UK Rate Ceiling Holds|AI Rout or Housing Floor?

BoE's Price Cap Warning

The Bank of England just handed Rachel Reeves a public rebuke — and in doing so, told the bond market something about UK rates that equity investors have not yet fully priced in.

On Wednesday, governor Andrew Bailey appeared before the Treasury Committee and called the government's plan to cap supermarket prices on bread, eggs and milk "not a sustainable thing in the long run." His colleague Swati Dhingra added that price controls across Asia and Europe "rarely worked." That is not a technical footnote — it is the Bank signalling it will not treat a government-imposed price floor as genuine disinflation.

The implication runs directly into the mortgage market. At the start of 2026, traders had priced several Bank of England rate cuts into short-dated gilts. The Iran war has since forced a repricing — but Bailey's Wednesday remarks suggest the Bank views this repricing as doing some of the tightening work for it, calling the jump in market rates "some breathing space to assess." That phrase alone should be read as a delay signal, not a pivot.

Capital flow from this is traceable. Retail and institutional demand that had been accumulating in housebuilder shares — anticipating rate cuts that would ease mortgage affordability — is now unwinding, interpreted from the share price action: Rolls-Royce's near-crash of 19.4% from its March high correlates with this same rate-repricing as investors sold assets most sensitive to higher-for-longer funding costs. FTSE 100 housebuilders are not yet at a new low, but the Bank's refusal to endorse the price cap as inflation-killing removes the last credible argument for an imminent cut.

What Bailey did not address is the mechanism that forces the next move — and that is where the rate thesis connects to something more structural in London property.

London Housing's Structural Crack

Bailey's rate-delay signal explains why London house prices fell 2.1% in the year to March — but it does not explain why Westminster fell 11% and why Berkeley, London's largest housebuilder, announced it could "no longer invest" in the capital after planners rejected its 900-home Peckham scheme for the second time.

That second fact is a causal extension of the first. Higher-for-longer rates shrink the addressable buyer pool at the top of the market — which is precisely where London is most exposed. The ONS data shows four in five first-time buyers in London now pay stamp duty, versus one in ten outside the capital. Stamp duty eats three per cent of a purchase price in London on average, against one per cent nationally. When rates stay elevated, the premium the market charges for the capital compresses rather than expands.

The Treasury added a structural layer this week. A consultation document confirmed that housebuilders will face the full mansion tax — up to £7,000 a year — on unsold luxury new builds after 12 months of completion. Berkeley's exit is not an isolated grievance. Molior data shows just 2,103 residential starts in London in Q1 2026, against the 22,000 per quarter needed to meet the mayor's own targets. The Home Builders Federation called the mansion tax liability the variable that tips sites from viable to unviable.

Capital is not returning here passively. Foreign net selling in prime central London — Westminster, Kensington, Hammersmith — has accelerated since the non-dom regime ended, while domestic institutional buyers have not filled the gap, interpreted from the 7.5% and 7.4% respective price falls in those boroughs with above-average transaction times. Berkeley's announcement functions as the sector's loudest verification that the supply shortfall will worsen before it recovers.

What the housing data cannot yet tell you is whether this supply destruction is actually a floor or a further trigger — and the answer depends on a different chain entirely, one that is running through London's office market at exactly the same moment.

AI's Double-Edged City Trade

Nvidia is expected to report quarterly revenue approaching $79 billion after the US market closes tonight — and the way London's data stocks move tomorrow morning will reveal which side of the AI trade the City has decided to stand on.

The tension is sharpest in two FTSE 100 names. Experian fell nearly 7% on Wednesday, hitting the bottom of the index despite reporting what it called a "record year" and announcing a $1 billion capital return. The sell-off was not about the numbers — it was about the narrative. Matt Dorset at Quilter Cheviot put Experian's valuation at 18 times forward earnings, well below its 10-year average, and named the cause: a derating driven by AI disruption fears. If Nvidia tonight delivers another "beat and raise" quarter — the eighth consecutive beat the market is expecting — the AI capex story is confirmed, and that confirmation cuts both ways for Experian. More AI spending means more credit data demand in the near term; it also means more in-house capability among Experian's largest customers over the medium term.

Relx fell 3.5% on the same day for the same structural reason. The passive flows that had supported both stocks as AI beneficiaries are now reversing into cash or directly into Nvidia-adjacent names, interpreted from the simultaneous derating of two high-margin data businesses with otherwise sound fundamentals. Morgan Stanley's recent downgrade of Relx, citing AI start-up competition, is the analyst community confirming what price action had already flagged.

British Land offered a contrasting read. The FTSE 100 developer reported a 32% jump in profit before tax and 95% office occupancy — and CEO Simon Carter named the source of new demand explicitly: "high growth AI and innovation-led businesses." That is direct evidence that AI capex is creating physical demand inside London's commercial property market even as it destroys valuation confidence in London's data incumbents.

The leaning here tilts toward continued pressure on Experian and Relx in the near term. HSBC's CEO Georges Elhedery said Wednesday that AI "will destroy certain jobs and will create new jobs" — the same week Standard Chartered announced 8,000 back-office cuts. If Nvidia's results tonight show revenue above $79 billion with margins holding near 75%, the case for accelerated AI deployment across financial services strengthens, and the timeline for Experian's disruption shortens. The verification benchmark to watch is Experian's share price relative to the 2,557p close — a break below 2,500p on above-average volume in the next two sessions would confirm institutional sellers, not just day-traders, are repositioning. What would prove the bear case wrong is a Nvidia miss: if Jensen Huang's numbers disappoint tonight, the entire AI-acceleration thesis stalls, and Experian's 18-times forward earnings re-rates upward as the disruption timeline extends.