What Could London’s PHV Fleet Look Like in 2031? Three Scenarios, Built From What’s Actually on the Record

TfL has told LDV directly that it holds no projection of how many robotaxis it expects to licence, or what happens to driver numbers, over the next five years. Nobody else has published one either. So this is LDV’s own attempt to build one — not a forecast, but three clearly labelled scenarios, built only from numbers already on the public record: today’s 15-vehicle trial, the regulatory milestones actually scheduled, and the growth framework the AV industry itself has already disclosed. Every assumption is stated. Where we don’t know something, we say so, rather than guess.

The starting point, precisely

As of August 2026: 94,258 licensed private hire vehicles in London (TfL, via FOI), roughly 110,000 licensed PHV drivers, and 15 licensed autonomous vehicles — a single one-year trial, Uber and Wayve’s, with a safety driver in every car. Robotaxis currently represent roughly 0.016% of London’s licensed PHV fleet. Two further regulatory milestones are already scheduled, not speculative: the Automated Vehicles Act 2024 is expected to be fully implemented in the second half of 2027, unlocking the DVSA licensing route for genuinely driverless operation; and Waymo has stated a target of a fully driverless commercial launch in Q4 2026, though it has not disclosed the fleet size that launch would involve.

Uber’s own framework, applied to the next five years

Uber’s own white paper, cited throughout our AV coverage, lays out a three-phase model for how it expects robotaxis to scale on its platform globally: Phase 1, “establish a foothold” — a baseline supply alongside human drivers in a handful of markets, which is where London sits today; Phase 2, “scale in dense metros” — which Uber’s own paper describes as unfolding “over the next five to ten years,” as costs fall and regulation allows; and Phase 3, “AV-heavy markets,” which Uber’s paper places in “the long run,” beyond even a ten-year horizon. By Uber’s own stated timeline, a five-year projection from today lands somewhere inside Phase 2 at best — meaningful growth from today’s near-zero base, but nowhere close to AVs becoming a dominant share of London’s PHV fleet.

Three scenarios for 2031

These are LDV’s own scenarios, built from publicly disclosed figures and stated corporate timelines, not projections issued by TfL, Uber, Waymo, Bolt or any government body. TfL has confirmed to LDV that it holds no five-year projection of its own. Treat the ranges as illustrative bounds, not predictions.

What this could mean for drivers — and the honest limits of saying so

Here the caution has to be greater still. Uber’s own white paper states that in California, one AV does the work of about four human drivers — but that is a California-specific figure describing a mature, high-utilisation AV market, and LDV asked Uber directly whether it applies to London. Uber did not answer. Applying that ratio mechanically to any of the scenarios above — for instance, treating 15,000 AVs as equivalent to 60,000 displaced driver-jobs — would be exactly the kind of unsupported extrapolation this article is trying to avoid, since nothing confirms the ratio holds outside the market it was measured in, and AV trip volume does not translate one-to-one into driver income lost, given some drivers work part-time, some AV trips may represent genuinely new demand rather than displaced trips, and TfL holds no data on any of this either.

What can be reported honestly is what London’s own unions have already estimated, independently of any of the scenarios above. ADCU has told LDV it estimates driver displacement could cost London £3.3 billion a year in lost driver earnings — a figure ADCU has called on TfL and Uber to formally assess and which, as of publication, nobody has publicly disputed. At GMB’s national congress, delegates warned that robotaxi adoption could put as many as 300,000 jobs at risk across the UK as a whole, not London alone. Those are the unions’ own published estimates, not LDV’s, and they are not tied to any specific five-year timeline — but they are the only quantified driver-impact figures currently on the public record from anyone.

The actual finding

The most defensible conclusion from this exercise isn’t a number — it’s the size of the gap. Even under LDV’s own high scenario, built by stretching every currently disclosed corporate ambition to its most aggressive plausible pace, autonomous vehicles would still represent a minority of London’s PHV fleet by 2031. Under the medium scenario — the one most consistent with Uber’s own stated timeline — AVs would still be a small fraction of the market five years from now. The five-year risk to drivers is therefore not that AVs will dominate London’s streets by 2031; on the evidence available, they almost certainly won’t. The real five-year risk is regulatory: that TfL enters this period, as it has entered this year, without ever having built the data — driver numbers, licence-volume forecasts, an economic impact assessment — that would let it, or anyone else, tell the difference between the low scenario and the high one until it has already happened.

This article’s scenarios are LDV’s own analysis, built from TfL’s FOI response on licensed PHV numbers, TfL’s direct statements to LDV regarding driver and licence projections, Uber’s policy white paper “Unlocking the Promise of Autonomy,” and corporate disclosures reported by TechCrunch, the Financial Times and others cited throughout LDV’s AV coverage. Union estimates are attributed to ADCU and GMB respectively and are not LDV’s own figures. These are scenarios illustrating a plausible range based on stated assumptions, not predictions, and should not be read as forecasts of what will occur.