Good evening from New Economy Brief.
There’s nothing new about technological innovations driving financial bubbles, as investors speculate on highly profitable future uses. The dotcom boom is a typical example; the crypto boom of 2020 another more recent one. It is more unusual for the CEO of a company at the centre of this boom to openly say ‘yes’, we are in a bubble. Sam Altman, CEO of OpenAI, did just this in August 2025.
This was before OpenAI closed deals with other AI-focused companies totalling $1 trillion. Commentators noted the circularity of many of these deals, tying the companies’ fates ever closer together. Major financial institutions like the IMF and the Bank of England now warn about the risks from a deflating AI bubble. This week, we're exploring how the fallout from an AI bust might affect the UK.
All aboard the AI express
Successive UK governments quickly embraced generative AI. Back in 2023, then-Prime Minister Rishi Sunak sought to position the UK as a frontrunner in AI safety. The Labour government is putting AI at the heart of its economic strategy, setting out its AI Opportunities Action Plan in January 2025.
This includes building out physical infrastructure, like data centres, to deploy AI at scale. The government has created ‘AI growth zones’ in Culham (Oxfordshire), Cobalt Park (Newcastle) and Blyth (Northumberland) to fast-track this. Companies like Microsoft and Blackstone could invest up to £30 billion in these projects.
This infrastructure could create a large number of jobs: both in the years-long construction process and through new opportunities for AI research and businesses more broadly. Big Tech companies are also in talks with the government to provide AI skills training to millions of UK workers, with the hope of boosting workforce efficiency.
The UK is also home to the largest start-up investment ecosystem in Europe. Venture capital (VC) investment reached £9 billion in 2024, with over a quarter of that going into AI companies. Globally, the UK ranks third in AI investments, a quantity that even affected demand for sterling. For the government, these AI companies are essential to keep the UK at the forefront of AI job and growth creation.
Hubble, bubble, AI and trouble
With near-zero growth in the UK and a challenging fiscal position, the government has fully embraced AI to turn the economy around. Underlining this, the Prime Minister took forward all 50 recommendations from the AI Opportunities Action Plan. But now experts are ringing the alarm bells about an AI bust, how might this impact the UK economy and one of Labour’s big political commitments?
The bust’s epicentre would likely be the USA. Stock markets might be affected drastically, with AI stocks representing 39% of S&P 500 market cap. This means that 39% of the total value of shares on the world’s most important stock index comes from AI. By contrast, the FTSE100 is less exposed to tech stock valuations with technology companies constituting closer to 5% of its total market capitalisation.
Nevertheless, major UK pension funds are invested in US technology stocks. An AI slowdown would tank these stocks and so affect retirement savings in the UK. Only 20% of workplace defined contribution pension schemes are invested in the UK, down from 50% 10 years earlier. This means UK workers now have a lot more to lose if big US companies start failing.
A more common effect of a stock market downturn is the rising cost of borrowing. The Bank of England has warned that a ‘sharp correction’ might impact the ‘cost and availability of finance for households and businesses’. This typically follows from hedge funds selling off their positions during a downturn, pushing down the value of the asset they are selling. Other investors holding these assets then receive ‘margin calls’ from their banks or brokers, asking them to post more collateral to keep holding their position or otherwise sell it off, too. Often forced to sell, this leads to a rapid liquidation of related assets which might cascade into a broader financial crisis. Yet while a possibility, this danger might be mitigated by new regulatory provisions following the Global Financial Crisis of 2008.
More likely than a broader crash is an immediate funding freeze for infrastructure projects (like largely debt-financed data centres) and AI training schemes. The £30 billion of investments pledged by Microsoft and co. are private company commitments that can be withdrawn should the tech giants feel the need to curb their spending. A significant drop in Big Tech shares might pressure them to improve profitability and focus spending on operations critical for their bottom line. Training programmes for UK workers also seem a likely target of early funding cuts following an AI downturn. Over 80,000 people work in the AI sector, and 2.2 million people work in tech-related positions – roughly 6% of the workforce. This means a significant number of workers might be abandoned mid-training.
Even more severely affected might be the AI start-up space. Historical examples from the dotcom bust and the 2022 reversal of monetary policy show that VC funding can quickly dry up in a downturn. When the Bank of England started hiking interest rates from 0.1% in December 2021 to over 5% 18 months later, late-stage VC funding dropped 75%. When investors want to see profitability, start-ups starved of funds are quickly forced into bankruptcy or consolidation. This would weaken domestic AI-driven growth and employment.
What Next?
The above points suggest that an AI downturn could be a genuine threat to the government’s growth plans. But there are a few immediate steps that could be taken:
Increase transparency about financial exposure. The focus here should be on large pension schemes with significant exposure to tech stocks. Stress tests might bring more clarity on how a significant drop in the US stock market could bring on ‘fire sales’ and put pension schemes under strain. This proposal is a more specific example of the broader recommendations made by the BoE’s Financial Policy Committee in 2024 to ‘to enable the effective monitoring of the systemic risks from AI’.
Create contingency plans for project financing. If big US funders were to pull out, who could step in to complete the projects? The government might consider public alternatives, like an AI growth fund, to put financing on a more reliable footing. Already existing examples of government-backed stabilisation for start-up financing include funds launched by Bpifrance in France and KfW in Germany.
Avoid overcapacity. While there are huge potential benefits to building out data centre capacity, this need should be re-assessed periodically and adjusted in line with rapid technological progress. Government payouts could be tied to hitting performance targets, e.g. a certain amount of capacity usage. Building data centre infrastructure is not like building railway infrastructure – a lot of components will be obsolete within years, not decades. Data centre overcapacity could therefore be very costly. Besides the financial dimension, building and running computing infrastructure is carbon intensive and can compete directly with the general population’s energy needs.
The UK faces significant exposure to an AI market correction through pension funds concentrated in US tech stocks, billions in potentially volatile data centre investments, and a start-up ecosystem dependent on venture capital. Historical crashes show how quickly tech bubbles deflate, threatening Labour's AI-driven growth strategy. Proactive policy interventions around transparency, contingency financing, and capacity management could help mitigate these interconnected risks. And with even Alphabet (Google’s parent company) CEO Sundar Pichai warning about ‘irrationality’ in the AI boom, this only seems more pressing.
Property tax reform. With the Autumn Budget rapidly approaching, IPPR’s Carsten Jung and Aditi Sriram have set out how the Chancellor can tax property more effectively and fairly. Their recommendations include raising rates on higher council tax bands and lowering those on lower ones, as well as raising the foreign buyer surcharge to ensure overseas investors contribute more to local communities.
Care and private equity. The Centre for Local Economic Strategies (CLES) has published a new report on ending extraction in the UK care system. It finds that in just three UK regions, private care providers extracted more than £250 million in profit over three years. Over a third of these are owned by private equity firms, organisations registered in tax havens, or both.
Scrapping the two-child limit. Analysis by the Women’s Budget Group shows how many children in each nation of the UK would be lifted out of poverty if the two-child limit is fully scrapped at the Budget next week. In England it’s 626,022 children, while in Scotland it’s 51,036. In Wales 54,722 children would be lifted out of poverty and in Northern Ireland the figure is 9,791.
Progressive energy bills. With the Chancellor exploring options to cut energy bills at next week’s Budget, a new briefing from the Joseph Rowntree Foundation (JRF) looks at a range of policy options. It finds that introducing a rising block tariff, which splits bills into a series of ‘blocks’ with tiered prices, offers the best balance of targeting spending at those who need it most, while still reducing bills for the majority of households.
