Bank of England warns mortgage shock is spreading as leveraged hedge fund bets swell AI bubble

The Bank of England has issued a stark double warning about the way global shocks are now travelling through Britain’s economy: first through household mortgages, then through the hidden plumbing of modern markets. The immediate alarm is familiar to millions of borrowers. The deeper concern is less visible, but potentially more destabilising: an investment boom in artificial intelligence, financed with rising levels of hedge fund borrowing and secured, in part, against government debt.

Threadneedle Street said that one million more homeowners are now expected to face higher mortgage bills as fixed-rate deals expire, lifting the total to five million households whose repayments will rise over the next two years. Only months ago, in December, the Bank had put the figure at four million. The change is not a statistical quirk. It is the Bank’s attempt to capture how quickly international events can reset the price of money, and how swiftly that reset lands on kitchen tables.

Officials attributed the deterioration in the outlook to higher borrowing costs linked to Donald Trump’s war on Iran, a conflict that pushed up global risk premiums and helped drive mortgage pricing higher in the UK. The Bank said that the average rate on a two-year fixed mortgage has risen from 4.2 per cent in December to 4.92 per cent now. For many households, that move is the difference between a manageable increase and a sustained squeeze on disposable income.

The Bank’s analysis suggests a split in the experience awaiting borrowers. Across those rolling off fixed-rate deals over the next two years, the average increase is put at about £45 a month, a rise that may be absorbed by some families but will still add pressure in an economy where essentials have remained expensive. For a large and identifiable cohort, the hit is sharper. Around 750,000 families who took out mortgages before 2022, when rates were unusually low, and who must remortgage by the end of this year, face an average rise of £170 a month. These are households whose fixed deals were struck at rates under 3 per cent and are now expiring into a markedly different world.

In the Bank’s telling, this is not simply the delayed effect of past rate rises. It is a reminder that the cost of credit remains vulnerable to external shocks, even when a conflict appears to have moved towards an ending. Mortgage lenders price off expectations for the path of interest rates, the availability of wholesale funding and the perceived risks sitting in the financial system. When those inputs worsen, the repricing can be abrupt, and the borrower only sees the final figure on the remortgage quote.

The political implications are difficult to ignore. The Bank’s warning lands as Andy Burnham, described as the presumptive prime minister, edges closer to power with the cost of living central to his pitch to the electorate. Any incoming government that promises relief will confront the reality that mortgage rates, while set in markets, are influenced by events far beyond Westminster’s reach. There is a familiar temptation in British politics to frame household hardship as a purely domestic failure, one that can be fixed by competence or will. The Bank’s message is that the UK is paying, in real time, for the volatility of an international order in which energy routes, security guarantees and geopolitical risk are being repriced.

Even if the average increase of £45 a month sounds modest beside the worst-case scenarios of recent years, the distribution matters. Mortgage stress is rarely uniform, and it is the clustering of higher payments among particular age groups, regions and income brackets that shapes wider economic outcomes. A household absorbing an extra £170 a month does not merely cut back on luxuries. It delays home improvements, reduces saving, and becomes more cautious about job moves, childcare choices and consumer spending. Multiply that across hundreds of thousands of families and the impact becomes macroeconomic, damping demand and complicating any growth strategy.

At the same time, Threadneedle Street is drawing attention to an entirely different source of fragility, one that sits in the shadow of the household story but is connected to it through borrowing costs and investor behaviour. In its latest assessment of financial stability risks, the Bank said there has been a “significant rise” in hedge fund leverage in equity markets, with officials pointing to borrowing used to invest in AI-related stocks. Rising equity prices, the Bank said, have been driven in part by a narrow set of companies linked to artificial intelligence, increasing concentration in some global indices. That combination, a crowded trade plus heavier borrowing, is the textbook recipe for a violent unwind.

The mechanics are worth spelling out because they go to the heart of why central banks worry about activities taking place outside the traditional banking system. Hedge funds can amplify exposure by borrowing against their portfolios, and in some structures the same asset can be used as collateral multiple times across different transactions. This chain can work smoothly when markets rise or remain calm. In a downturn, it can become an accelerant. Falling share prices trigger margin calls. To meet them, funds sell liquid assets. If those liquid assets include government bonds, the selling pressure can push yields higher, tightening financial conditions for everyone, including mortgage borrowers.

The Bank said the hedge funds it is worried about are not merely borrowing to buy equities. They are also borrowing against gilts and other forms of government debt. Sarah Breeden, the deputy governor for financial stability, warned that this could push up Britain’s borrowing costs in a crisis. “The hedge funds that we’re talking about are active in government bond markets in the UK, which are critically important to financial stability, and they are active across a whole lot of markets internationally,” she said. She warned of “the risk of a spillover, either from one international bond market to another, or from the equity markets and AI firms to UK government bond markets”.

This is the uncomfortable bridge between the two halves of the Bank’s warning. Households feel stress through mortgage rates. Markets can magnify that stress if leveraged investors are forced into rapid sales of safe assets, pushing up gilt yields and raising the baseline cost of capital. The Bank noted that during the Middle East crisis, hedge funds offloading debts pushed up gilt yields. It is an example of how a shock in one market can ricochet into another and end up setting the terms for households and businesses trying to refinance.

The Bank also highlighted how concentrated the AI trade has become geographically and in index weightings. It said AI stocks are increasingly concentrated in specific regions, making up more than half of market capitalisation in the United States, South Korea and Taiwan. Concentration can generate a misleading sense of strength, because broad indices can rise even when most constituents are stagnant. It can also turn a market correction into something more systemic, because index funds, derivatives and leveraged strategies all become tied to the performance of the same narrow set of names.

Volatility is already evident. The Bank said this concentration has increased volatility over the past year as AI stocks have swung sharply. In a normal cycle, that would be a story for traders and tech analysts. With leverage layered on top, it becomes a question for financial stability officials. The Bank is, in effect, asking whether investors have built a market structure that cannot tolerate disappointment, whether from earnings, regulation, geopolitics or a shift in interest rate expectations.

There is a further, more modern anxiety threaded through the Bank’s report: the collision between artificial intelligence and cyber risk. Officials warned of a “significant increase in risks” to financial stability from AI-enabled cyber-attacks against banks and the wider financial system. The concern is not that cyber threats are new, but that AI could lower the cost and raise the sophistication of attacks, widening the gap between the scale of threats and the capacity of firms to defend themselves.

The Bank pointed to a recent warning from the US technology company Anthropic, which said its new system, Mythos, was too powerful for public release because of its ability to identify weaknesses in cyber defences. Whether or not any single system becomes widely available, the direction of travel is clear: tools that can search for vulnerabilities, generate convincing social engineering and automate reconnaissance will spread. In that world, the safety of the financial system depends not only on capital buffers and liquidity, but on operational resilience and the security of third-party providers that banks rely on for payments, cloud services and data.

The Bank set out a severe scenario in which firms and key providers fail to keep pace with the innovation of attackers. Vulnerabilities could accumulate and the risk of a systemic cyber event could rise materially, it said, with the potential for blackouts and attacks that interrupt core functions. Even in more benign scenarios, the Bank argued, a sustained increase in the volume and complexity of attack capabilities would force firms to identify, assess and remediate issues at far greater speed and magnitude than many are accustomed to.

There is an inherent tension here for policymakers. On one hand, the Bank is warning about market fragility, leverage and cyber risk, an argument for caution. On the other, it is signalling a willingness to ease some demands on systemically important UK lenders, including NatWest, Lloyds and Santander UK, despite unease among some officials. Threadneedle Street said it will allow these lenders to run down their emergency buffers more in a crisis and rebuild them more slowly. It will also consult on changes that would allow lenders to hold slightly lower levels of capital relative to their loan book, enabling them to borrow and lend more.

The stated rationale is growth and the desire to cut some of the red tape installed after the financial crisis. There is a legitimate debate to be had about whether the UK has carried an excessively conservative set of requirements compared with peers, and whether the cumulative effect has been to restrain lending unnecessarily. Yet the timing of this shift, alongside warnings about leverage outside the banking system, raises a difficult question about where risk is being allowed to accumulate. If banks are encouraged to lend more with slightly thinner buffers, while hedge funds borrow more to speculate in concentrated equity themes, the system may be shifting risk rather than reducing it.

It also throws into relief the distribution of pain and gain. Households remortgaging into higher rates experience the direct cost of a tighter financial environment. Meanwhile, the frothier edges of markets are being fuelled by borrowed money chasing a narrow set of winners. The Bank’s task is to ensure that the second does not tip over into a crisis that worsens the first, by driving up gilt yields, freezing credit or triggering a sudden tightening in conditions.

In the background sits a larger question about what the AI boom represents. It may be the beginning of a sustained productivity revolution. It may also contain elements of a valuation bubble, inflated by optimism, index dynamics and leverage. The Bank is not in the business of calling market tops. Its language is careful, directed at concentration and the build-up of debt rather than the merits of the technology. But when it says that rising equity prices have been driven by a narrow set of AI-related companies, and that hedge fund leverage has risen significantly, it is indicating that the market’s structure is becoming less forgiving.

For policymakers, including any incoming government, the message is that the cost of living cannot be separated from the stability of the financial system. Mortgage payments are a household line item, but they are also a transmission mechanism for global risk. Market plumbing, often dismissed as technical, can determine whether shocks remain contained or spill into everyday borrowing costs. Cyber resilience, once treated as an IT concern, is now being framed as a systemic financial risk.

The Bank’s report does not offer comforting certainties. It offers a map of vulnerabilities: millions of households facing higher repayments, a leveraged bet on AI valuations that could unwind into government bond markets, and a rising threat of AI-enabled cyber disruption. In combination, these are reminders that Britain’s economic prospects are tied not only to domestic policy choices but to the resilience of institutions and infrastructure in a world where shocks are faster, more interconnected and harder to insure against.