Connected Vehicle & Mobility Observatory, seventh edition: mobility beyond the private car

Interview by Paolo Barbato with Giulio Salvadori, Director of the Connected Vehicle & Mobility Observatory at Politecnico di Milano.

In the United States, autonomous vehicles carry out more than 450,000 commercial rides a week, in China around 250,000. In Europe, over the same period, 36 pilot projects have taken place and no large-scale commercial service. This is the gap that opens the seventh edition of the Connected Vehicle & Mobility Observatory at Politecnico di Milano, presented in May 2026 with a title that is also a diagnosis: “Driving the Future: mobility at the decisive bend”.

The bend is the one in which technologies stop being promises and become services. To read the results beyond the headlines, we spoke with Giulio Salvadori, Director of the Observatory, who led the research on connected vehicles, autonomous driving, smart mobility and charging infrastructure.

The robotaxi before the private car

Paolo Barbato: Giulio, let’s start with the most striking figure: the United States and China are running hundreds of thousands of autonomous rides a week, while Europe is stuck at pilots. How structural is this gap?

Giulio Salvadori: It’s a real gap, and for now it’s widening. China and the United States maintain a structural advantage: 67% and 44% of already-commercialized level 4 projects are Chinese and American respectively. But the most instructive figure isn’t quantitative, it’s the time it takes to bring a project from launch to commercialization. In China it’s 2.7 years on average, in the United States 6.4, in Europe 6.6. The European delay doesn’t stem from technology, it stems from regulatory waiting times in the go-to-market phase. We surveyed 47 level 4 autonomous driving projects globally, and the dominant service model, in 53% of already-commercialized cases, is the robotaxi. This confirms something worth stating clearly: the first scalable application of autonomous driving won’t be the private car, it will be autonomous ride-hailing. In terms of cost and profitability, the privately owned autonomous car lags behind.

Paolo Barbato: And in Europe, is something really moving or are we just watching?

Giulio Salvadori: Something is moving, even if in scattered fashion. The fastest markets are the United Kingdom and Switzerland, both outside the Union, which have built more agile regulatory frameworks. The first commercial robotaxi service open to the public within the European Union arrived on April 8, 2026, when China’s Pony.ai launched an operational service in Zagreb. Waymo has announced a launch in London by the end of 2026, and in the same weeks Baidu Apollo started trials in Switzerland. On the institutional front, in September 2025 the European Commission launched the European Connected and Autonomous Vehicle Alliance, an industrial forum to coordinate development and investment, with large-scale trials planned from 2026. The push is there. What’s missing is regulatory speed.

The Italian response

Paolo Barbato: In Italy the news of recent weeks is Niulinx. How do you read it against the picture you describe?

Giulio Salvadori: Niulinx is the sign that research can become enterprise here too. It’s a spin-off of Politecnico di Milano born from the group led by Sergio Savaresi, and in April 2026 it closed a €38 million round, the largest ever in Italy in the sector, with A2A and CDP Venture Capital as lead investors. The model isn’t the pure robotaxi but robosharing: an electric vehicle that autonomously reaches the customer, is driven during the trip and then repositions itself toward charging or the next user. The low-speed technology, up to 30 km/h, was already tested on the road in Brescia in 2025. It’s interesting because robosharing eliminates the main operating cost of traditional car sharing, the manual relocation of vehicles. It’s a model designed for European rules, not against them.

Paolo Barbato: There’s a point that recurs often in the report: autonomous driving isn’t the product of a single player. What does that mean in practice?

Giulio Salvadori: It means the ecosystem is structured on three interdependent levels. There are the OEMs, which supply the certified vehicles and the physical platform. There are the technology enablers, which develop the vehicle’s brain: the AI models, perception, simulation and validation systems. And there are the service platforms, which aggregate demand and manage fleets, turning technology into usable mobility. No one covers the entire chain alone efficiently. That’s why we’re seeing major alliances among manufacturers, sharing platforms and algorithm developers. Those who try to do everything alone, like Tesla, are the exception, not the rule.

The benefits, beyond the tip of the iceberg

Paolo Barbato: The Observatory has quantified the benefits of autonomous driving in Italy. Which numbers really count?

Giulio Salvadori: We modeled the gradual introduction of robotaxis and robosharing over a 2028-2050 horizon, distinguishing two orders of effects. The direct effect is the replacement of existing professional fleets, taxis and car sharing, with more efficient autonomous vehicles: eliminating the human factor reduces injuries per kilometer by up to 70%, and the more intensive use of vehicles makes it possible to cover the same demand with a fleet reduced to about a third. This is worth about €168 million in avoided social cost and 3,000 fewer injuries. But the real impact is the indirect effect. As services become efficient and fares fall, a growing share of households gives up the second car. Applied to an urban fleet of about 18 million vehicles, this means about 900,000 fewer cars, 100,000 injuries avoided, 41,000 fewer tonnes of emissions and an avoided social cost of €5.9 billion. That’s 97% of the €6.1 billion total estimated for the period. The message is that we shouldn’t look at the tip of the iceberg, that is the single autonomous vehicle, but at the whole block: value comes from the change in the mobility model, not from the technology itself.

Paolo Barbato: There’s also a figure on people’s willingness to use it.

Giulio Salvadori: Yes, and for anyone thinking about work-related mobility it’s the most relevant one. Today 54% of Italians say they would like to use an autonomous car, with a clear preference for repetitive trips: the home-work commute and urban journeys in heavy traffic conditions. It’s exactly the profile of daily commuting. The latent demand, on the workers’ side, already exists.

From the connected vehicle to the vehicle that delivers services

Paolo Barbato: Let’s move on to connectivity. The market is growing, but you say the frontier has shifted. Where to?

Giulio Salvadori: The connected vehicle market in Italy is now worth almost €3.7 billion, up 14%, driven by ADAS driver-assistance systems, which alone are worth €1.52 billion and grow by 27%. At the end of 2025 there were 18.8 million connected vehicles out of 46 million on the road. Connectivity, however, is no longer the frontier: by now all new vehicles are born connected. The real shift is toward the software-defined vehicle, the car as a software platform able to deliver services and manage onboard artificial intelligence. 92% of companies consider this transition inevitable, 81% expect software value equal to that of hardware. But there’s an elephant in the room: today only 13% of a manufacturer’s revenue comes from software, and in Italy we’re at 5% of turnover, €180 million. The direction is clear, monetization isn’t. Finding how to create value from data and services is the knot of the coming years.

The infrastructure that has to keep pace

Paolo Barbato: Let’s close with electrification and infrastructure. What’s the state of the art?

Giulio Salvadori: On electric there are encouraging signs. In April 2026 pure electric cars reached 8.5% of registrations, plug-in hybrids 9.1%, both clearly up on the year before, supported by incentives and company fringe-benefit rules. We remain distant from Germany, the United Kingdom and France, all over 32% of the overall share, but the direction is right. On infrastructure, a technology set to change the charging experience is Plug & Charge: since January 8, 2025 the ISO 15118-2 standard has been mandatory on new charging stations, and it cuts session start time from 30-120 seconds to 3-5, raising the first-connection success rate from 70% to 95%. The critical point remains the regulatory framework for Smart Roads, stuck at the 2018 decree. The 2026-2027 two-year period will be decisive: we need more flexible tools, like regulatory sandboxes, to close the gap with European partners and make the infrastructure ready for autonomous driving. We’re facing a decisive bend. The technology is there. What Europe has to recover is the ability to bring it onto its own roads before 2030.

A closing note

There’s a shift in perspective, in this year’s results, that’s worth making explicit. For years the Observatory’s discourse on connected vehicles and on automotive technological advances centered on the vehicle: the driver’s comfort and safety, smart onboard functions, driver-assistance systems. Starting with this edition the center of gravity has clearly shifted to autonomous driving and its impact on the entire mobility system. The two figures Salvadori cited say so. The dominant service model, in 53% of already-commercialized projects, is the robotaxi, confirming that the first scalable L4 application will be autonomous ride-hailing and not the privately owned car. And people’s demand says so too: 54% of Italians say they would like to use an autonomous vehicle precisely for repetitive trips, the home-work commute or urban journeys in heavy traffic conditions.

In his recent keynote at the Next Mobility Exhibition, professor Sergio Savaresi defined Niulinx as intermodal by design. It’s a distinction that shifts the focus from the micro to the macro and forces us to reason about new scenarios of modal integration. A truly autonomous car is no longer a good to own, but a service to use only when needed, with zero initial investment for the end user. It’s not a detail: the major European manufacturers have little interest in going beyond so-called level 2+, that is the systems that still require a driver at the wheel, precisely because a fully autonomous and shared car would reduce the consumer’s dependence on the privately owned car. This was observed by MEP Pierfrancesco Maran, today chair of the Environment Committee of the European Parliament, and the point captures exactly the industrial stakes.

Intermodal autonomous driving, by contrast, has the potential to change the car’s relationship with the rest of the system. From a competitor of public transport it can become its complement, covering the first and last mile where the network doesn’t reach. And instead of discouraging the bicycle, it can make it more central and attractive in urban mobility: if I no longer own a car and pay only for the kilometers I travel, for short systematic trips the bike makes sense, private or shared, while robosharing covers occasional or long-range trips.

When it comes to autonomous driving, it no longer makes much sense to split into supporters and opponents. In the first four months of 2026 alone, startups in the sector raised a global record of over $21 billion, $16 of which in Waymo’s round in February alone. It will arrive, in Europe too, in Italy too. The responsibility lies in ensuring it integrates into the mobility system that already exists, completing it rather than cannibalizing it, and that it is, ideally, European full-stack technology.

Giulio Salvadori

Giulio Salvadori is Director of the Connected Vehicle & Mobility Observatory at Politecnico di Milano, part of the Digital Innovation Observatories of the School of Management. The Observatory has reached its seventh edition and analyzes the evolution of the market for connected vehicles, autonomous driving and smart mobility services, in collaboration with the Department of Electronics, Information and Bioengineering, the Department of Mechanical Engineering and the IoT Lab of Politecnico di Milano.

The data cited in this article are drawn from the Executive Summary of the 2025-2026 Research “Driving the Future: mobility at the decisive bend”, available on the osservatori.net website.