inDrive and Pony.ai Plan Kazakhstan Robotaxis: What Is Confirmed

inDrive and Pony.ai are exploring robotaxi services in Kazakhstan, with Astana identified as the first city and NVIDIA providing supporting technology. This is a proposed deployment, not a service passengers can book today. The important question is how the partners turn a feasibility project into an affordable, reliable ride.

What the partners have announced

NVIDIA's October 22 announcement confirms that the companies are assessing the integration of Pony.ai's seventh-generation robotaxis into the inDrive app. It identifies NVIDIA Hyperion as the sensor and compute platform, alongside the Halos safety system.

The project separates the customer-facing ride service from the autonomous vehicle technology. That could let inDrive introduce a new type of vehicle through a familiar booking interface, rather than asking passengers to adopt a completely separate service. It does not establish that the integration is finished.

Dow Jones reports, citing the companies' joint statement, that they signed a memorandum of understanding to assess technical, regulatory and commercial feasibility. Pony.ai would supply the vehicles and inDrive would provide the platform and operate the service. An initial Astana launch remains subject to an agreement with Kazakhstan's government.

When could inDrive robotaxi testing begin?

inDrive CTO Yuri Misnik told TechCrunch that testing with safety operators is planned for the first half of 2027.

A testing target is not a public driverless launch date. The distinction changes what a reader should expect: trials can evaluate a proposed service without opening unrestricted bookings to paying passengers.

Dow Jones says the announcement did not disclose fleet size, investment or a launch date. There is consequently no confirmed passenger price or booking schedule to compare with conventional rides. The current evidence supports a partnership and a testing ambition, not a completed rollout.

How this fits inDrive's autonomous vehicle strategy

The company's AV initiative describes a partnership-led approach focused on emerging markets. inDrive says it intends to contribute its platform, operational data and market access instead of making large investments in vehicle manufacturing or fleet ownership.

It also wants existing drivers to have a route into owning and operating autonomous fleets. That is a stated strategic aim, not an ownership offer announced for Kazakhstan. The Astana proposal does not establish who could buy a vehicle, what financing would involve or how revenue would be divided.

For a potential owner-operator, those details would matter as much as driving performance. Removing a person from the driving seat does not remove the need to clean, maintain, charge and recover a vehicle. Someone must still account for those costs and the time a car spends unavailable.

What would make the Astana project meaningful?

A useful pilot would establish three things together: the conditions in which the vehicles can operate, how much human assistance the service needs, and the cost of delivering an actual passenger trip. Measuring only the driving system would leave the commercial question unanswered.

For example, a vehicle that completes a route successfully may still be expensive to operate if it spends long periods waiting for support or travelling without a passenger. Conversely, a tightly defined service area could be useful if it covers repeatable demand at a sustainable cost. These are evaluation criteria, not results the partners have published.

The next meaningful milestone is therefore a defined operating plan: where rides can happen, under what supervision and with what access for passengers. That would turn the announcement from an intention to enter a market into something users and fleet operators can assess.

I'm Paras. I work at the intersection of AI, engineering and distribution.

I help technical founders explain complex AI products and launches in clear stories that developers, customers and investors can understand and share.

Book a 30-minute call.