Uber Gets Green Light to Operate Robotaxis in Madrid: Level 4 Lands in Spain

The Directorate-General for Traffic (DGT) has taken a historic step by authorizing Uber to operate level 4 autonomous cars in Spain. This means that vehicles can drive without a human driver in limited environments, albeit with remote supervision and specific conditions. The company has announced the initial deployment of 20 robotaxis in Madrid before the end of 2026, adapting to demand and different areas of the capital.

With this authorization, Madrid positions itself as the third European city where Uber will operate autonomous vehicles, after London and Zagreb. A strategic move that places the Spanish capital at the forefront of autonomous transport on the Old Continent.

Technology Partners and Roadmap

To carry out this ambitious project, Uber has established key alliances. On the one hand, the Chinese company WeRide, with previous experience in bus tests in Spain, although this will be its first foray into passenger cars. On the other, Avomo, the mobility division of MooveCars. The initial phase includes detailed mapping of the city, route validation, and preparation tests under the DGT's ES-AV regulatory framework, specific to automated vehicles. During this stage, vehicles will always have a specialist on board for supervision.

Uber hopes that this project "establishes an important foundation for future deployments of autonomous transport services throughout the country and in the European market in general," according to a company statement.

Expert Vision: Mature Technology, But Not for Europe

COMPUTERWORLD has consulted Manuel García, director of the Master's in 'Big Data' and researcher in AI and Autonomous Driving at the European University, who provides a fundamental technical perspective. According to García, from a technological standpoint, robotaxis already operate commercially in several US states, such as California and Texas. "Therefore, at a technological level, there is already quite mature technology," he acknowledges. However, he qualifies: "Nevertheless, this technology is proven in a traffic environment and road situation like the American one, which is very different from the European one."

The expert highlights critical differences: in the US, traffic signs, road markings, and traffic lights differ from European ones, and there are no roundabouts. "It is conceivable that before seeing fully autonomous robotaxis, there will be two initial phases: one of data collection to adapt existing technology to the European market; and another of testing in which this new adapted technology is verified to determine if there is any human failure, so that it can take control and correct the situation."

García emphasizes that the presence of a human on board is necessary during the initial phase of the testing plan. "They must take control of the vehicle if the autonomous car does not behave correctly. Once the testing plan concludes satisfactorily without human intervention, it could disappear," he adds.

"Level 4 autonomous driving is already proven and mature. Level 5 is very far off in time."

— Manuel García, director of the Master's in 'Big Data' and researcher in AI and Autonomous Driving at the European University.

Impact on Urban Mobility

Regarding the expected effects on urban mobility, congestion, and public transport when robotaxis begin operating, this expert does not believe there will be widespread relevant disruptions. "It could be that, in the testing phases, a robotaxi encounters an isolated situation it does not know how to handle, which will lead the vehicle to enter safe mode and become blocked awaiting human intervention." However, as he adds, "it is expected to be very occasional and limited."

Level 5: A Distant and Complex Horizon

Regarding level 5, which would involve travel anywhere and under any condition without a steering wheel or pedals, Manuel García considers that "that is very far off in time." In his opinion, technological research is dominated by American and Chinese institutions and companies. "Therefore, much of the developed technology needs adaptation to the European market, since there are road elements, such as roundabouts, that do not exist or are different."

Consequently, the expert maintains that level 5 will take time to be seen on our streets "given that general driving on open roads entails many dangers and diverse and changing casuistries that must be incorporated into the decisions the vehicle makes." To explain this, he refers to Moravec's Paradox, a key concept in AI: tasks that are easy for a person are very difficult for a machine and vice versa. "In the end, driving for a person is a simple task (since it is designed by and for humans) and for a machine it is very difficult," he concludes.

Implications for the IT and Business Sector

The arrival of robotaxis in Madrid will not only transform mobility but also have a significant impact on the technology and business sector. Managing autonomous fleets will require robust data infrastructures, with real-time processing capabilities and ultra-reliable communications. Companies will need to adapt their IT strategies to integrate these vehicles into their supply chains and mobility services.

Furthermore, cybersecurity becomes a fundamental pillar: protecting autonomous driving systems from malicious attacks is critical. In this sense, initiatives such as security guides for Microsoft 365 can serve as a reference for establishing data protection and access policies in cloud environments, something that will be essential for platforms managing robotaxis. Also, server virtualization with Proxmox can facilitate the creation of scalable and isolated test environments to validate autonomous driving algorithms.

The integration of AI agents, such as those described by AWS Agents, which suggest decisions but leave final execution to code, is an example of how hybrid systems are being designed in which AI assists but does not completely replace deterministic logic. This approach could be applied to autonomous driving, where AI makes decisions but always under verifiable safety frameworks.

On the other hand, managing permissions and contexts in enterprise RAG systems, as explored in this article, anticipates the security challenges that will arise when handling large volumes of vehicle and user data in autonomous driving environments.

Finally, the consolidation of technology providers, such as the case of V-Valley and HPE, reflects the trend towards integrated ecosystems that can support the complexity of large-scale autonomous operations.

Original source: ComputerWorld. Analysis and adaptation by ForgeNEX.