Last week I visited the Cambridge Connector, an autonomous bus trial powered by Fusion Processing and Alexander Dennis (ADL), running a 20 minutes fixed route between a local Park & Ride and a business park.
The ride was OK (more on this later) - but top of mind for me was how do we commercialise autonomous buses. Daniel Clarke, Head of Technology and Innovation at the Greater Cambridge Partnership, explained that past and existing Connector trials focused on validating the tech and use cases; the next Pathfinder (/UK government) grant, planned to begin in September 2026 and continue for two years, will explore commercialisation.
Autonomous buses bring many advantages: No driver means lower operating costs and extended work hours, including nights and weekends; higher utilisation equals better service, bringing societal advantages; there are also lower maintenance and energy costs brought on by efficient driving, Fusion & ADL quoted 7% savings; add to that potential safety advantages.
Cities & governments could benefit from these advantages - by either ‘just’ saving money or by operating more services, in places where it wasn’t profitable enough to operate before. For example, the existing Cambridge route was chosen BECAUSE it wasn’t profitable and therefore wasn’t served, although it connects a Park & Ride to a business park. A societal justification (take cars off the road) takes place.
Yet global adoption of autonomous buses has been slow. A vicious cycle exists: with extremely very few public transport authorities (PTAs) willing to budget scaled autonomous bus operations, public operators are not buying in; this in turn makes it hard for tech companies in the field - such as Fusion, Imagry and ADASTEC, to raise funds and finance product development. And this, in turn, makes the solution less attractive to PTAs…
Global commercialisation of autonomous buses has been limited; see my June 2025 article for places who have been able to commercialise:
To commercialise autonomous buses we need to take the driver out. What’s stopping us?
1) People, or riders. Taking the driver away generates new challenges, such as ensuring the safety of elderly people who take time to board the bus and take their seat; offering a ramp for disabled people; and facing anti-social behaviour and sexual harassment concerns. By design, public transport is aimed at serving all, especially vulnerable people, which means taking the driver out isn’t so simple. In addition, there is the issue of revenue protection - how can we be sure people are paying?
2) Regulation. In the UK, the law mandates a safety driver on public roads; Jim Hutchinson, Fusion Processing's CEO, tells me they expect this to change by H2/2027. Until then the driver cannot be taken out, hence no commercial viability. I brought the UK as an example - but each of you think for yourself - “does my country allow driverless bus operations?”
3) Road Infrastructure is not ready. Autonomous buses cannot break the rules (yet), and this leads to hindering response to edge cases. A sharp turn in which the bus needs to slightly cross a white line to pass becomes impossible for an autonomous bus. Cyclists require 1.5 meter distance from behind, but that means that sometimes autonomous buses could be stuck for long periods of time unable to pass, with incoming traffic on the other lane. There are many use cases we never even think about, because human drivers solve those instinctively. Those come to life in trials.
Without solving the above (& probably other) challenges - commercialised operations will remain a distant future. There are some solutions:
Tech maturity is not slowing the adoption & commercialisation of autonomous buses. This has to do with buses operating on fixed routes, i.e. limited ODDs, which makes it easier for autonomous technologies to prepare for. Of those many routes, some are simpler than others, which means PTA/Os can gradually implement an autonomous service on a per-route basis. This helps to solve challenge #3 above.
“Not one ‘driver’ size fits all”: different communities and different routes can have different solutions - some services will be fully autonomous, while others could have a staff member onboard to assist passengers, either fully onboard or only at specific times. For example, a service can run autonomously during the day, and transition to include a staff member (not a driver) at night. The level of remote interference can also vary, from a centralised control centre based in a different city, to multiple ‘rider-assistance stations’ throughout the city, aimed at supporting riders at short notice. This mitigates - not fully solves - the challenges raised in #1 above.
#movingpeople is a part of Mobility Business - a consultancy focused on Innovation, Growth and Autonomy in the Mobility industry.
The ride: it was OK.
Reasonably boring. Ask anyone in autonomy, the story repeats itself: whenever people get on an autonomous bus, they're excited, taking pictures, looking at this 8th wonder. Fast forward 5 minutes later, everyone is on social media, not minding the bus. Autonomy is good enough for people to not care about it.
Not 100% autonomous throughout the route. There are two parts of the route in which the driver is designed to take control - one is a S-bend that requires briefly crossing a white line; the other a tight corner with visibility limitations. By current safety design, the bus just isn’t able to complete autonomously. You may ask yourself why then this route was chosen? Well, human drivers manage those two parts intuitively without a worry, so it didn't surface when the route planning took place. To Fusion’s defence - route design was determined before they entered the consortium.
Cyclists cause trouble. During the ride I noticed that when encountering a cyclist on the road, the driver immediately took control. It turns out that this doesn’t happen every time, only when there is incoming traffic (such as was the case). The reason is that in autonomous mode and with incoming traffic, the bus may find itself stuck behind a cyclist for long, as it must obey the 1.5 meter distance from the cyclist at all times. See point #3 on road infrastructure.
Hard breaks every now and then. There were several instances during the ride; Jim Hutchinson tells me that this is a known issue and that they are working to fix this.
Thank you to Daniel Clarke and Jim Hutchinson (mentioned in the post) and also to Jamie Wilson, Head of Concepts & Advanced Engineering, and Stefan Baguette, Group Marketing Communications Manager, both from Alexander Dennis.
For more on the ride you can read:
Neil Kennett from Cars Of The Future shared the ride with me and lived to write about it: https://carsofthefuture.co.uk/uk-self-driving-enviro100aev-bus/
Roger French took the bus on March 5th, and was much more impressed than he is with DRT: https://busandtrainuser.com/2026/03/05/more-autonomous-bus-travels-in-cambridge/
Like what you’re reading? Please share #movingpeople with your friends & colleagues. Every like and share matters. And if you’re not already a subscriber, now it the time to join:
