Scuter has closed a €3.5 million seed extension funding round as the Italian mobility startup now enters a new phase of business.
Founded in 2015, Scuter has developed a three-wheeled electric vehicle designed to be ridden on the road. The company operates on a leasing model, renting vehicles out directly to customers, supplying fleets to organisations and partnering with cities and dealers to rent vehicles out locally while remaining under the Scuter platform.
Having now completed the second and final tranche of its seed extension financing round, valued at €1 million, Scuter concludes an eighteen-month investment phase that has been dedicated to commercially and operationally validating the vehicle and business model.
Next, the company is turning its attention to autonomy.
“Today, shared mobility vehicles spend much of their time parked in the wrong place or unavailable when demand exists elsewhere,” Scuter CEO Gianmarco Carnovale told Zag Daily.
“Autonomous repositioning allows vehicles to move where demand is emerging, reducing waiting times for users while significantly increasing vehicle utilisation. Higher utilisation means more rides per vehicle, lower operating cost per ride and ultimately lower prices for customers. We see a real opportunity to arrive somewhere very close to the ticket’s cost of a mass transit ride.”
Scuter spent the first €2.5 million tranche of seed financing when it arrived last year. Over the last eighteen months, the company has invested in designing and developing its fleet of shared vehicles, securing EU type approval and beginning pilot operations in Milan.
A commercial pilot of 200 vehicles is currently operational in Italy’s fashion capital, with Scuter reporting 1,900 active riders, zero accidents and an average ride distance of 3.8 kilometres.
Scuter’s vehicle has a width of 78cm and weighs 145 kilograms, reaching an extended range of 90 kilometres. It has three wheels for stability and is advertised as not requiring a helmet.
The first phase of business, Carnovale explained, has been about commercially and operationally validating the vehicle itself before the focus shifts to autonomy. This is where he believes much of the industry has gone wrong, arguing that many players have revolved their product around autonomy before proving that the “basic foundations” of the vehicle itself can commercially scale.
Scuter’s approach to autonomy flips the model. Carnovale believes the company has validated its shared mobility vehicle as a non-autonomous product, with Scuter now prepared to add the Physical AI layer to provide autonomous capabilities.
“Rather than starting with autonomy and searching for a viable commercial model afterwards, we deliberately chose to validate the vehicle, operations and business model first before introducing the AI layer,” Carnovale said.
The AI layer isn’t confined to mobility either, Carnovale explained, instead highlighting that Scuter sees itself as developing a Physical AI platform whose first application “happens to be urban mobility” but can be extended to urban infrastructure more broadly.
Whatever the application, Carnovale believes that the vital ingredient for Physical AI is commercial feasibility.
“We don’t believe leadership in Physical AI will be defined by who demonstrates the most impressive prototype,” he said.
“We believe it will belong to those who build the most effective physical platform, industrialise it successfully, earn genuine customer adoption and continuously compound operational intelligence through real-world use. Our ambition is not simply to make it possible but to make it economically scalable for the mass market.”
Adding the autonomous capabilities
Scuter’s focus is now on integrating the Physical AI layer that will give its fleet the autonomous capabilities.
Carnovale envisions Scuter’s vehicles autonomously rebalancing themselves to even out distribution and having a vehicle drive itself directly to the customer rather than requiring the customer to locate the vehicle. The aim is to lower operational costs from field team dispatching which, in turn, means lower ticket prices for users.
“This is one of the key differences between Scuter and most autonomous mobility projects,” Carnovale said. “Many robotaxi companies are pursuing premium transportation services. Our ambition is to use autonomy to make shared urban mobility affordable enough for mass-market adoption.”
While the seed extension period was focused on completing product validation, Scuter has also begun laying the foundations for autonomy. The company has defined its Physical AI architecture, developed a phased development roadmap in line with budget requirements and began engaging specialist talent, with an upcoming funding round aimed to accelerate AI development.
In the first phase of Physical AI integration, the company will retrofit its existing vehicles with onboard GPUs, sensors and electronics required for perception. At this stage the vehicles will not drive autonomously but will be able to perceive their surroundings, with Carnovale citing capabilities such as verifying parking conditions, performing biometric driver identification, collecting operational data and building the dataset that will train future autonomous capabilities.
Customer behaviour will firstly be mapped in Scuter’s current operational fleet in Milan, with the vehicles providing the dataset that will train the Physical AI model whilst simultaneously generating revenue for Scuter.
“Unlike many autonomy projects, we’re not building an expensive dataset before having a business,” Carnovale said. “The business itself will generate both revenue and the dataset.”
Stage two will then introduce the hardware required for autonomous motion itself, including actuators controlling steering, braking and vehicle stabilisation. From that point onwards, vehicles will be physically capable of operating without a driver onboard.
The company plans to build on open-source models and integrate them into its own architecture, with Scuter’s current technology direction based on NVIDIA’s Alpamayo autonomous stack.
“Our vision is supervised autonomy,” Carnovale said. “Vehicles will operate autonomously in normal conditions while a remote operator will intervene only in exceptional or uncertain situations where the AI requests human supervision.”
According to Carnovale, there’s no plan in the pipeline for Scuter’s vehicles to travel autonomously with a rider onboard – and this is simply not the mission.
“We don’t see autonomy as the product,” he explained. “We see autonomy as an operational layer that makes urban mobility dramatically more efficient.
“In other words, our ambition isn’t to build another autonomous vehicle. It’s to build one of the first commercially scalable Physical AI platforms for urban fleet orchestration.”
