produce better products. Meanwhile, more than a third report that, in the past two years, they decided not to offer a product because of an incumbent technology giant.
WE HAVE THE RIGHT LAW – WE JUST NEED TO USE IT
The Digital Markets, Competition and Consumers Act 2024 (DMCCA) is a flexible, robust and proportionate tool for dealing with this. The last Conservative government enacted it to improve competition and growth. It lets the regulator – the Competition and Markets Authority (CMA) – move fast on incumbents who abuse their market power, while leaving most businesses alone. The problem, then, is not the law but senior politicians’ lack of support for it. 1
This survey included businesses that significantly rely on technology platforms / digital infrastructure for revenue, reach, operations, distribution, software, compute or data storage. Full details of the survey methodology can be found in the appendix.
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
5
Over the past two years, senior officials have resigned or been sacked from the CMA, and in interviews they describe a collapse in ministerial support for their work. In 2025, the government explicitly asked the regulator to prioritise inward investment (DBT 2025), which was widely interpreted as a call to go easy on ‘big tech’. And subsequently, multiple investigations have led to voluntary commitments rather than binding rules, despite evidence of harm. The CMA continues to be one of the UK’s most expert regulators, curbing some of the worst behaviour of technology giants while managing a hostile geopolitical climate and industry lobbying. But it could move faster if it had politicians’ clear backing.
COMPETITION IS CENTRAL TO GROWTH AND SOVEREIGNTY
A narrative has taken hold that paints all digital regulation as the enemy of technological progress. But while some regulation could be improved to foster innovation, this blanket argument does not apply to competition legislation. Quite the opposite: competition policy spurs innovation and most businesses experience it as an enabler, not as a blocker. More open, competitive markets push businesses to innovate and drive growth across every postcode. Open markets also help with sovereignty. There is little point investing £500 million in a sovereign AI fund to build national champions while leaving untouched a market structure that will see them acquired or outcompeted. The trade-offs do remain a challenge. Investment in the UK from technology giants is important and the UK should try to avoid diplomatic retaliation from the US where possible. But the UK can be smart about enforcement such that the CMA moves faster without being perceived as an activist regulator that spooks investors. Meanwhile, international coordination can strengthen governments’ hand in resisting corporate pressure while red tape is removed elsewhere, so the UK becomes a more attractive capital investment prospect overall. Where investment risks remain, the government must take the trade-off seriously and make the big choices. Proportionality is the right test to avoid over-enforcement, but the government cannot outsource that judgement to the incumbents it is supposed to regulate.
BOLD ENFORCEMENT COULD UNLOCK GROWTH AND SOVEREIGNTY
We recommend the UK government publishes an updated strategic steer for the CMA. This should clarify that proactive, rapid and bold enforcement is pro-growth. The priority should be reducing barriers to entry for competitors, not maximising inward investment from incumbents. Greater emphasis is also needed on interventions that genuinely boost growth. So far, the CMA has made life fairer for dependants (for example, publishers relying on Google or app developers relying on Apple and Google), but it has done little to create conditions where competitors thrive, even though these conditions would deliver growth and sovereignty. In practice, this means considering the following: • Search. Keep Google’s designation under close review, with a view to bringing Gemini within scope, especially as AI transforms the search landscape. • Cloud. Review the decision to accept voluntary commitments from the major cloud providers instead of designating them with strategic market status. 6
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
• •
Economies of scale mean cloud will never be a highly competitive market, but even one more player could be enough to bring prices down. App stores. Structural change in near-term market shares is unlikely, so focus on conduct requirements that protect dependants. Capping commissions and direct developer-to-consumer app downloads would help. AI. The market structure is not yet settled, and the CMA’s potential impact is huge, especially where incumbents use existing dominance to gain ground in AI.
Finally, complementary industrial policy is essential. The CMA could do everything right and we would still be left without UK competitors if we do not have the right industrial policy. Combining competition enforcement with bottom-up support for UK champions, including via initiatives such as sovereign AI and reforms to public procurement, is essential.
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
7
1. INTRODUCTION Only a handful of companies own the infrastructure on which much of the UK economy depends. To give just a few examples: • Google accounts for more than 90 per cent of UK general internet search queries (CMA 2026a). • Microsoft and Amazon Web Services (AWS) each have a share of 30–40 per cent of UK customer spend in cloud, with Google as the third provider (CMA 2025a). • Microsoft software is now under investigation for its market power (CMA 2026b). • Apple and Google have a tight grip over all UK app distribution (CMA 2021). The harms this causes are well documented and include artificially high prices, degraded product quality, low investment and slowed innovation (Furman et al 2019). We also know that traditional competition law is too slow to deal with digital markets. By the time an investigation concludes, the harm is often entrenched and hard to reverse (Furman et al 2019). But the UK is actually ahead on this, with the Digital Markets, Competition and Consumers Act (DMCCA) passed in 2024. This law is more flexible and proportionate than its EU counterpart. It lets the regulator design bespoke remedies for every market (for example, they could tackle high fees in cloud while addressing fair attribution in search), ensuring that it can address the problem without subjecting companies to irrelevant requirements. The regulator can update its rules to reflect rapid technological change, meaning it can easily tackle AI. And it is time-bound, transparent and built around industry collaboration. The DMCCA gives the Competition and Markets Authority (CMA) three main tools. The CMA can: • designate a company with entrenched market power as having ‘strategic market status’ (SMS) • police those firms’ behaviour via bespoke ‘conduct requirements’ (CRs) • introduce ‘pro-competitive interventions’ (PCIs) to address the root causes of market power and reshape markets.2 This might sound dry and procedural. Competition authorities tend not to intervene and the remedies they introduce – such as ‘interoperability’, ‘fair ranking’ or ‘steering’ – can look like technical fixes, far away from the big measures politicians point to when speaking about growth. But the DMCCA powers add up to help address exactly the problems the UK government has staked itself on. As this report will explore, digital competition can contribute to growth, sovereignty, the cost of living and democracy. And in today’s geopolitical climate, a technocratic approach based on economic evidence is far more appropriate than any moralistic response.
2
8
Beyond these measures, the DMCCA sets out rules on merger reporting and the enforcement powers of the CMA, for example in introducing penalties. This detail is beyond the scope of this report.
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
2. THE UK’S BIG TECH RULES HAVE NOT LIVED UP TO THEIR PROMISE POLITICAL BACKING HAS WAVERED
A good law and an expert regulator are necessary but not sufficient to address the harms that concentrated markets cause. Standing up to the largest companies in the world also requires sustained political backing. For a time, that backing existed. The DMCCA 2024 passed with cross-party support and the CMA was ready to move fast. As the CMA’s chair at the time told us: “The CMA was ready to implement the new law quickly and effectively. Its Digital Markets Unit had been operating for years, its capabilities were market-leading, and the teams had already completed extensive market studies and investigations into search and search advertising, mobile platforms and cloud services. The business and innovationexperienced CMA board of the time made clear its intent to tackle all three areas within the first year, launching SMS [strategic market status] investigations into Google and Apple in search and mobile platforms immediately, to be followed by Amazon and Microsoft in cloud services.”3 Marcus Bokkerink, former chair of the CMA
Yet, since then, political backing for the Act has wavered, raising concerns at the top of the CMA, as reflected in the quotes below: “The Starmer government leadership of the time placed serving and former executives and advisors of the so-called ‘big tech’ firms in positions of influence across industrial strategy and business and investment policy, took over the CMA board, and gave it a revised steer, precisely to slow enforcement of the digital markets competition regime. There seemed to be a belief at the top of government that our economy could only be saved by investment from a handful of global technology firms – and that their investment would come to a halt if they had to operate in open, competitive markets where customers are free to choose and innovative challengers are free to compete.” Marcus Bokkerink, former chair of the CMA
“Obviously there was a threat to independence and a message received loud and clear in the CMA that government was involved. It was clear that the US tech companies were a constituency that the CMA had to treat with kid gloves.” Kip Meek, former chair of the CMA cloud inquiry
3
The quotes in this report are from semi-structured interviews conducted with academics, industry experts and officials. They represent the views of key officials working within the CMA during the early phases of the implementation of the DMCCA and these interviewees provided permission for their names to be used for the quotes. More details on the interview methodology are available in the appendix.
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
9
“Implementation has been very slow. We have got to some remedies in Google Search, which are perfectly sensible, but they aren’t moving the dial. There has been no attempt to actually get to the heart of market power.” Mike Walker, former chief economic advisor to the CMA
The pattern these accounts describe is the same: a law passed by parliament that had become a political inconvenience. Due to the DMCCA’s flexibility, it is highly vulnerable to shifting priorities. The CMA has significant discretion to set its own agenda, rather than following a fixed EUstyle rulebook. But this means the CMA’s pace is set as much by political appetite as by law. Without sustained backing, flexibility becomes a lever that industry can lean on to slow the regulator down. The CMA was already facing a stark asymmetry in power, dealing with the best-resourced companies in the world, but if you add in a government actively taking incumbents’ side, it becomes almost impossible for the CMA to move fast. This shift showed publicly too. In May 2025, the Starmer government published a new strategic steer for the CMA and advised it to focus on growth. Specifically, it called on the CMA to enhance the “attractiveness of the UK as a destination for international investment” (DBT 2025). This signal has been widely interpreted as an instruction to tread lightly on big tech (Pratley 2025, Ring and Pickard 2025). Now, more than a year later, there may be a chance for the political will behind the DMCCA to return, but the UK needs to rebuild momentum. Again, in the words of the CMA’s former chair: “There are some chinks of light emerging. UK competition law is still in place and businesses and consumers are challenging technology platform incumbents’ anti-competitive conduct directly at the Competition Appeal Tribunal. The UK’s ‘middle power’ allies in Europe, the G7 and beyond are all continuing their public authorities’ enforcement of competition law in digital markets. And more parts of UK government are beginning to see that unrivalled market power in the country’s digital infrastructure and services leads to lower (not higher) productivity, more economic dependency, and weakened security. It is encouraging that as a result of these and other pressures, the CMA is starting to consider whether Microsoft’s business software ecosystem could be opened up to greater customer choice and competition.” Marcus Bokkerink, former chair of the CMA
ENFORCEMENT HAS WON NARROW VICTORIES
Fortunately, shifting political tides have not stopped the CMA from making some progress. The first conduct requirements, rules that explicitly police the largest tech companies’ behaviour, have arrived. Publishers are now allowed to opt out of having their content appear in Google AI Overviews and Google is required to attribute content fairly (CMA 2026c). Additionally, there has been movement on Apple and Google’s app store duopoly, with the CMA consulting on conduct requirements that allow app developers to avoid fees of up to 30 per cent for in-app purchases (CMA 2026d). And, on Microsoft business software, a new investigation has been announced, signalling greater ambition for the future (CMA 2026b).
10
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
This is progress, but it is narrow. A relatively small set of businesses that depend on big tech for survival have hopes of being treated more fairly, but the measures set out above do little to make the wider digital economy more competitive. No pro-competitive measures have been floated – a power explicitly granted to the CMA to enable it to tackle the root causes of market power. And there has not been sufficient focus on structural remedies, choice screens, interoperability or data access. There are signs the CMA has held back. There has been a continued reliance on voluntary commitments over statutory tools. Most notably, there were no binding commitments or strategic market status (SMS) designations introduced in cloud markets. Instead, the CMA accepted voluntary commitments on egress fees and interoperability (Kirkwood 2026, Scanlan 2026). And several meaningful interventions have been delayed or ruled out, including any requirement for Google to pay publishers for content and any inclusion of Gemini within Google’s SMS status. In addition, Amazon Marketplace, YouTube and Meta have received no attention under the SMS regime. And, many stages of tiered consultation, which can take years, continue to hold enforcement back. Some caution is, of course, defensible as the regulator builds consensus and navigates a new regime. But the pattern here is enough to suggest a problem.
AI IS LEAVING US BEHIND
On top of this, digital markets have transformed since the DMCCA was passed. So far, the CMA has largely been working through low-hanging fruit, cases where harm is easiest to prove. But with digital markets moving so quickly, this leaves the regulator struggling to catch up with the markets where the potential for impact is greatest. If AI becomes a monopoly – or more likely an oligopoly – the UK’s leverage will be tiny and its financial returns constrained. If AI instead becomes a commodity, more of the productivity windfall will flow to workers, and UK businesses will reap the rewards. The AI infrastructure layer is already highly concentrated: • NVIDIA dominates the graphic processing unit (GPU) market. • The three ‘hyperscalers’ (AWS, Microsoft and Google) account for a large portion of compute access. • ASML is the sole manufacturer of extreme ultraviolet (EUV) lithography machines. • The Taiwan Semiconductor Manufacturing Company (TSMC) manufactures the vast majority of advanced chips (Widder and Kim 2025, Hacker 2026, Mitchell 2026). The foundation model layer is structurally prone to concentration. Training costs for frontier AI have grown 2.4 times a year since 2016 and on current trends will exceed a billion dollars by 2027 (Cottier et al 2024). Furthermore, vertical integration with incumbent companies could further entrench concentration. Microsoft, Google and Amazon have collectively invested more than US$20 billion in the leading AI developers, causing the CMA in the UK and the Federal Trade Commission in the US to flag these partnerships as a risk (CMA 2023, Office of Technology Staff 2025). And, as offerings such as Google AI Mode and the free version of Microsoft Copilot come integrated with the technology so many of us already use, incumbents have an advantage in winning the AI race.
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
11
But an alternative market structure is still possible. Open-source AI and smaller AI models are just a few months behind. Users are turning to cheaper AI products to suit their needs. The application layer remains competitive. And, where the technology is changing fastest, for example in domains such as hardware design, new competitors can succeed. However, the window for structural intervention is narrow. Whenever a new technology emerges, incumbents are likely to squash the innovators. And despite early and expert commentary from the CMA on the state of AI markets, the current conduct requirements do not speak to the urgency of this issue.
12
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
3. UK BUSINESSES ARE PAYING THE PRICE The concentration of large technology giants within digital markets is not just a problem for the narrow group of actors who tend to be vocal on competition issues – publishers, app developers, think tanks and digital rights activists. In fact, UK businesses of all sizes pay the price, even if they don’t often speak up. We surveyed a broad sample of UK businesses (n=417), selecting participants to include all businesses that are significantly reliant on large technology platforms or digital infrastructure providers for their work – for anything from data storage, cloud compute and operations to marketing and revenue. This filter was designed to include a large proportion of the UK economy.4 Of these businesses, 79 per cent are concerned about large technology companies using their dominance in digital markets to limit competition. They cite several concrete harms that affect them directly. High on the list of problems is a lack of resilience. Of the businesses in the sample, 62 per cent say they would find it “costly and disruptive” to switch away from their main technology provider, and 7 per cent say they would not be able to switch at all. In parallel, 50 per cent report that they could not continue operating if they lost access to their main technology supplier or they could only do so for a matter of days (see figure 3.1). FIGURE 3.1 Half of businesses would cease to function within days of losing access to their most important technology provider How long businesses could function without access to their most important technology provider
17%
Not at all
33%
Days 22%
Weeks 15%
Months A year
3% 8%
Longer than a year Don’t know
2%
Source: Opinium polling for IPPR (2026, n=417) 4
The following question was used to screen out businesses that would not interact significantly with the companies in question. “Is your business significantly reliant on large technology platforms and/or digital infrastructure for any of the following? Generating sales or revenue, marketing and acquiring customers, core business operations (such as payments, logistics, communications), software, cloud computing or data storage, selling or distributing products and services online”. We also excluded sole traders. Full details of the survey method are available in the appendix.
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
13
Businesses report direct experience of many of the extractive practices that the Digital Markets, Competition and Consumers Act 2024 (DMCCA) is designed to address (see figure 3.2) and this is having a dampening impact on innovation. More than a third of businesses report that they decided not to offer a product in the past two years because of an incumbent technology giant. FIGURE 3.2 UK businesses experience the harms of market concentration The percentage of businesses experiencing a range of market concentration impacts in the past two years Paid fees to a platform they felt they had no realistic choice but to accept
62%
Had to change how their business works because a platform's rules required it
47%
Experienced a platform changing its algorithm without explanation
46%
Found that a platform favoured its own products or services over theirs
46%
Decided not to offer a new product because a tech giant already dominated that space
35%
Been unable to access their own customer or sales data
34%
Source: Opinium polling for IPPR (2026, n=417)
Finally, these results show that businesses would back many of the specific conduct requirements that the DMCCA makes possible, with support consistently high and opposition consistently low (see figure 3.3). Not only that, an average of 72 per cent of businesses say they would produce a better product or lower their prices if these measures were introduced. FIGURE 3.3 Businesses’ perceptions of potential DMCCA conduct requirements are consistently positive The percentage of businesses that would find three proposed DMCCA rules useful Rules improving interoperability, so businesses can switch provider Rules requiring more transparency (eg explaining algorithms) Rules giving more notice explanation before pricing changes
7% 92% 8% 92% 7%
NET: Not useful
92% NET: Useful
Source: Opinium polling for IPPR (2026, n=417) Note: DMCCA = Digital Markets, Competition and Consumers Act 2024.
14
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
4. GROWTH AND SOVEREIGNTY: WEIGHING THE TRADE-OFFS Despite all the evidence that the status quo is materially harming UK businesses, there has not been a push from government on competition enforcement. This is in large part due to fears around the side effects of doing so. What if big tech withdraws investment, or product launches get delayed, or diplomatic retaliation places the UK on the back foot? But, the positive case for enforcement outweighs the risks. Proportionality remains the right test where investment and diplomatic risks exist, but that test still points to enforcement that is bolder than what we have seen to date.
GOOD GROWTH AND INWARD INVESTMENT
The positive case Concentration in digital markets is already extracting value from the UK and holding back businesses’ growth. Among the UK businesses surveyed for this research, concentration in digital markets was ranked as the second highest constraint on their growth, above access to finance, talent and macroeconomic conditions (see figure 4.1). This finding holds for startups and for the UK’s most profitable companies. Among small businesses (those with 10–49 employees), 33 per cent report this as an issue, while 32 per cent of companies with an annual turnover of more than £50 million say this has had an impact on them. FIGURE 4.1 Market power ranks high as a constraint on UK businesses’ growth Percentage of businesses that have experienced various constraints on growth in the past two years
41%
Regulation and compliance costs Market power of large technology companies or digital platforms
34% 32%
Access to finance
29%
Access to talent 25%
Macroeconomic conditions 14%
None of the above Don’t know
1%
Source: Opinium polling for IPPR (2026, n=417)
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
15
Furthermore, 66 per cent of businesses report that they send at least half of their digital spend straight to the largest US technology giants (Google, Apple, Microsoft, Amazon, Meta and NDIVIA) (see figure 4.2). This shows how much the ‘dependency tax’ adds up. FIGURE 4.2 Many UK businesses send the majority of their digital budget to US big tech Percentage of businesses sending their digit spend to the US, by proportion of spend
11%
Less than a quarter
22%
Around a quarter
34%
Around half 21%
Around three quarters 10%
Almost all of it Don’t know
2%
Source: Opinium polling for IPPR (2026, n=417)
But clawing back this dependency tax is not even the main prize. Concentration harms today’s businesses, but it kills tomorrow’s. Platform dominance suppresses investment in adjacent markets too, making it hard for innovative businesses to scale (Kamepalli et al 2022). Research from 2003 to 2021 shows that after a technology giant acquires a company, both the flow of venture capital funding and the rate of new startup formation drop in that market. In the US, venture capital investment was six times lower in affected markets and in Europe it was 60 per cent lower (Braesemann 2026). Investors have expressed this concern before, with 80 per cent of those surveyed in 2021 saying they worried about incumbent firms making it hard for new players to enter the market (Hallas and de Coverly Veale 2023). Founders might, of course, set out hoping to be acquired, but investors price those rounds accordingly, and what the wider economy needs is investment aimed at scaling, not selling. The Digital Markets, Competition and Consumers Act 2024 (DMCCA) is designed with the concerns of startups and investors in mind. Only companies with a global revenue of more than £25 billion or UK revenue of more than £1 billion are in scope for the DMCCA’s strategic market status (SMS) regime and, rightly, this keeps almost everyone out of its remit (House of Commons 2024). Consolidation deals between medium-sized firms can go ahead. For startups, the remaining concern is that competition enforcement could close off their most profitable route to exit – being acquired by a US technology giant. The Competition and Markets Authority (CMA) should monitor this closely and engage with the startup community to understand how its work can help startups profit without them needing to be acquired by the technology giants.
16
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
The UK also needs to prioritise good growth in every postcode over a short-term ‘sugar rush’. Competition enforcement encourages investment where companies grow and scale in the UK (Coyle 2025). And for growth to be broadly shared, it is essential to tackle the pattern whereby concentration leads to dominant firms capturing an increasing share of profits, while wages stagnate and labour share falls (Autor et al 2020, Dibb and Valletti 2025). Finally, it is worth remembering the extent to which the public feel competition interventions in pounds saved. The CMA reports average consumer savings of more than £3 billion a year from its work (CMA 2025b). And the DMCCA savings are potentially huge. Google Search alone earned between £3 billion and £4 billion in excess profits in 2024 (CMA 2025c). If similar excess profits could be avoided across the technology giants, the cost-of-living impact could be huge. The trade-off The trade-off here is not hard to find. If enforcement has an impact on the profits of technology giants, the fear is they will withdraw investment at exactly the moment when the UK most needs access to their technology and investment in the UK’s infrastructure. Managing the trade-off The question becomes: Will big tech retaliate? And where would withdrawal hit hardest? Big tech firms are known to use threats as a negotiation tactic. They make empty promises of market exit even where the UK clearly remains attractive enough that the government can confidently call their bluff. In Australia, Google threatened to withdraw following a law that required it to compensate news providers. It didn’t leave (BBC 2021). In the EU, fines have been imposed on companies such as Apple and Meta, without withdrawal from these companies (EC 2025). And, even in the context of regulation that genuinely poses a large compliance burden on big tech firms – that is, European-style data protection – research shows a small effect. Only 7 per cent of model releases were delayed in the UK, based on data from 2018 to 2026, and the main culprit was data protection (Lidiard et al 2026). There are, of course, cases where regulators have gone too far, and technology companies have exited or delayed product launches. Apple, for example, delayed Siri AI in the EU due to the Digital Markets Act 2022 (Apple 2026). The point here, though, is that governments should predict whether threats are credible based on the impact of the regulatory intervention and the value of the UK market to these companies, not by taking the technology giants at their word. Additionally, far more specificity is needed on what sort of regulation affects what sort of investment. The reality of the UK’s position means data centres and investment from frontier AI companies should be a top priority. But there isn’t much evidence that the DMCCA threatens this. Competition regulation is low on the list of reasons why a hyperscaler might withdraw investment. First, it will complain about energy costs, planning regulation, employment rights and data protection. Consider the landmark £31 billion promised as part of the UK–US tech prosperity deal that was suspended in December 2025. The grievances that the US cited were the UK’s digital services tax, food standards and Online Safety Act 2023, not the DMCCA (Courea 2025). And the AI companies the UK most wants to attract are not the ones this regime currently constrains. The CMA is targeting entrenched gatekeepers while the AI labs are moving to reduce their dependence on established players. For example, OpenAI (an AI research organisation) has diversified its cloud partnerships and increasingly these companies are spreading compute across multiple providers. Enforcement IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
17
that keeps their access to deeper layers of the stack competitive can in some ways be good for them. The conduct requirements introduced for Google Search, for instance, prevent Google from obtaining an unfair advantage over the likes of Anthropic (an AI safety and research company) and OpenAI. Eventually, the frontier AI companies might come in scope, but it takes time for market power to become ‘entrenched’ – a key requirement for consideration under the DMCCA. And, were investments to be reduced, smart industrial policy could make up some shortfall. CMA interventions do have an impact on big tech’s bottom line – for example by preventing cloud providers or app stores from overcharging – but they make no difference to the value for money of infrastructure investment in the UK. So, if this is the sort of investment the UK needs most, policy should target that specifically, offering accelerated grid connections and targeted planning reform that helps technology giants see a return on their investment sooner. The relevant question is not whether enforcement carries any short-term cost – almost any structural reform does – but whether that cost is (a) large, (b) unavoidable or (c) not outweighed by the productivity return. On the current evidence, it is none of these. It would be naïve to suggest that bolder enforcement comes with no cost, and some investment decisions are sensitive to the CMA’s approach. For this reason, the UK is fortunate to have a regulator that already works closely with industry to design interventions with broad buy-in. Sometimes, though, we will need to accept that a short-term investment cost is worth it if the harm is significant and the longterm gains are clear.
THE UK’S SOVEREIGNTY AND GEOPOLITICAL REALITY
The positive case For the UK, an equally urgent goal is digital sovereignty and resilience. The UK needs to build a strong domestic ecosystem so that more of the profits from emerging technologies accrue in the UK and the UK does not end up fully at the mercy of US export controls. If the UK continues to prioritise investment from US technology giants, it is, to an extent, trading that investment off against the growth of UK businesses. And in the meantime, the UK’s supposed technology champions will continue to pay a dependency tax to incumbents, only to ultimately be outcompeted or acquired. In figure 4.3 we illustrate how this may play out for a fictional UK startup.
18
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
FIGURE 4.3 One year to create a UK AI champion
January
February
March
Choose cloud infrastructure You need cloud infrastructure to get started. In practice, there are three hyperscale choices: Amazon Web Services (AWS), Google or Microsoft. The public alternative has a queue and the European ones never scaled. You sign with AWS. Everyone else you know has.
Pick a foundation model You need a foundation model. Three US labs make the best ones that will keep your product competitive. Whatever they charge per token, you pay. There is no one to negotiate with, because there is nowhere else to go.
April Access graphics processsing unit (GPU) compute for training May
June
Fine-tuning time. Now you are paying twice: the chip-maker’s margin, then the cloud provider’s markup on top. Two tolls, limited options and no guarantee of competitive pricing.
Investment is chilled Your product has started working. You go back to your investors to seek more funding. But they hesitate. Not because your product is bad. Instead, they flag that an incumbent with an existing customer base will simply copy your product and win. They do not think it is worth the risk.
July
August
September
Launch your app Now is the time to launch your product with real customers, via the app store. However, 15–30 per cent of customers’ payment is siphoned off before it reaches you. The CMA is consulting on measures that could stop this, but the decision will not come soon enough.
October
November
December
Big tech buyout Against the odds, you are doing well. You even have investment from the UK’s sovereign AI fund. At this point, one of the incumbents offers to buy your product. At the time, this feels like the only choice – how can you compete with a technology organisation so big?
Source: IPPR analysis of a hypothetical AI startup
Our survey shows the extent to which market concentration already prevents the UK from reaping the economic rewards of new technologies. Not only are 66 per cent of businesses sending at least half of their digital spend straight to
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
19
the US technology giants, but with AI, they are also buying from a narrow set of companies, none of which are based in the UK. The most popular AI product is ChatGPT, followed by two AI tools that incumbent technology giants have built into tools that businesses already use, namely Microsoft Copilot and Google Gemini (see figure 4.4). So far, the foundation model market remains quite competitive, but there is a risk that incumbents are using their existing customer base to outcompete others. FIGURE 4.4 UK businesses pay for the ChatGPT AI tool the most, followed by Microsoft Copilot and Google Gemini The AI tools British businesses pay for (%)
74%
ChatGPT / OpenAI Microsoft Copilot
61%
Google Gemini
59% 38%
Claude / Anthropic Perplexity
18%
Deepseek
16%
Mistral
5%
Other 0% Source: Opinium polling for IPPR (2026, n=253 – that is, all businesses that pay for at least one AI tool)
The UK government knows it needs to fix its lack of leverage across the AI stack. The supply-side agenda is growing fast, with a £500 million sovereign AI fund and advanced market commitments to help UK companies compete and grow (DSIT 2026). But without competition enforcement to allow these companies to scale, this agenda will ultimately fail as companies get outcompeted or acquired. Competition, of course, is not a guarantee that companies will succeed, but it is a guarantee of a fair playing field. The CMA can ensure UK companies have access to essential inputs (the cloud, app distribution and so on) on fair terms. It can prevent large digital platforms from preferencing their own products, meaning that technology giants that already have all the customers cannot simply copy startups’ ideas. It can also prevent dominant firms buying out and shutting down budding UK technology companies before they grow. Coyle (2026) has described today’s strategy, whereby industrial policy creates emergent champions, only for the government to pull back from protecting those champions, as “incoherence”. The trade-off Where the UK’s attempt to build sovereign capability comes at a cost for the US technology giants, there is a risk of diplomatic retaliation that the UK is unable to absorb.
20
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
Managing the trade-off The question then becomes: How far can the UK go before diplomatic retaliation becomes a genuine risk? Even with greater levels of enforcement, the risk is lower than is often assumed, for two reasons. First, the EU comparison is reassuring. Despite a more forceful regulatory regime, the EU has not seen significant geopolitical implications as it has gone ahead on implementation. So, the idea of this happening to the UK – with its more flexible regime and closer CMA engagement with industry – is unlikely, especially as the UK remains one of the largest markets for US technology giants. Of course, the UK is at risk when its competition regulator is out on a limb, but where greater enforcement simply means catching up with other jurisdictions, the risk is lower. Second, where diplomatic retaliation has surfaced within UK technology policy, for example on the UK–US tech prosperity partnership, the pattern has been to complain about a broad suite of UK regulation, from the digital services tax to food standards. And, as noted above, digital competition has not yet ranked highly as a concern. Of course, this calculus could change if the CMA became much bolder and this is a reason for a thorough risk assessment ahead of any intervention. Where diplomatic retaliation is a risk, synchronised activity from multiple competition authorities across countries becomes essential. Consistent implementation across jurisdictions can be good for technology giants, as consistent requirements make compliance easier for them. Coordination is also useful as a safeguard against retaliation because the shared bargaining power of multiple international markets reduces the threat from any one actor. But coordination is not easy. It requires a deliberate balance. The CMA is right to prioritise UK interests and lead in markets where the UK is best placed to intervene (CMA 2025d). Yet, at the same time, this can sometimes lead to a ‘game of chicken’ where every competition authority defers action and it waits for the others to move. The UK’s approach must therefore be to actively build coalitions that strengthen the UK’s hand rather than treating a lack of international comparators as a reason for delay. The diplomatic costs of competition enforcement are clearly not zero. But any government serious about sovereignty has to accept that efforts to build UK technology might come with US friction. The UK should not be reckless, and for this reason, proportionality does remain a key test for the CMA. But proportionality should not be used as an excuse to fall back on more traditional approaches to competition policy. The DMCCA was created for a reason, and the reason was not to ensure the CMA finds a compromise with big tech at every stage.
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
21
5. RECOMMENDATIONS ISSUE A NEW STRATEGIC STEER FOR THE CMA
The government should issue an updated strategic steer for the Competition and Markets Authority (CMA) to provide the regulator with a reset and signal to the market that implementation of the digital markets regime will be consistent in the future. The new strategic steer should redefine what pro-growth means in digital markets and add language that clarifies the CMA’s role. The steer should: • request strong action to open up digital markets to competition, favouring legally binding requirements rather than voluntary measures • reference digital sovereignty and the need to reduce the market power of gatekeeper firms • commit explicitly to preserving the CMA’s independence • set pace expectations via a requirement that the CMA publishes and consults on an annual forward plan of markets under consideration for strategic market status (SMS) investigation • direct the CMA to evaluate cumulative leverage across the digital and AI stack layers rather than simply examining markets in silos. Then, the government should step back. Too much intervention could easily backfire if politics is again seen to intervene in the independence of the CMA.
PRIORITISE THE REMEDIES THAT BOOST GROWTH
So far, the CMA has made more progress towards helping dependants than it has towards helping competitors. If it is to truly support the UK’s growth, this must change. But different markets allow for different things and prioritisation must reflect that. In some markets, competition can realistically be strengthened, especially right now, because new technologies such as AI are cracking open markets that looked closed for years. In others, network effects and deep integration make the structure hard to shift. And in some markets, the growth potential is bigger than in others, for example in cloud and AI, because these are inputs across the rest of the economy. In other markets, such as internet searches and app stores, the gains are narrower and it does make sense to focus specifically on the businesses that are most severely harmed by big tech giants (for example, search publishers and app developers for mobile ecosystems). Below we consider four illustrative cases: • Search. The CMA should continue protecting dependants (for example, search publishers) as a near-term priority, but keep this work under close review, with a view to bringing Gemini within scope. Google has dominated search for more than a decade and looked like a natural monopoly, but with AI transforming search, Google’s grip could be eroded and competition opened up in AIenhanced search.
22
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
•
•
•
Cloud. The CMA should review its decision to accept voluntary commitments from the major cloud providers instead of designating them with strategic market status. Economies of scale mean that the cloud will never be a fragmented or highly competitive market, but given this is a service everyone relies on – from UK businesses, to the NHS, to members of the public – it is a key market to prioritise. Furthermore, creating conditions that encourage even one more player to enter the market could be enough to bring down prices, improve public sector resilience and support UK business growth. App stores. Structural change in near-term market structure is unlikely, so the CMA should focus on conduct requirements that give dependants real protections. Capping commissions, providing alternative in-app payment methods and enabling direct developer-to-consumer app downloads would all help here. AI. The structure of the AI market is not yet settled and the potential for the CMA to have an impact on this is huge. The CMA should act early where incumbents use dominance in adjacent markets to gain ground in AI and it should be ready to intervene far faster than it has in previous markets, given that market concentration could tip quickly.
Requiring dominant firms to let data, customers and services move in and out on fair terms lowers the switching costs and removes the lock-in that keeps these markets closed. Given how directly interoperability serves the growth and sovereignty case, across all digital markets the CMA should treat interoperability as the priority remedy wherever market structure allows.
INTRODUCE REPORTING REQUIREMENTS THAT SUSTAIN THE CMA’S MANDATE
To sustain this regime, the CMA needs to demonstrate progress by publishing disaggregated measures of success. We propose that, in addition to annual reports that detail the CMA’s overall benefit to the UK’s economy, a new regular progress report on digital competition is introduced, reporting progress on four fronts: • Consumers. How has the Digital Markets, Competition and Consumers Act 2024 (DMCCA) saved consumers money and eased the cost of living? • Dependants. How has the DMCCA improved conditions for the businesses most dependent on big tech for survival? • Competitors. Have rivals emerged, have market shares shifted and have sovereign competitors succeeded? • Democracy. Has progress been made on the political consequences of market concentration? And are there improvements for our information environment?
ALIGN BROADER TECH POLICY PRIORITIES TO ENSURE IMPACT
The new steer and set of priorities for the CMA will only work if other government policies are sufficiently coordinated. For this to all come together, we recommend the government focusses on the following. Back up competition enforcement with bold industrial policy Opening markets is only half the job. UK companies need to be ready to fill the space if this is to work as part of the UK’s digital sovereignty strategy. This means doubling down on efforts within the sovereign AI unit, providing the compute access and procurement guarantees needed to give UK challengers a genuine runway. Ultimately, if the CMA succeeds in prising open markets but there is no UK alternative ready to step in, the market will remain concentrated.
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
23
Protect the CMA’s independence The past year has shown how easily political signals can chill enforcement. In this context, the competition reform bill – currently before parliament – is an opportunity to cement the regulator’s independence. For example, the government could remove the power for the secretary of state to approve (or withhold approval for) updates to the CMA’s guidance on the digital markets regime, and not take forward plans for this power to be extended to other CMA guidance documents. Changes to markets and merger regimes must also preserve the regulator’s independence. Coordinate with international partners to make competition enforcement faster, not slower Acting in step with the EU and others can make conduct requirements more consistent for platforms and blunt the threat of retaliation. But this must not be used as an excuse to wait for others – especially the US – to act. There are two key reasons for this. First, competition authorities or courts in other countries will not be constructing remedies with UK consumers or businesses in mind. Second, through the DMCCA, the UK has superior powers to drive targeted yet fundamental changes in digital markets. So, while other jurisdictions may act, the outcomes are likely to be suboptimal for the UK market.
MAINTAIN A PROPORTIONATE ENFORCEMENT FRAMEWORK
Bold action must not be confused with over-regulation. To safeguard the UK’s wider investment climate, the CMA’s interventions must remain strictly evidence based, proportionate and transparent. The CMA must avoid unnecessary regulatory burdens on scaling firms and attempts to micro-manage fast-moving technology markets, while continuing its industry engagement, which to date has helped ease the introduction of new conduct requirements. The goal should be spurring innovation and enabling challengers, not penalising scale itself.
24
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
REFERENCES Apple (2026) ‘Due to DMA, Siri AI delayed in EU for iOS 27 and iPadOS 27’, Apple website, 8 June 2026. https://www.apple.com/uk/newsroom/2026/06/due-to-dma-siri-aidelayed-in-eu-for-ios-27-and-ipados-27 Autor D, Dorn D, Katz L, Patterson C and Van Reenen J (2020) ‘The fall of the labor share and the rise of superstar firms’, The Quarterly Journal of Economics: 645–709. https://economics.mit.edu/sites/default/files/publications/Autor%20et%20al_2020_ The%20Fall%20of%20the%20Labor%20Share%20and%20t.pdf BBC (2021) ‘Google threatens to withdraw search engine from Australia’, BBC News website, 22 January 2021. https://www.bbc.co.uk/news/world-australia-55760673 Braesemann F (2026) ‘When big tech plays like Man United: how kill zones threaten Europe’s future’, Oxford Internet Institute website, 16 January 2026. https://www.oii.ox.ac.uk/ when-big-tech-plays-like-man-united-how-kill-zones-threaten-europes-future Competition and Markets Authority [CMA] (2021) ‘Mobile ecosystems market study’, GOV.UK website, 15 June 2021. https://www.gov.uk/cma-cases/mobile-ecosystems-market-study Competition and Markets Authority [CMA] (2023) ‘AI Foundation Models: initial review’, GOV.UK website, 4 May 2023. https://www.gov.uk/cma-cases/ai-foundation-modelsinitial-review Competition and Markets Authority [CMA] (2025a) ‘Cloud Services Market Investigation: Summary of final decision’, press release, 28 July 2025. https://www.gov.uk/cma-cases/ cloud-services-market-investigation#final-report Competition and Markets Authority [CMA] (2025b) ‘Annual report and accounts 2024 to 2025’, GOV.UK website, 10 July 2025. https://www.gov.uk/government/publications/cmaannual-report-and-accounts-2024-to-2025/annual-report-and-accounts-2024-to-2025 Competition and Markets Authority [CMA] (2025c) ‘Appendix C: Profitability analysis’ in Strategic market status investigation into Google’s general search services: Final decision. https://assets.publishing.service.gov.uk/media/68e8b32257038b5739b9861a/ Appendix_C_-_profitability_analysis.pdf Competition and Markets Authority [CMA] (2025d) ‘CMA strategy 2026 to 2029’, GOV.UK website, 20 November 2025. https://www.gov.uk/government/publications/cmastrategy-2026-to-2029/cma-strategy-2026-to-2029 Competition and Markets Authority [CMA] (2026a) ‘CMA proposes package of measures to improve Google search services in UK’, GOV.UK website, 28 January 2026. https://www.gov.uk/government/news/cma-proposes-package-of-measures-toimprove-google-search-services-in-uk Competition and Markets Authority [CMA] (2026b) ‘CMA launches strategic market status investigation into Microsoft’s business software ecosystem’, press release, 14 May 2026. https://www.gov.uk/government/news/cma-launches-strategic-market-statusinvestigation-into-microsofts-business-software-ecosystem Competition and Markets Authority [CMA] (2026c) ‘Further CMA action to secure a fairer deal for businesses and improve Google search services in UK’, press release, GOV.UK website, 17 June 2026. https://www.gov.uk/government/news/further-cma-action-to-secure-afairer-deal-for-businesses-and-improve-google-search-services-in-uk Competition and Markets Authority [CMA] (2026d) ‘CMA consults on new requirements for Apple and Google’s mobile platforms’, press release, GOV.UK website, 30 June 2026. https://www.gov.uk/government/news/cma-consults-on-new-requirements-for-appleand-googles-mobile-platforms Cottier B, Rahman R, Fattorini L, Maslej N and Owen D (2024) ‘The rising costs of training frontier AI models’, ResearchGate website. https://www.researchgate.net/ publication/381108862_The_rising_costs_of_training_frontier_AI_models
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
25
Courea E (2025) ‘US puts £31bn tech “prosperity deal” with Britain on ice’, Guardian, 15 December 2025. https://www.theguardian.com/us-news/2025/dec/15/us-pauses-techprosperity-deal-britain-donald-trump-keir-starmer Coyle D (2025) ‘The relationship between competition policy and industrial policy in an era of structural change’, Intereconomics, 60(4): 205–209. https://www.intereconomics.eu/ contents/year/2025/number/4/article/the-relationship-between-competition-policyand-industrial-policy-in-an-era-of-structural-change.html Coyle D (2026) ‘Reimagining the AI arms race’, Bennett School of Public Policy website, 29 June 2026. https://www.bennettschool.cam.ac.uk/blog/reimagining-the-ai-arms-race Department for Business and Trade [DBT] (2025) ‘Strategic steer to the Competition and Markets Authority’, GOV.UK website, 15 May 2025. https://www.gov.uk/government/ publications/strategic-steer-to-the-competition-and-markets-authority/strategic-steerto-the-competition-and-markets-authority Department for Science, Innovation and Technology [DSIT] (2026) ‘AI firms pioneering drug discovery, cheaper supercomputing and more get first backing through UK’s sovereign AI’, press release, GOV.UK website, 16 April 2026. https://www.gov.uk/government/ news/ai-firms-pioneering-drug-discovery-cheaper-supercomputing-and-more-getfirst-backing-through-uks-sovereign-ai Dibb G and Valletti T (2025) Fair play: How competition policy can drive growth, IPPR. https://www.ippr.org/articles/fair-play European Commission [EC] (2025) ‘Commission finds Apple and Meta breach Digital Markets Act’, press release, 23 April 2025. https://digital-strategy.ec.europa.eu/en/news/ commission-finds-apple-and-meta-breach-digital-markets-act Furman J, Coyle D, Fletcher A, McAuley D and Marsden P (2019) Unlocking digital competition: Report of the Digital Competition Expert Panel, HM Treasury. https://www.gov.uk/ government/publications/unlocking-digital-competition-report-of-the-digitalcompetition-expert-panel Hacker P (2026) ‘The world’s most complex machine’, Works in Progress, (13), 23 April 2026. https://worksinprogress.co/issue/the-worlds-most-complex-machine Hallas S and de Coverly Veale H (2023) Building competition: A new lens for digital competition policy in the UK, Coadec. https://coadec.com/wp-content/ uploads/2023/06/Building-Competition_-Report-040623.pdf House of Commons (2024) Digital Markets, Competition and Consumers Act 2024, The Stationery Office. https://www.legislation.gov.uk/ukpga/2024/13/pdfs/ ukpgaen_20240013_en.pdf Kamepalli SK, Rajan R and Zingales L (2022) ‘Kill zone’, working paper 27146, National Bureau of Economic Research. https://www.nber.org/system/files/working_papers/ w27146/w27146.pdf Kirkwood M (2026) ‘UK regulator probes Microsoft while backing voluntary cloud rules’, Tech Policy Press website, 2 April 2026. https://www.techpolicy.press/uk-cloud-regulatoropts-for-voluntary-commitments-launches-microsoft-investigation Lidiard J, Vereschak O, Gibbs T and Anderljung M (2026) ‘Delays to frontier AI in the EU and UK’, Centre for the Governance of AI website, 28 June 2026. https://www.governance.ai/ research-paper/delays-to-frontier-ai-in-the-eu-and-uk Mitchell A (2026) ‘Nvidia statistics 2026’, Axis Intelligence website, 25 June 2026. https://axis-intelligence.com/nvidia-statistics Office of Technology Staff (2025) ‘Behind the FTC’s 6(b) report on large AI partnerships & investments’, technology blog, Federal Trade Commission website, 17 January 2025. https://www.ftc.gov/policy/advocacy-research/tech-at-ftc/2025/01/behind-ftcs-6breport-large-ai-partnerships-investments Pratley N (2025) ‘Does the government really know what it wants from the CMA?’, Guardian, 13 February 2025. https://www.theguardian.com/business/nils-pratley-on-finance/2025/ feb/13/does-the-government-really-know-what-it-wants-from-the-cma Ring S and Pickard J (2025) ‘How the UK’s competition regulator lost the trust of ministers’, Financial Times, 16 February 2025. https://www.ft.com/content/0cc18c6e-bab5-43deae10-d72bbe129294?syn-25a6b1a6=1
26
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
Scanlan R (2026) ‘UK Competition and Markets Authority accepts voluntary commitments to address cloud market concerns’, Steptoe website, 10 April 2026. https://www.steptoe.com/en/news-publications/stepahead-antitrust-and-competitioninsights/uk-competition-and-markets-authority-accepts-voluntary-commitments-toaddress-cloud-market-concerns.html Widder D and Kim M (2025) ‘How big cloud becomes bigger: scrutinising Google, Microsoft and Amazon’s investments’, SSRN website, 4 August 2025. https://papers.ssrn.com/sol3/ papers.cfm?abstract_id=5377426
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
27
APPENDIX: METHODS This research drew on two novel sources of evidence: an online survey of UK businesses conducted by the market research agency Opinium; and a series of qualitative interviews with businesses, academics, civil society, legal experts and officials.
THE SURVEY
Opinium conducted an online survey, designed by IPPR, of 417 UK businesses between 26 June and 3 July 2026. This sample was targeted towards businesses that interface in some way with large technology providers and platforms. Because of this, the following screening question was used to filter out businesses without relevant experience of these providers and platforms: <bn>“Is your business significantly reliant on large technology platforms and/ or digital infrastructure for any of the following? Generating sales or revenue, marketing and acquiring customers, core business operations (such as payments, logistics, communications), software, cloud computing or data storage, selling or distributing products and services online.” Sole traders were excluded from the sample, which spanned a wide range of business sizes: 63 micro (1–9 employees), 88 small (10–49 employees), 105 medium (50–249 employees) and 161 large (250+ employees) businesses. The sample also spanned a range of business functions: • 108 consumer facing and trade • 93 knowledge and professional services • 90 technology and digital • 83 production and industry • 20 healthcare and life sciences • 13 education • 10 transport. Senior figures in the businesses provided responses to the survey.
THE INTERVIEWS
We conducted 12 semi-structured interviews between 22 June and 9 July 2026. Participants included: • businesses affected by competition issues • academic and legal experts on digital competition • officials and former officials working closely on the Digital Markets, Competition and Consumers Act 2024 (DMCCA). The sample was not intended to be representative, and participants were recruited through existing networks. Findings from the interviews should be read as illustrative of the experiences and attitudes of those with direct experience of uncompetitive digital markets.
28
IPPR | Bottleneck Britain How digital competition could power UK growth and sovereignty
Institute for Public Policy Research
GET IN TOUCH For more information about the Institute for Public Policy Research, please go to www.ippr.org You can also e-mail info@ippr.org or find us on X @ippr
Institute for Public Policy Research Registered Charity no. 800065 (England & Wales), SC046557 (Scotland), Company no, 2292601 (England & Wales)
The progressive policy think tank