Uber rival inDrive eyes 2027 robotaxi tests in Central Asia with Pony.ai and Nvidia

inDrive, Pony.ai, and Nvidia plan to test robotaxis in Kazakhstan beginning in the first half of 2027. The trial will start in Astana, the capital, and initially include a safety operator. The eventual goal is fully driverless rides. This matters because it would extend autonomous ride-hailing beyond the United States and China.

Pony.ai would provide its Gen-7 robotaxis. inDrive would list the rides in its existing ride-hailing app and operate them in Kazakhstan. Nvidia would supply the autonomous-driving technology, including its Hyperion platform and Halos safety system. The companies have signed a memorandum of understanding to explore the partnership.

The plan fits Kazakhstan’s own move toward autonomous taxis. The country announced plans for autonomous taxi services in 2027 and introduced rules for testing on certain public roads. Astana’s snow and ice could provide a demanding environment before any wider rollout. The companies have not announced a fully driverless launch date.

A robotaxi is an autonomous vehicle offered as a taxi or ride-hailing service. It uses sensors, computers, and artificial-intelligence systems to handle driving tasks. An ordinary ride-hailing car depends on a human driver to steer, observe traffic, and make decisions. A robotaxi is designed to perform those tasks itself.

The Kazakhstan proposal shows the difference in practice. inDrive would offer Pony.ai’s Gen-7 robotaxis through its app. The first trials in Astana would still include a safety operator. That person would provide an additional safeguard while the companies evaluate the system. The longer-term objective is fully driverless rides.

Robotaxis could change how ride-hailing services operate, but they must prove they can work safely and affordably. Astana’s harsh winters create a difficult test because snow and ice can challenge autonomous-driving systems. The article does not provide a final definition of commercial approval or say when driverless service would begin.

The proposed partnership gives each company a different job. Pony.ai would provide the physical vehicles, specifically its Gen-7 robotaxis. inDrive would bring the service to passengers through its ride-hailing platform and operate it in Kazakhstan. Nvidia would supply the technology needed for autonomous driving.

Nvidia’s contribution would include its Hyperion platform and Halos safety system. Hyperion is described as a ready-made foundation combining sensors and computing technology. That could reduce the need for partners to build and integrate every technical component themselves. Pony.ai’s vehicles would then be offered through the inDrive app as rides.

The model reflects inDrive’s “capital-light” approach. The company says it combines its expertise with partners in different markets instead of doing everything itself. The agreement is currently a memorandum of understanding, so it describes a plan to explore deployment rather than confirming a completed commercial service. Initial rides would include a safety operator.

The proposed Kazakhstan test could become part of a much larger international expansion. Pony.ai aims to operate more than 3,500 robotaxis in over 20 cities worldwide by the end of 2026. That target is far beyond a single Astana pilot. It shows the company is building a multi-city robotaxi network.

inDrive supplies a broad distribution channel. It operates in more than 1,200 cities across 47 countries. If the Kazakhstan model works, the company could examine similar partnerships in other markets. Its separate agreement with Saudi Arabia’s AI Driver already explores autonomous vehicles powered by WeRide, beginning with Riyadh.

These figures describe potential reach, not a confirmed global deployment. The Kazakhstan plan remains an exploration under a memorandum of understanding. Successful expansion would also depend on local rules, road testing, affordability, and safety. inDrive says a favorable and collaborative regulatory environment is one major requirement for launching autonomous services.

Astana could be an important test site because its winters are harsh. The city faces months of snow and ice, creating difficult road conditions for autonomous-driving systems. Testing there would examine how well robotaxis operate when visibility, road surfaces, and driving conditions become more challenging.

The proposed trials would begin with a safety operator, giving the companies a way to evaluate the vehicles while adding human oversight. Pony.ai would provide the Gen-7 robotaxis, while Nvidia would supply systems including Hyperion and Halos. The results could show where the technology performs reliably and where it needs improvement.

The article presents winter testing as a possible source of broader lessons. inDrive founder Arsen Tomsky said success in Astana could help expansion to locations facing similar challenges. However, the article does not report test results yet. Testing is planned for the first half of 2027, so the winter performance remains to be evaluated.

Robotaxis could affect three important parts of ride-hailing: cost, safety, and availability. Nvidia says its Hyperion platform could lower development costs by providing sensors and computing technology in a ready-made foundation. Lower technology costs may matter especially where ride-hailing is already affordable and autonomous services must compete on price.

A concrete example is the planned Astana service. Pony.ai would provide the vehicles, inDrive would connect them with passengers, and Nvidia would supply the autonomous systems. Early rides would include a safety operator. That approach could support careful testing before the eventual goal of fully driverless trips, while also showing whether the service works in demanding conditions.

Success could help inDrive bring autonomous rides to more markets through partnerships. It might make service available in places that attract fewer human drivers, but the article does not promise that outcome. Expansion requires safe performance and supportive rules. inDrive specifically identifies a favorable, collaborative regulatory environment as a key requirement.

Autonomous vehicles combine sensors, computing systems, and AI models to drive without continuous human control. Sensors gather information about the vehicle’s surroundings. Computing systems process that information quickly. AI models and driving software then help interpret the road, identify hazards, predict what may happen, and choose actions such as steering, slowing, or stopping.

The article highlights Nvidia’s role in this chain. Its Hyperion platform provides a ready-made foundation of sensors and computing technology. Nvidia also offers the Halos safety system and its Alpamayo model. Tools for synthetic data and testing can expose autonomous-driving systems to different scenarios before or during road trials.

Pony.ai would use this technology in its Gen-7 robotaxis, while inDrive would provide the passenger service. Kazakhstan’s testing would begin with a safety operator and take place in Astana’s snowy, icy conditions. The article does not detail each sensor or decision-making algorithm, but it identifies the main technology layers and testing tools.