London, battleground of both driverless vehicle models

LondonTwo models of artificial intelligence. Two visions of autonomous driving. The automotive and technology giants are fiercely competing and, at the same time, converging on the same stage: London. The British capital, already immersed in pilot programs, is preparing to see driverless vehicles begin to make their way through its streets. The debate is inevitable: what future awaits flesh-and-blood taxi drivers?

Imagine it's Friday night on Charing Cross Road, in the heart of the West End. The audience is leaving the theaters, rain is sliding onto the asphalt, and a cyclist decides to cross outside the marked crosswalk. A double-decker bus brakes sharply to avoid a collision. In the middle of this perfectly recognizable scene at the start of any London weekend, a white vehicle from the Uber fleet is cautiously approaching. There are no hands on the wheel. The car detects the cyclist's hesitation, stops precisely, and the incident ends with just a scare.

The big news is that this vehicle doesn't rely on a fixed map. The system learns from its surroundings and processes the information in real time. This capability is the result of a multi-million dollar investment. Wayve, a start-up A British company founded in Cambridge in 2017 announced this week a €1.02 billion funding round, raising its valuation to €7.305 billion. This essential capital—contributed by companies such as Nissan, Mercedes-Benz, and Nvidia—is needed to sustain the computing power required to analyze data in milliseconds. Without this development, London would have to rely exclusively on technology imported from the United States or China. Uber and Wayve have begun a new phase of autonomous vehicle trials in the British capital, as part of the collaboration between the mobility platform and the British company. The tests are being conducted on real streets and use technology capable of managing driving autonomously under certain conditions. However, for the time being, the vehicles are being driven with a safety operator on board, in accordance with current regulatory requirements. Alternative ways of understanding the world

These initiatives are part of the UK's strategy to move towards the commercial implementation of so-called robotaxisIn this context, London has designated 20 of the 33 boroughs (boroughsThe City of London is being used as the main testing ground for Uber and Wayve vehicles. These areas combine the heavy traffic of the city center with narrow, complex streets and residential areas. The boroughs included in this phase are: City of London, City of Westminster, Camden, Islington, Hackney, Haringey, Waltham Forest, Tower Hamlets, Newham, Redbridge, Barking and Dagenham, Southwark, Lambeth, Lewisham, Greenwich, Hammersmith and Fulham, Kensington, and Chel. Although Waymo (Google) vehicles will also arrive in London in September 2026, their technology is almost the antithesis of Wayve's. And that's where the real battle is being fought: in the different approaches. Waymo has built its system on high-definition maps, which require scanning each street beforehand with millimeter detail. This method guarantees high accuracy but implies dependence on a pre-mapped environment. Faced with sudden changes – roadworks, detours, or unforeseen alterations – the vehicle can lose its bearings and reduce its operational capacity. Wayve, on the other hand, is betting on what is known as integrated AI (embodied AI), which does away with HD maps and learns to drive by observation, like a human. This approach allows driving in new scenarios without specific prior training, known as "driving zero-shot

The differences are also reflected in the cost and complexity of the machinery needed to make these vehicles possible. Waymo's models integrate advanced sensors such as the latest generation LiDAR, a laser technology capable of generating three-dimensional representations of the environment, but whose additional cost can reach around $100,000 per car. Wayve, on the other hand, advocates a minimalist approach based primarily on cameras and more affordable sensors. This simplification facilitates the integration of the system into production cars from manufacturers like Nissan or Mercedes, a key factor for commercial scalability. In short, it licenses its system to automotive giants.

Another distinguishing element is the interpretation of decisions. Waymo operates mostly under programmed rules and deterministic models. Wayve has developed LINGO-1, an AI model capable of verbalizing the vehicle's reasoning. Thus, the car can justify actions with phrases like "I'm braking because the pedestrian seems hesitant," for example. A development that could improve user confidence, facilitate incident diagnosis, and contribute to regulatory validation processes.

Also in Catalan?

Beyond safety, another issue emerges: language. The LINGO-1 model allows interaction with the vehicle. Although initial training is conducted in English, these systems are based on multilingual architectures. Technically, there are no impediments to passengers being able to communicate with the car in their own language, including Catalan, in the near future. The incorporation of this feature into the fleet operating in London will depend largely on Uber's decision to activate language packages in a market as globalized as the UK capital. On the other hand, the impact on the job market is another major unknown of autonomous driving. Nearly one million people in the UK—3% of the workforce—earn their living by driving. Uber finds itself in a paradox: the platform that generated thousands of jobs is now funding the technology that could reduce them.

Meanwhile, in cities like San Francisco and Beijing, robotaxis are already part of the urban landscape. When the final tests are completed after the summer, London is preparing to enter its year zero in 2027. Is there a way to reverse the journey towards a driverless city?