Sadiq Khan Wants a Robotaxi Cap. The Real Fight Is Over Who Regulates London’s Streets

Sadiq Khan's intervention on robotaxis marks a rare moment of municipal resistance inside an otherwise accelerating national programme. The London Mayor has asked for a cap on the number of driverless vehicles permitted on the capital's streets, precisely as the UK government prepares legislation allowing fully autonomous operations.

His argument is not opposition to the technology itself, but a demand that deployment velocity match regulatory capacity. That distinction matters enormously, because it reframes the debate from whether robotaxis should exist to who controls their arrival.

The timing is instructive. Uber already runs limited self-driving rides in London using Wayve technology, with a safety driver present, under a Transport for London licence. Freenow by Lyft intends to launch fully driverless Baidu vehicles as early as next year through a national licensing scheme, while Waymo has signalled interest in the UK market.

Roughly 24 Baidu Apollo Go cars are being tested in Brent. Each of these developments arrived before the regulatory architecture that would govern them at scale, which is exactly the sequencing Khan criticises.

The structural asymmetry at the heart of the dispute is jurisdictional. Transport for London can approve or reject fully autonomous vehicles, but it cannot manage or license robotaxi services. Ministers, by contrast, hold powers to accelerate rollout nationally.

Khan wants additional authority devolved to City Hall so that London can set volume limits, conditions and service standards. The conflict therefore concerns governance design as much as traffic management, and it will likely set a precedent for other British cities watching closely.

TL;DR Sadiq Khan has called for a cap on robotaxi numbers in London as the UK government prepares to permit fully driverless vehicles. He warns that rapid, unregulated deployment could threaten driver livelihoods, worsen congestion and affect air quality, and argues London needs devolved powers to license and limit services. Uber, Freenow by Lyft and Waymo are all positioning for the market, while driver unions protest. The dispute is fundamentally about who governs autonomous mobility at scale.

The Regulatory Vacuum Behind London's Robotaxi Debate

Autonomous vehicle policy in Britain has developed in layers that do not yet interlock. National legislation is being drafted to permit fully driverless operation, while local authorities retain only binary approval powers over individual vehicle types.

The result is a framework in which ministers can liberalise supply but city governments cannot calibrate demand, geography or service volume. Khan's request for a cap is, in practical terms, a request to close that gap before commercial operators exploit it.

Why the Current Powers Are Insufficient

Transport for London's mandate covers safety approval, not commercial licensing. It can determine whether a vehicle is fit to operate autonomously, but it cannot decide how many such vehicles may run, where they may rank, or what service obligations attach to their operation.

Private hire licensing rules were written for human drivers and do not map cleanly onto fleets managed by remote supervisors and software updates.

This limitation becomes acute when operators can obtain national licences that bypass municipal conditions. A company licensed nationally could, in principle, deploy vehicles in London without TfL controlling fleet size or geographic distribution.

Khan's warning about a "big bang without proper regulation" describes precisely this scenario: supply arriving faster than the city's ability to shape it.

The comparison with San Francisco is frequently invoked but imperfectly understood. San Francisco permitted robotaxi expansion and subsequently confronted congestion incidents, emergency vehicle interference and public backlash, prompting regulators to restrict fleet growth.

London's density, narrow Victorian road geometry and near-ten-million population make those risks more pronounced, not less, which strengthens the case for pre-emptive volume controls.

The Operators Already in Position

Three distinct commercial pathways are visible. Uber's arrangement with Wayve uses a TfL licence and retains a safety driver, positioning the company as a gradualist. Freenow by Lyft plans to import Baidu's Apollo Go platform under a national scheme, targeting fully driverless service next year. Waymo, with extensive American operating experience, is evaluating entry without a confirmed timeline.

Baidu's Brent trials provide the most concrete evidence of technical readiness. Around two dozen Apollo Go vehicles are being tested, and the company reports fully autonomous commercial service in Chinese and Emirati cities.

Independent observation during October test drives suggested competent handling of London streets, though human intervention remained common around roadworks and complex junctions.

That intervention rate is the critical variable regulators will scrutinise. A vehicle that requires remote assistance at roadworks is not equivalent to one operating unsupervised through the same environment. The gap between demonstration and dependable service defines the risk window, and it is within that window that Khan wants volume limits applied.

Labour Market Exposure and Union Pressure

Driver unions have protested outside the London headquarters of Wayve and Waymo, and the anxiety is rational rather than reflexive. London's private hire sector employs tens of thousands of drivers whose income depends on trip volume.

A capped, gradual rollout preserves transition time; an uncapped one compresses adjustment into a few years, with limited retraining infrastructure in place.

Khan explicitly linked autonomous deployment to the livelihoods of private hire and black taxi drivers. The black cab trade carries additional cultural weight, and its knowledge-based licensing model is difficult to reconcile with algorithmic navigation. Any credible policy must therefore address compensation, retraining and licensing reform, not merely vehicle counts.

Operator rhetoric emphasises partnership. Freenow's UK and Ireland general manager, Danny O'Gorman, said the company is working closely with TfL and wants to proceed properly, while conceding that post-launch regulation remains unresolved. That concession is revealing: even the operators acknowledge the governance model is incomplete.

Congestion, Air Quality and the Urban Case for Restraint

Khan's environmental argument deserves scrutiny because it cuts against the industry's usual framing. Robotaxis are marketed as congestion-reducing shared mobility, yet early evidence from American deployments suggests empty repositioning miles and induced demand can offset those gains.

A fleet that circulates while awaiting fares adds vehicle kilometres without carrying passengers, which is precisely the pattern London's congestion charge was designed to discourage.

The Empty-Mile Problem

Autonomous fleets optimise for availability, not for kerb space. Vehicles may circle rather than park, because parking costs money and waiting does not. In dense districts, that behaviour compounds congestion at exactly the moments when road capacity is scarcest. Without volume caps or kerb pricing, operators have no commercial incentive to internalise that cost.

Air quality follows the same logic. Electric robotaxis eliminate tailpipe emissions but not particulate matter from tyres and brakes, nor the congestion effects that concentrate pollution at street level.

London's narrow roads trap pollutants in ways that San Francisco's grid does not, which makes the case for measured deployment stronger rather than weaker.

Khan's framing of gradual change is therefore technically defensible. A phased rollout allows regulators to measure actual vehicle kilometres travelled, observe kerb utilisation and adjust caps accordingly. A sudden liberalisation removes that feedback loop, and correcting afterwards requires the politically difficult step of withdrawing permissions already granted.

Learning From San Francisco's Experience

San Francisco functions as the natural experiment. Robotaxis expanded rapidly, then encountered incidents involving emergency vehicles, construction zones and stalled vehicles blocking transit routes. Regulators responded by restricting fleet size and requiring reporting. The sequence, expansion followed by correction, is exactly what Khan wants London to avoid by sequencing regulation first.

The counterargument holds that San Francisco's problems reflected early technology rather than structural flaws, and that intervention rates have fallen as systems matured. That may be true, but it is a claim operators must evidence in London conditions, with its roadworks, bus lanes and complex junctions, before volume limits are lifted rather than after.

Air Quality Claims Under Examination

Electric drivetrains remove exhaust emissions, and that is a genuine gain in a city with nitrogen dioxide problems. But congestion-related pollution persists, and induced demand can increase total vehicle kilometres.

The net air quality effect therefore depends on fleet utilisation rates and whether robotaxis substitute for private cars or for public transport journeys.

Substitution risk is underappreciated. If autonomous fares undercut buses and tubes, passengers may shift from high-occupancy to low-occupancy modes, raising emissions per passenger journey. Khan's concern about air quality is thus not rhetorical; it reflects a plausible modal shift that volume caps and pricing could mitigate.

The Political Economy of Who Controls Autonomous Mobility

Beneath the specific dispute lies a broader question about the distribution of authority in technologically disrupted sectors. National governments typically liberalise to attract investment, while cities absorb the operational consequences.

Khan's demand for devolved licensing powers is an attempt to rebalance that arrangement, and it will be contested because it directly constrains the pace at which capital can deploy.

Investment Signals and Operator Patience

Freenow's statement that the national framework is an important investment factor is a signal, not a threat. Operators need regulatory certainty to commit capital, and fragmented municipal rules raise costs. But certainty can be delivered through clear national standards that explicitly permit city-level volume conditions, rather than through pre-emption of local authority.

The design choice matters. A framework granting cities licensing power while maintaining national safety standards would satisfy both investment needs and local accountability. A framework that centralises licensing nationally would accelerate deployment but concentrate risk, and would likely generate the kind of backlash that ultimately slows adoption more severely.

What a Workable Settlement Might Look Like

A credible settlement would combine national safety certification with municipal service licensing, volume caps tied to measured performance, mandatory reporting of intervention rates, and a transition fund for displaced drivers.

Each element addresses a specific failure mode identified in the London debate, and together they convert an adversarial standoff into a managed sequence.

Khan's call for a cap is best understood as an opening negotiating position rather than a permanent ceiling. The substantive demand is for London to hold the levers, so that expansion tracks evidence rather than ambition.

Whether ministers concede that principle will determine not just robotaxi volumes but the precedent for governing automated systems in British cities.

Precedent Beyond London

Other British cities face identical structural constraints, and most lack London's administrative capacity to negotiate. If the capital secures devolved licensing powers, comparable authorities will seek parity, producing a patchwork of local conditions. If it does not, national liberalisation will apply uniformly, and local objections will have no operational effect.

The wider implication concerns automated systems generally. Delivery robots, autonomous shuttles and eventually personal autonomous vehicles will raise the same question: whether cities can shape deployment or merely react to it.

London's robotaxi dispute is the first substantial test, and its resolution will inform how Britain governs automation in public space for years.

Related By Tags

• 01
  France's Third Weekend of Unrest and the Wider Test of Institutional Patience

  WORLD NEWS & EVENTS
• 02
  Okpebholo's New Nickname and the Economic Case He Is Making for Tinubu's Reforms

  POLITICS
• 03
  Ikogosi Marathon Bars Sowore on Race Day as Sanctions Demand Reaches World Athletics

  POLITICS
• 04
  Sanwo-Olu Walks the Lagos Fashion Week Runway and What It Signals for Nigeria's Creative Economy

  ANALYSIS
• 05
  Ardern Ties Public Trust to Financial Security in Sydney Address

  POLITICS
• 06
  Thousands Protest in India's Capital Over Voter Roll Revision

  POLITICS
• 07
  Election Officials Face Foreign Threats With the Weakest Federal Pipeline Since 2016

  POLITICS
• 08
  Netanyahu's Coalition Arithmetic After Israel's Inconclusive Election

  POLITICS