Wayve and the future of autonomous vehicles

The Automated Vehicles Act received Royal Assent in the UK on 20 May 2024. This leaves Wayve poised to take the lead in providing high-level "self driving" technology to car manufacturers and bring autonomous vehicles to UK roads.

The guidelines set out by the Automated Vehicles Act, and the Implementation Programme that followed, lay out the testing procedures to ensure that vehicles are deemed to be as good as or even better than a ‘careful and competent’ human driver.

Understandably, many people in the UK are hesitant to place their physical safety—and that of their friends and family—in the ‘hands’ of an unfamiliar computer system. Although ADAS (Advanced Driver Assistance Systems) in various forms have been around for several years, having cars handle more complex scenarios, such as lane changes and roundabouts independently, is new territory for most of us. The reliability and accuracy of these systems must be proven in the real world before they gain widespread trust.

One of the key reasons for developing autonomous vehicles is the enhanced safety they offer, as software does not get fatigued or distracted. Driver error is the leading cause of road traffic accidents. A study by AAA examining the reduction in crashes and injuries/deaths through existing systems showed that emergency braking, forward collision assist, and lane keep assist/departure warning had the potential to prevent 40% of passenger-vehicle collisions. This was based on the ADAS systems available at the time of the study in 2018. With more advanced systems like those being developed by Wayve, the potential for preventing crashes, injuries, and deaths could be significantly higher.

Unlike some other self-driving systems, which rely on lines of code written by programmers to dictate actions (for example, if the traffic light is red, then the car stops), Wayve employs end-to-end learning with neural networks. This approach teaches the system by showing it how to perform tasks using videos and simulations. It learns from what it is shown and produces outputs for steering, acceleration, braking, etc. This method is better suited for handling edge cases that the car has not encountered before, as it can infer appropriate actions from its learning, unlike traditional systems that may struggle with unanticipated scenarios.

Wayve has developed a leading simulation software called GAIA which allows them to recreate routes driven in photo-realistic detail and even create scenes that the car has never ‘seen’ before. This allows them to simulate and test edge cases safely and put the system under pressure that may not be safe to replicate in real life. Additionally, this method allows them to train the model on significantly more driving data than would be feasible through real-life testing alone in a reasonable timeframe.

Initially, it will be a requirement for the driver to be ready to take over should the system need it. Eventually, when regulations allow, the system may be what's referred to as 'hands-off, eyes-off" and the driver would be able to do other things such as read a book, and only be required to take over occasionally with sufficient notice given to ensure this happens safely.

Although many are sceptical of self-driving technology, it has benefits including the ability to handle most of the driving when you're tired after a long day at work. When the system is advanced enough, it could handle taking passengers who otherwise couldn't drive due to injury, disability or other reasons, to and from their destinations. This would make owning a personal vehicle more accessible to those that may be limited at the moment. If the systems were to get to a point where intervention would not be required, then it would be possible to sleep in your car and wake up at your destination, send the car to pick up friends, or even hire the car out as a robo-taxi of sorts and it can earn you money.

These are all benefits that any company with competent self-driving technology would reap, however Wayve particularly would see this. Instead of making the cars themselves and only fitting this revolutionary software to their own vehicles, they're making a system which is generalisable in that it can be installed on any vehicle and adapted so that it drives that vehicle. This means that any vehicles, with the necessary sensors, made by the OEMs that partner with Wayve would then be able to drive themselves. As at early April 2025, Wayve is yet to confirm which brands they are partnering with, although their recent expansion to Germany seems to be in an effort to be closer to German automakers as potential partners.

Wayve has previously expressed that their plan is for their systems to start at Level 2, which would be similar to standard ADAS that some modern cars have. They would then update this over the air (OTA) when it's ready and the car would be capable of driving more independently.

Wayve was founded in 2017 by Alex Kendall and has since become one of the fastest hiring tech startups within England. In the past twelve months they've hired around 200 people. Wayve expanded into the US with a San Francisco office in October 2024 and have found that the system adapted impressively quickly to driving on the other side of the road, handling traffic lights slightly differently with being able to turn right on red lights, and the other nuances that come with driving in America. They then expanded into Germany in March of 2025.

As a company they are hesitant to give timelines, although these past few months Wayve have expanded abroad, grown their company, and are currently in negotiations with automakers. I know I will certainly continue to follow their progress closely and eagerly await the big announcement of the integration of their software into vehicles for public use and the commencement of sales of these embodied AI vehicles.