-10.png&w=3840&q=75)
-10-2.png&w=3840&q=75)
WhatsApp us
For two years, prime central London has been the market everyone explained and nobody bought. That changed in the second quarter. Transactions above £5m rose 18% year on year across Mayfair, Knightsbridge, Belgravia and Chelsea, the strongest quarter since 2022, and the buyer register has shifted decisively international: dollar- and dirham-denominated purchasers accounted for a little over half of super-prime exchanges, against a third a year earlier.
What's new this cycle is how the top agencies are pricing into that demand. Several of the largest prime London agents have quietly rolled out AI-assisted valuation models over the past year, trained on decades of transaction, planning and renovation-spend data, to sanity-check the traditional comparables-led approach at a price point where comparables are often thin — there may be only two or three genuinely similar houses sold in a Belgravia garden square in five years. The models flag when an agent's instinct and the data-driven estimate diverge, rather than replacing the agent's judgement outright.
"At £20m a comparable from three years ago on a different garden square isn't really a comparable. The model gives us a second opinion built from a thousand data points instead of three anecdotes — we still make the call, but we make it with better information."
Agents report the return of a buyer type largely absent since 2016: the portfolio purchaser acquiring two or three units in a single quarter, and this buyer in particular is pushing agencies toward data-led pricing, since they are comparing London against valuation-model-driven markets like Manhattan and Singapore and expect the same rigour. Stock, not demand, is now the constraint — new instructions in the core postcodes fell 9% over the same period, and the best houses are increasingly trading off-market, sourced through the same data models flagging owners likely to sell.
The risks are political as much as financial. Non-dom reform continues to reshape who holds London property and through what structures, and any autumn move on council tax banding or stamp duty at the top end would test the recovery's depth. Our base case remains constructive: pricing power returns to vendors in the core squares first, and the gap to the 2014 peak closes from the top down.
