In April 2022, British autonomous-driving startup FiveAI ceased operations. Bosch acquired the business for an undisclosed sum, framing the transaction as an acqui-hire for talent rather than a continuation of FiveAI’s car-sharing effort. The London driverless trial that FiveAI had been developing with UK government backing never reached a commercial launch, and FiveAI’s self-driving stack did not reach the market under its own brand.
Before the acquisition, FiveAI had won a UK government grant through the Centre for Connected and Autonomous Vehicles (CCAV) to introduce a shared, on-demand driverless car service in London. The scheme, branded StreetWise, set out to prove that shared autonomous fleets could shrink private car ownership in cities, freeing up parking and cutting congestion. The trial was scheduled for the London boroughs of Bromley and Croydon, with Transport for London (TfL) handling the regulatory interface.
FiveAI operated from Cambridge and Bristol. It built its own self-driving stack from the ground up, refusing third-party systems. The StreetWise consortium pulled together several partners to test whether a driverless car-sharing operation could work on public roads with real passengers aboard.
The StreetWise Consortium and Government Funding
The StreetWise programme ran as a consortium backed by a grant from the Centre for Connected and Autonomous Vehicles, the UK government body that funds autonomous-driving research and on-road trials. FiveAI led the group, and partners handled distinct slices of the work. The exact grant figure for the StreetWise phase remains undisclosed, but the effort sat inside a wider CCAV funding envelope for connected and self-driving demonstrations across the UK.
FiveAI’s stated goal was to shrink private car ownership in cities through shared autonomous fleets. The logic: a pool of self-driving cars, available on demand, could replace several privately owned motors, reclaim parking real estate and ease traffic. The business model rested on per-ride payments, mirroring existing ride-hailing services but stripping out the human driver. A rider would book a vehicle through an app; the car would pilot itself to the pickup point; passengers would travel to their destination with nobody behind the wheel.
The consortium members brought expertise in autonomous driving, urban mobility, insurance modelling, and regulatory compliance. FiveAI contributed its own stack covering perception, planning and control. Other partners supplied the base vehicles, charging infrastructure, roadside sensing, and data-analysis horsepower.
Trial Scope and Technology
Why Bromley and Croydon
The trial was scheduled for the London boroughs of Bromley and Croydon. Planners chose these areas because they delivered a dense mix of urban and suburban driving: residential streets, arterial main roads, tricky junctions and heavy pedestrian activity near retail centres all sat inside a manageable geographic boundary.
The self-driving stack
FiveAI planned to deploy its own end-to-end self-driving pipeline, not bought-in components. That meant it controlled everything from sensor input to wheel commands. The system fused cameras, radar and lidar to perceive the world around it. The code was tuned for London’s messy traffic patterns: pedestrians, cyclists, double-decker buses and narrow Victorian-era streets. Before the StreetWise phase, FiveAI had already run its platform on public roads in London and other UK environments, gathering miles in multiple boroughs and weather conditions.
Fleet size and public access
The number of cars that operated during the London programme remains unconfirmed. Plans called for a small circulating fleet, but how many units actually ran on public roads under the StreetWise banner is not publicly documented. Likewise, no record has surfaced showing whether members of the general public climbed inside during the trial window. FiveAI had indicated passenger-carrying stages were in scope, but the final implementation details stayed opaque.
Regulatory Approvals and Safety Framework
The TfL interface
Transport for London sat inside the regulatory process from the start, working with FiveAI and its consortium partners to define the conditions under which self-driving cars could carry passengers on public roads. The UK regulatory architecture at that time demanded that operators secure clearance from both the Department for Transport and the relevant local roads authority before a single passenger could step aboard.
Proving safety
FiveAI needed to demonstrate its machines could operate safely across the trial zones. That safety case pulled together evidence from earlier testing, simulation runs across thousands of edge-case scenarios, and deep analysis of how the autonomy system behaved around cyclists, school crossings and emergency vehicles. The firm also had to detail its disengagement protocol: how the human safety driver would regain control, the latency involved, and the chain-of-command inside the cabin.
The trial was designed as a data-gathering exercise. FiveAI intended to feed real-world performance logs back into its stack to refine behaviour and to supply regulators with an auditable safety record. The company’s argument, repeated to ministers and the press, was that real-world passenger-carrying mileage was the only way to build public trust and prove that shared autonomous mobility could function in a dense, chaotic city such as London.
Outcome: Acquisition by Bosch
The end of the road
FiveAI’s London driverless car-sharing trial did not translate into a commercial service. In April 2022, Germany’s Bosch acquired the outfit in a deal that treated the startup as a source of engineering talent, not a going concern. The purchase price stayed secret.
What stopped
After the deal closed, FiveAI’s self-driving stack never reached customers under the FiveAI name. The StreetWise programme and the planned Bromley-and-Croydon car-sharing service were wound down. Bosch folded the FiveAI team into its own autonomous-driving division, where the engineers turned their attention to the parent group’s existing roadmaps and abandoned the car-sharing startup’s concept entirely.
The wider lesson
StreetWise’s journey from government-backed trial to acqui-hire captures the headwinds that confronted autonomous-driving ventures between the late 2010s and early 2020s. Building a safe, reliable system that can navigate a complex city without a human behind the wheel proved tougher and costlier than many startups’ original projections. FiveAI’s approach, a proprietary stack paired with a per-ride car-sharing model, demanded simultaneous investment in deep-tech R&D, fleet operations and regulatory negotiation. Without a line of sight to revenue, the operation became more valuable as a talent pool for an established automotive tier-one than as a standalone business trying to go it alone.
