The proposed Astana project would combine inDrive’s ride-hailing platform, Pony.ai’s autonomous vehicles and NVIDIA’s computing technology as the companies assess opportunities in Central Asia.
Kazakhstan | Artificial Intelligence, Technology & Business
Information checked on 26 October 2026.
The proposed inDrive robotaxi project is moving towards testing in Kazakhstan, following a memorandum of understanding between inDrive, Pony.ai and NVIDIA. Announced on 23 October 2026, the collaboration identifies Astana as the first planned location.
TechCrunch reported a first-half 2027 start, initially with a safety operator.
The agreement establishes a framework for evaluating autonomous ride-hailing. It does not confirm that a fully driverless commercial service is available in Kazakhstan. The companies say launch would follow an agreement with the country’s government.
For Central Asia, the project presents an opportunity to examine how autonomous vehicles could operate within an existing ride-hailing service—and what would be required to make that service practical.
What the inDrive Robotaxi Agreement Covers
The companies will assess the technical, operational, regulatory and commercial feasibility of offering autonomous rides through the inDrive app.
Each partner would contribute a different part of the service:
| Company | Proposed contribution |
|---|---|
| inDrive | Provide the ride-hailing platform and operate the service |
| Pony.ai | Supply its seventh-generation robotaxis and autonomous-driving technology |
| NVIDIA | Support the project with AI infrastructure and autonomous-vehicle technology |
This division brings together passenger access, vehicle automation and computing infrastructure.
The business challenge is making those components work as one service. A successful journey requires more than an autonomous vehicle completing a route: customers also need clear booking information, understandable pickup instructions and support when something goes wrong.
Pony.ai’s Gen-7 Robotaxis Provide the Vehicle Foundation
The planned project would use Pony.ai’s Gen-7 robotaxis.
Pony.ai’s relationship with NVIDIA predates the Kazakhstan announcement. In an April 2026 update, the company said their collaboration began in 2017 and described the computing hardware supporting its autonomous vehicles.
According to that announcement, Pony.ai began mass production of a robotaxi computing controller using four NVIDIA DRIVE AGX Orin chips in 2025. The company said that controller powers its Gen-7 fleet.
Pony.ai also announced a newer controller based on NVIDIA DRIVE Hyperion and DRIVE AGX Thor. However, that separate product announcement should not be treated as confirmation of the precise hardware configuration for Astana.
For the Kazakhstan project, local testing will need to establish how the selected vehicles and systems perform on the intended routes and under the conditions permitted for operation.
How NVIDIA’s Technology Supports Autonomous Driving
NVIDIA DRIVE Hyperion is a reference architecture that brings together computing hardware, software and a validated sensor suite for autonomous-vehicle development.
Sensors provide information about the vehicle’s surroundings, while onboard computing supports the software that interprets those inputs and determines driving actions.
A reference architecture gives developers a common technical foundation. It does not, by itself, establish that a particular service is ready for public operation in every location.
NVIDIA also describes Halos as a system connecting safety engineering, evaluation, validation and inspection across autonomous-vehicle development.
For the proposed pilot, the practical question is how the partners use these tools to produce evidence of reliable operation. Technology specifications and local operating performance are different measures, and both matter when assessing readiness.
What Level 4 Autonomy Means
The announcement refers to Level 4-ready autonomous vehicles.
Under the automation framework explained by the US National Highway Traffic Safety Administration, a Level 4 system performs the driving task within specified operating conditions. Within those limits, it does not depend on a human occupant to take over driving.
Those boundaries are essential. They can define where and under what circumstances an automated service operates.
Describing a vehicle as Level 4-ready therefore does not establish unrestricted availability across a city. The service still needs a defined operating area, appropriate validation and the relevant arrangements for deployment.
For passengers, clear communication about these limits would be as important as the booking process itself.
Why Astana Is an Important Testing Location
Astana would be the project’s first city, with Kazakhstan serving as the initial test market.
In its announcement, inDrive highlighted the city’s long winters, snow and ice as valuable conditions for evaluating autonomous mobility.
That makes weather a central part of the project’s practical assessment. The announcement identifies a challenging environment; it does not establish that the proposed fleet has already demonstrated reliable operation throughout an Astana winter.
A useful evaluation would examine service availability, operating restrictions and the circumstances in which vehicles cannot complete requested journeys.
For customers, dependability would mean knowing when the service is available and what assistance they can expect if conditions change.
The Business Opportunity Extends Beyond Vehicle Automation
From a business perspective, the partnership could connect autonomous-driving technology with an established way for passengers to request transport.
However, providing access through an app is only one part of the commercial task.
A robotaxi operation also needs workable arrangements for vehicle maintenance, cleaning, customer assistance and fleet availability. Its economics depend on the relationship between those costs and the number of useful journeys vehicles can complete.
For example, a vehicle that operates reliably for only a limited portion of the day presents a different business case from one that can serve passengers consistently across its permitted operating hours.
These are considerations for assessing the proposed service, rather than performance results disclosed by the partners.
What Passengers Still Need to Know
The joint announcement does not specify the initial fleet size, passenger fares, detailed service boundaries or a confirmed public booking date.
Those details will determine how the proposal translates into an everyday transport option.
Passengers would benefit from clear information about where rides are available, how to identify their vehicle, how to contact support and what happens when a journey cannot proceed.
Accessibility will also be an important design consideration. Booking, boarding and receiving assistance can present different needs for different passengers, so convenience should be evaluated across the complete journey.
What Comes After the Astana Pilot?
Following the initial launch, the companies intend to assess whether the approach could extend to other Kazakh cities and additional countries where inDrive operates.
That makes Astana a proposed starting point for broader evaluation. It is not a confirmed timetable for deployment across Central Asia.
The most useful signs of progress would be an agreed operating framework, clearly defined trial conditions and evidence showing how the service performs locally.
For inDrive and its partners, the opportunity is to turn a technology collaboration into dependable transport. The next stage will be measured through the experience of booking, taking and completing a journey.
Explore more global business, technology and innovation coverage in The Empire Magazine’s article archive.
Read our related coverage: https://theempiremagazine.com/us-halloween-candy-sales/
Follow The Empire Magazine:
Facebook | Instagram
The Empire Magazine | Crown For Global Insights
