Nvidia H200 China and Kazakhstan: New Robotaxi Partnership Announced

The ride-hailing platform inDrive, the Chinese autonomous driving firm Pony.ai, and the technology giant Nvidia have signed a memorandum of understanding to test robotaxis in Kazakhstan starting in 2027. This partnership aims to introduce driverless vehicles to the streets of Astana, the Kazakh capital, to evaluate the commercial and technical viability of autonomous mobility in Central Asia. The initiative marks a significant step in testing high-level automation in challenging environments, specifically targeting the extreme winter weather conditions found in the region.
What is the plan for robotaxi testing in Kazakhstan?
The three companies intend to pilot autonomous vehicles in the first half of 2027 within the city of Astana. According to KuCoin, the initial testing phase will involve safety operators inside the vehicles, with the ultimate goal of transitioning to fully driverless rides. The project will utilize a "sandbox" regulatory framework established through agreements between inDrive and Kazakh authorities throughout 2025 and 2026.
The collaboration is structured as a three-way division of labor. Pony.ai will supply its seventh-generation (Gen-7) autonomous vehicles, while inDrive will provide the existing ride-hailing platform and the user base. Nvidia will provide the underlying AI infrastructure, including its Hyperion platform and Halos safety system. According to TradingView, riders in Astana will not need to download a new application to access these services, as the technology will be integrated directly into the current inDrive app.
Why was Kazakhstan chosen for this project?
Kazakhstan offers a unique combination of regulatory readiness and harsh environmental testing grounds. The country has already implemented a structured legal framework for driverless transport as part of its national "Year of Digitalization and Artificial Intelligence." This legal foundation allows for controlled testing on specified public roads.
The specific choice of Astana provides a rigorous technical challenge for the autonomous systems. As the second-coldest capital in the world, the city experiences heavy snow and ice.
Astana has a population of approximately 1.7 million people, according to AOL, creating a substantial urban environment to assess how autonomous ride-hailing can complement existing mobility services.
How will the technology work under the hood?
The robotaxis will rely on the Nvidia Hyperion platform to manage the complex tasks of autonomous driving. This platform serves as a reference for both sensors and computing, designed to reduce the high costs typically associated with developing autonomous vehicles. By providing a pre-integrated foundation, Nvidia allows partners to avoid the expense of building every individual component from scratch.
The safety and processing stack includes several specific Nvidia components:
According to KuCoin, these cost-saving measures are particularly vital for emerging markets where autonomous services must remain price-competitive against traditional, inexpensive ride-hailing options.
- The Hyperion platform for sensor and compute integration.
- The Halos full-stack safety system.
- The Alpamayo Stack for operational support.
- Tools for generating synthetic data to test various driving scenarios.
Who are the key players involved in this deal?
The partnership brings together a ride-hailing veteran, a specialized robotics leader, and a global computing powerhouse. Each entity brings a specific market role and technical capability to the Kazakh pilot program.
The organizations involved include:
- inDrive: A global mobility platform operating in over 1,200 cities across 47 countries. It is known for a unique business model where riders and drivers negotiate fares directly.
- Pony.ai: A global leader in autonomous driving technology. According to KuCoin, the company had a fleet of 1,975 vehicles at the end of June and aims to operate over 3,500 robotaxis in 20 cities worldwide by the end of 2026.
- Nvidia: A semiconductor and AI infrastructure provider. The company is positioning itself as the foundational platform layer for the autonomous driving industry.
How does this fit into Kazakhstan's national AI strategy?
The robotaxi project is not an isolated event but part of a broader national push toward artificial intelligence and digitalization. The Kazakh government is actively building the hardware infrastructure necessary to support high-level AI operations across the country.
The deployment of advanced hardware in the region is already underway. According to TradingView, Kazakhtelecom has already installed 400 Nvidia H200 and 480 B300 GPUs in Astana and Almaty. To keep pace with technological demands, Kazakhtelecom expects to add more than 1,000 additional units, including 1,000 Vera Rubin GPUs, by 2027. This hardware buildup aligns directly with the timeline for the inDrive and Pony.ai robotaxi tests.
What happens next for autonomous mobility in the region?
The immediate future focuses on regulatory milestones and the establishment of specific test routes. The agreements signed between the companies and Kazakh authorities include plans for a potential pilot route originating from the airport in Astana.
Future developments will be measured by several key practical indicators:
- The official launch of the airport pilot route.
- The total number of vehicles that join the initial fleet.
- The determination of whether riders can use the inDrive haggling model to bid on driverless trips.
- The potential expansion of services to other Kazakh cities or other international markets where inDrive operates.
Frequently Asked Questions
When will robotaxis actually start operating in Kazakhstan? Testing is scheduled to begin in the first half of 2027 in the capital city of Astana.
Will the robotaxis be fully driverless from the start? No, the initial tests will include a safety operator in the vehicle before moving toward fully driverless operations.
How will the price of a robotaxi ride be determined? The service intends to use inDrive's existing model where riders propose a fare and drivers can accept, decline, or counter the offer.
Sources
- 1.tradingview.comInDrive plans 2027 robotaxi tests in Kazakhstan with Pony.ai and Nvidia
- 2.aol.cominDrive and Pony.ai Announce Collaboration to Explore Autonomous Mobility in Kazakhstan Built on NVIDIA Hyperion
- 3.cryptobriefing.comInDrive plans 2027 robotaxi tests in Kazakhstan with Pony.ai and Nvidia
- 4.kucoin.cominDrive to Test Robotaxis in Kazakhstan with Pony.ai and Nvidia in 2027


