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

October 23, 2026 · DigitalEdu

inDrive, the ride-hailing platform known for its fare-negotiation model, has announced plans to begin testing Pony.ai's autonomous vehicles in Astana, Kazakhstan, during the first half of 2027. The initiative marks inDrive's most concrete step yet toward deploying robotaxi services and signals a strategic bet on Central Asia as a proving ground for self-driving technology. With Nvidia providing the underlying compute infrastructure, the partnership brings together a regional mobility leader, a Chinese autonomous driving specialist, and the dominant supplier of AI hardware.

What Happened

inDrive confirmed it will launch a pilot program using Pony.ai's Level 4 autonomous vehicles in Astana, the capital of Kazakhstan, targeting the first half of 2027 for initial testing. The vehicles will operate on Pony.ai's virtual driver system, which handles perception, planning, and control without human intervention within designated operational domains. Nvidia's DRIVE Thor platform, its next-generation centralized compute architecture for autonomous vehicles, will power the onboard AI workloads. The companies have not disclosed the size of the initial fleet, the specific geographic boundaries of the test zone, or whether the service will be open to the public or limited to invited users. inDrive, which operates in over 700 cities across 46 countries, has historically differentiated itself through a peer-to-peer fare negotiation model rather than algorithmic pricing. This pilot represents its first public commitment to integrating fully driverless vehicles into that network.

Why It Matters for Businesses

The partnership illustrates how ride-hailing platforms outside the U.S. and China are pursuing autonomous strategies tailored to their operational realities. inDrive's strongest markets — across Central Asia, Latin America, Africa, and South Asia — often feature complex traffic environments, less standardized road infrastructure, and regulatory frameworks still evolving for autonomous vehicles. By selecting Astana for its first robotaxi tests, inDrive is signaling confidence that Kazakhstan's regulatory environment can support iterative deployment, and that the city's climate and urban layout offer a meaningful stress test for Pony.ai's technology. For businesses in logistics, last-mile delivery, and fleet management, the pilot offers a rare window into how Level 4 autonomy performs outside the controlled conditions of Phoenix, San Francisco, or Beijing. Nvidia's involvement also underscores a hardware trend: centralized, high-compute platforms like DRIVE Thor are becoming the default choice for companies skipping Level 2+ driver-assist systems and targeting full autonomy directly. Companies evaluating autonomous partnerships should note that inDrive's model — integrating a third-party virtual driver into its own dispatch and customer-facing stack — may become a more common go-to-market path than vertical integration.

The Technology Stack and Regional Context

Pony.ai brings a virtual driver system already tested in Guangzhou, Shenzhen, Beijing, and Irvine, California, with over 3.5 million kilometers of autonomous driving logged across its fleets. The company received one of China's first permits for driverless robotaxi services on public roads and has operated a commercial service in Guangzhou's Nansha district since 2022. Its expansion into Central Asia via inDrive represents its first major deployment outside China and the United States. Nvidia's DRIVE Thor, announced in 2022 and sampling to automakers in 2024, delivers up to 2,000 teraflops of AI performance on a single system-on-chip, consolidating autonomous driving, parking, and in-cabin AI functions. For inDrive, the choice of Thor suggests a preference for a scalable, software-defined architecture that can be updated over-the-air as Pony.ai's models improve. Kazakhstan has been actively courting technology investment through its Astana Hub technopark and a national AI strategy adopted in 2024. The government has signaled openness to autonomous vehicle testing, though a comprehensive legal framework for commercial driverless operations remains under development. The pilot will likely proceed under a special regulatory sandbox arrangement, a model used in other jurisdictions to enable limited deployment while permanent rules are drafted.

What To Watch

Several factors will determine whether this pilot scales into a commercial service. First, the regulatory trajectory in Kazakhstan: if the government establishes clear liability, insurance, and data localization rules for autonomous vehicles by 2027, inDrive and Pony.ai will have a template for expansion to other Central Asian capitals where inDrive holds dominant market share, including Tashkent, Bishkek, and Dushanbe. Second, technical performance in Astana's winter conditions — temperatures regularly drop below -20°C — will test sensor reliability, battery management, and perception algorithms trained primarily in subtropical and temperate climates. Third, the commercial terms between inDrive and Pony.ai will set a precedent for revenue-sharing models between platform operators and autonomy providers. Unlike Uber's now-divested ATG or Lyft's partnership with Motional, inDrive is not building its own stack; it is buying autonomy as a service. If the unit economics work in a market with lower average fares than the U.S. or China, the model could accelerate robotaxi adoption in emerging markets globally. Finally, Nvidia's ability to deliver Thor at volume and support the software integration timeline will be a dependency not just for this pilot but for the broader autonomy ecosystem counting on Thor as the post-Orin standard.

The Bottom Line

inDrive's 2027 robotaxi pilot in Astana is more than a regional experiment — it is a test case for deploying Level 4 autonomy in the complex, cost-sensitive environments where most of the world's ride-hailing trips actually occur. Business owners in mobility, logistics, and AI infrastructure should track this pilot not for its immediate scale but for what it reveals about regulatory readiness, hardware-software integration, and the viability of platform-provider partnerships outside the traditional autonomy hubs. If inDrive and Pony.ai can demonstrate safe, reliable, and economically sustainable driverless operations in Central Asia, the playbook they write will be studied by every fleet operator eyeing the next decade of automation.

Source: Original Article

Related: Blackstone’s Jas Khaira says fears over AI-buildout debt are overblown, TechCrunch Founder Summit 2026: Everything you need to know , AI promises productivity, but for managers, it’s mostly a headache for now.

For more on this, see Anthropic CEO outlines plan to ‘pace the frontier’.

Key AI Automation Facts

• AI chatbots resolve 60–80% of routine customer-support questions without human handoff.
• Workflow automation saves teams roughly 15–20 hours per employee per week (DigitalEdu client averages).
• Small businesses adopting AI automation report ~2.5x faster growth than comparable non-adopters (industry analysis).

AI-generated facts — rounded industry estimates; verify independently before republication. See more in our AI Automation Statistics hub or book a free strategy call.