What’s Really Holding Autonomous Mobility Back?

Welcome to Smart Mobility Spotlight, Imagry’s newsletter delivering the latest in autonomous driving and smart mobility solutions. Subscribe to stay informed with curated insights, expert perspectives, and emerging trends in the ever-evolving world of transportation.

Beyond the Autonomous Bus and Shuttle Pilot

Over the past decade, autonomous shuttles and buses have been piloted in hundreds of cities around the world. Most of these pilots were considered successful. The vehicles ran their routes, passengers rode them, safety records held, and final reports were written. And then, in the vast majority of cases, the service ended, the vehicles were shipped to a different destination, and the area resumed its original function.

The common explanation has been technological maturity: once the systems are good enough, deployment will follow naturally. Yet the closer one looks at the past decade, the less complete that explanation appears. The technology has continued to prove itself in pilot after pilot, while proven pilots have rarely turned into permanent service, which suggests the gap lies somewhere beyond the technology itself. So, a question worth examining more closely is: if the technology works, why do successful pilots so rarely become lasting service?

🔁 The Pilot Trap: Why Success Doesn’t Convert

There is a name for what happens when pilot projects stop being a step toward deployment and become a destination in themselves: the pilot trap. A pilot is announced, runs well, concludes, and is followed by another pilot somewhere else, with no pathway from any of them to lasting service. The technology keeps passing the test, and the test keeps being repeated.

Research confirms how deep this pattern runs. A peer-reviewed study published in Nature’s Humanities and Social Sciences Communications assessed how 58 U.S. cities were using autonomous vehicle pilot projects and interviewed planners across 20 of them. Most pilots were funded by short-term grants, which means the project ends when the money does, regardless of how well it performed. The most common goal cities set for their pilots was not solving a transport problem at all, but introducing the technology to the public. The researchers also found a consistent disconnect between what the pilots measured and the transportation goals cities actually held, and little evidence of learning carried from one pilot to the next. Their conclusion: pilots, as currently structured, exist in a state of “permanent temporariness”.

In other words, the typical pilot is designed as an endpoint. Success is defined as completing the demonstration, not continuing the service, so there is nothing on the other side of a successful pilot except another pilot. And the technology is only one part of the equation. As we explored in a previous Imagry newsletter What’s Really Holding Mobility Innovation Back?, most of the barriers standing between autonomous mobility and the cities that need it are not technology questions at all. A pilot that proves the technology has therefore addressed only a fraction of the problem.

🌍 When Every City is its Own Market

If the trap is so well documented, why does the industry keep falling into it? A large part of the answer is regulatory fragmentation. A pilot proves the technology in one place, under one set of rules. Deployment means running that same technology across many places. But when every jurisdiction has its own regulations, its own approval procedures, and its own requirements, nothing carries over. Each new city means new permits, new adaptations, new integration work, as if the previous pilots never existed. The technology may be ready to scale, but there is no single market to scale into, only fragments, each too small to justify the investment.

Europe is the clearest illustration. The region has run more autonomous shuttle and bus pilots than anywhere else, yet at a European Parliament event on autonomous vehicles in December 2025, the contrast was stated plainly: operators in the U.S. and China have already logged more than 150 million autonomous kilometers, while Europe remains stuck in small-scale pilots, held back in part by the stalemate of 27 different national traffic rules.

And fragmentation has a price tag. When no unified market exists, suppliers cannot build a sustainable business on scattered pilots, and some eventually give up on the segment entirely. When they do, the cities running their vehicles are left stranded. That is exactly what happened in the German city of Monheim, which had to suspend a functioning autonomous bus route, one its operator described as an essential part of the network, after the vehicle supplier withdrew from passenger transport to focus on markets where customers were ready to move beyond pilots. The route worked successfully and riders used it. What was missing was a market on the other side of the pilot.

📈 Deployment is the Classroom

There is a second reason the industry, which as adopted AI-based technology, cannot stay in pilot mode. It has to do with how autonomous systems learn. An autonomous vehicle improves by encountering situations on the road. The common ones, intersections, pedestrians, merging traffic, appear within the first weeks of any pilot. But the situations that matter most for safety are rare (these are known as “edge cases”) and may occur only once in a million kilometers. A typical pilot covers only a few thousand kilometers, which is not nearly enough to enable the vehicle to encounter a sufficient number of edge cases to satisfy requirements for safety and reliability.

The strongest evidence for this comes from the deployments that did scale. A 2025 peer-reviewed study in Traffic Injury Prevention analyzed 56.7 million driverless miles of Waymo‘s ride service against human-driver benchmarks on the same road types, finding statistically significant reductions across crash categories, including a 96 percent reduction in injury-involving intersection crashes. But the finding that matters most for this discussion is methodological: the authors note this was the first retrospective assessment of a driverless system able to draw statistical conclusions about serious crash outcomes at all, because only now had any system accumulated enough real-world mileage for the analysis to be possible. The safety evidence demanded from the entire industry cannot be produced inside a pilot. It is a product of scale, and today it exists only where deployment has actually happened.

The same is true of everything around the vehicle. Remote supervision, maintenance regimes, incident handling, passenger service, insurance models: these are operational capabilities, and operational capabilities are built through operation. This is why two different approaches have emerged across the industry. One treats deployment as the final step: pilot, prove maturity, then scale. The other treats deployment as the source of maturity: scale under careful supervision, and let the software, the operations, and the safety evidence improve through real daily service. The latter approach attaches a hidden price to the former: every year a market spends waiting for the technology to finish maturing is a year in which operators deploying elsewhere collect the experience that actually does the maturing.

🚌 What Operators Can Do Now

For public transport operators and authorities, this reframing carries a practical message: preparing for autonomy is not a question of waiting longer. It is a question of planning earlier.

The sector’s own guidance points the same way. A 2026 publication from UITP examines the benefits, barriers, and catalysts of deploying automated vehicles at scale, and highlights that operators need to prepare the systems around the vehicle, from legacy IT and operational technology to workforce and processes, well before large fleets arrive. Operational readiness is built in years, not procurement cycles.

The most concrete starting point is one operators already control: fleet renewal. Every bus in service today has a retirement date, and every operator already plans long-term replacement budgets around those dates. That turns the deployment question from “where do we find new money?” into a far more manageable one: “what do we buy next, when vehicles must be replaced anyway?” Treating autonomy as part of the normal renewal cycle, and building the long-term planning, procurement signals, and operational readiness around that decision, is how deployment becomes an evolution of existing budgets rather than a leap into new ones. As we argued in a previous Imagry newsletter entitled Ten Commandments, One Question: What Kind of Transport System Are We Building?, the technology should be chosen to serve a defined goal. For operators facing aging fleets and deepening driver shortages, that goal is already defined.

🚍 Imagry’s Approach: Ready for the Other Side of the Pilot

The pilot trap is structural, and no technology company can break it alone. Markets, regulation, and procurement have to open the pathway. What technology can do is make sure that when the pathway opens, nothing has to be rebuilt from city to city. That is the principle guiding the design of Imagry Autonomous Buses solutions:

• AI-based, HD-mapless autonomy. Our system runs on real-time, vision-centric perception, with no dependency on HD maps, LiDAR, cloud connectivity, or road infrastructure changes. As a result, deployment is faster and far less costly to scale from one city to the next.
• Flexible, location-independent technology. The same software stack operates across buses and shuttles, in left- and right-hand-drive markets, with vehicles operating on public roads in Europe, Israel, the U.S., and Japan.
• Built for public transport use cases. The platform is optimized for predictable, high-frequency service: complementing human-driven fleets at peak hours, shuttling people across medical, university, and industrial campuses, and closing first and last mile gaps to transit hubs to serve the underserved.

The pilot question is real, and the industry has answered it many times over. But it was always the first question, not the last. The next one belongs to planning, procurement, and readiness, and the operators, cities, and systems that thrive will be the ones answering it now.

Imagry in the News

Click here to see the latest news and events featuring Imagry’s autonomous driving solutions.

Video Spotlight

Recorded live at MOVE 2026, Imagry CEO Eran Ofir joins the first Move Monday episode to discuss one of public transportation’s biggest challenges: the global shortage of bus drivers. The conversation explores how autonomous buses, mapless AI, and scalable deployment could help shape the future of urban mobility.

Autonomous Mobility Career Opportunities

We’re building more than autonomy. We’re building a team that dares to do what others say is impossible.

We value people who chase hard problems not credit. Who ask better questions. Who stay curious. Who care about the mission, not job titles. And we know that to build the future, we need all kinds of minds.

If that sounds like you, we’d love to meet you.

See our open positions here.

Want to receive information about automated mobility on a regular basis?