London's Triple Robotaxi Experiment Puts Business Models, Not Just Technology, on Trial

The arrival of Baidu's Apollo Go and Freenow's RT6 robotaxis on London's roads, reported by Automotive World, is being framed correctly as more than a third entrant joining a crowded field. With safety drivers aboard and public rides pencilled in for 2027 subject to regulatory approval, the deployment sits alongside Waymo's standalone operation and the Uber-Wayve partnership to give London something no other city currently has: three structurally distinct robotaxi business models operating under a single regulator at the same time.

That structural divergence is the real story for automotive leaders. Uber and Wayve represent the partnership model, pairing a dominant ride-hailing platform with an AV specialist that supplies the technology stack. Waymo has chosen the opposite path, running its own app with fleet partner Moove and betting that riders will download a dedicated service rather than dispatch through an incumbent one, a bet made harder by its recent, well-publicised break with Uber and the wind-down of that arrangement in Phoenix, Atlanta and Austin. Freenow and Baidu occupy a third position: a licensed private-hire operator with deep London roots pairing with a purpose-built Chinese hardware platform. Each model carries its own distribution logic, and London will now generate genuine comparative evidence on which one actually works when novelty wears off and riders start comparing fares.

Cost economics deserve particular attention. Automotive World cites Apollo Go's RT6 build cost at roughly US$35,000-37,000, against an estimated US$100,000-plus for Waymo's sensor-heavy Jaguar I-Pace retrofits. If that gap holds once Baidu's fleet scales, it points to a familiar automotive lesson applied to autonomy: purpose-built platforms tend to beat retrofits on unit economics, however much of a head start the retrofit approach bought in development time. For OEMs and suppliers watching from the sidelines, the question is less about who wins in London and more about which manufacturing philosophy the winning economics ultimately reward.

Distribution, however, still favours incumbency. Uber's existing dominance in London gives Wayve access to a far larger pool of habitual riders than either Freenow or Apollo Go can currently draw on, even though Freenow operates across 180 European cities and Apollo Go brings more than 22 million cumulative rides and a 27-city global footprint, most of it accrued in China. Freenow's advantage may lie elsewhere: its roots in London's licensed taxi and private-hire ecosystem could translate into a steadier regulatory relationship with Transport for London than Uber has managed, given the latter's history of licensing disputes with the authority.

The regulatory dimension is where London's significance extends well beyond its own streets. This is the first time Chinese autonomous vehicle technology has been tested in the city, and Transport for London together with the Centre for Connected and Autonomous Vehicles are likely to apply a level of scrutiny on data routing, cybersecurity and mapping compliance that neither Waymo nor Wayve will face in the same way. That scrutiny arrives against the backdrop of an unexplained incident in Wuhan in which dozens of Apollo Go vehicles reportedly stopped functioning simultaneously, prompting Chinese authorities to pause new robotaxi permits temporarily. How London's regulators respond, and whether Hong Kong's recent first driverless Level 4 permit for Apollo Go proves a reliable dress rehearsal, will do more than decide Baidu's UK prospects. It will set the template other Chinese AV operators can expect as they eye Western Europe, making London's verdict on Apollo Go a signal well worth tracking regardless of how the London business itself performs.