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

Ride-hailing company inDrive, together with Chinese autonomous driving firm Pony.ai and chipmaker Nvidia, announced on Friday that it has signed a memorandum of understanding to explore bringing robotaxi service to Kazakhstan. According to inDrive CTO Yuri Misnik, testing will begin in the country's capital, Astana, in the first half of 2027. The first phase will put a human safety operator in the vehicle, with fully driverless rides described as the eventual goal.
The division of labor under the proposed plan is straightforward on paper. Pony.ai would supply its Gen-7 robotaxis, inDrive would surface them inside its existing ride-hailing app and run local operations, and Nvidia would contribute AI technology, specifically its Hyperion platform and its Halos safety system. A memorandum of understanding is a non-binding statement of intent rather than a contract, so the 2027 timeline and the commercial structure remain plans rather than commitments.
This is a notably different model from the one most Western robotaxi efforts have followed. Nvidia describes Hyperion as a ready-made foundation of sensors and computing hardware that lets companies avoid building and integrating every component themselves, which the company says can lower development costs. It also offers the Alpamayo model along with tools for generating synthetic data and testing driving systems against varied scenarios. Ali Kani, Nvidia's vice president of automotive, framed the cost angle as especially salient in emerging markets, where ride-hailing is already inexpensive and an autonomous service has to compete on price rather than novelty. That is a sensible reading of the economics: a driverless car carries heavy upfront hardware costs, so the only path to a viable fare in a low-price market is amortizing that hardware across high utilization and driving down per-vehicle integration work. Whether Nvidia's off-the-shelf stack gets costs low enough is the open question.
inDrive's own framing leans on the same theme. Misnik described the company's approach as capital-light by design, built on partnering with specialists in each market rather than owning the full autonomous stack. This is consistent with inDrive's history: the company built its brand on negotiated fares and affordability, and it operates in more than 1,200 cities across 47 countries, having expanded into delivery, intercity transport and financial services. In practice, a capital-light posture means inDrive is betting on being the demand and operations layer while partners carry the vehicle and autonomy risk.
Astana looks like a deliberate stress test rather than an easy first market. The city has a population of roughly 1.7 million and endures months of snow and ice each year, conditions that are hard on perception systems, lane markings and sensor cleanliness. inDrive founder and CEO Arsen Tomsky said launching there would provide an opportunity to evaluate the technology in challenging winter conditions, and that success could yield insights for other locations with similar challenges. The logic is that a system proven on ice and in low-visibility weather travels well to other northern cities; the risk is that winter edge cases slow the program down.
Kazakhstan's regulatory posture supports the plan. Last month the country announced intentions to launch autonomous taxi services in 2027, with initial trials slated for Astana and two other cities, and it has introduced rules allowing autonomous vehicles to be tested on certain public roads. Misnik cited a favorable and collaborative regulatory environment as one of the main requirements for a successful launch, and said inDrive is examining additional markets for future deployments.
Kazakhstan is not inDrive's first autonomous foray. In December the company signed a separate agreement with Saudi Arabia's AI Driver to explore deploying vehicles powered by Chinese self-driving firm WeRide, starting in Riyadh, with a pilot targeted for the first half of 2026. The new deal adds a second Chinese autonomy partner and a second geography. Pony.ai, for its part, is pushing hard outside China: it aims to have more than 3,500 robotaxis running in over 20 cities worldwide by the end of 2026, up from a fleet of 1,975 vehicles at the end of June. Uber has also said it will test Pony.ai robotaxis in London and has begun supervised autonomous rides there with British startup Wayve, and it is working with multiple robotaxi companies in other markets including the Middle East. The competitive picture, then, is a scramble for partners and cities rather than a single winner-take-all race.
Friday's announcement coincided with Nvidia CEO Jensen Huang's first visit to Kazakhstan, which hints at how much of this kind of dealmaking now runs through Nvidia's automotive relationships. A likely trade-off for inDrive is control: relying on Pony.ai for vehicles and Nvidia for the compute and safety stack leaves inDrive owning the customer relationship and local operations but not the core technology, which limits both differentiation and the ability to fix problems without a partner.
Why it matters: this is an early signal that robotaxi deployment is spreading beyond the US and China into Central Asia, with a local ride-hailing platform acting as the distribution layer rather than a tech giant. For developers and product teams, the interesting piece is the stack: Nvidia's Hyperion plus Halos, paired with Pony.ai's Gen-7 vehicles, is being pitched as a way to skip years of integration work, which could make autonomous fleets accessible to smaller operators. Whether that holds up in Astana's winters by 2027 is inference at this stage, but a second inDrive partnership after the Riyadh deal suggests the company intends to run this playbook across multiple markets.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
The company plans to begin testing Pony.ai's robotaxis in Kazakhstan's capital in the first half of 2027.