Voxel51 Partners With Porsche and VW Group on AI Driving Data

Voxel51 Partners With Porsche and VW Group on AI Driving Data

On October 21, Voxel51, a US-based company, announced a collaboration with Porsche, Volkswagen Group brands, and CARIAD to develop a shared artificial intelligence data system for automated driving. The initiative was announced at GTC Berlin in Germany, with Porsche leading the initial implementation alongside NVIDIA and Microsoft. The project is designed to connect real-world driving data with simulation, AI training, and testing in a common development framework. By bringing these activities together, the participating organizations aim to help automakers develop and evaluate advanced driving systems more efficiently while reducing reliance on repeated physical test drives.

How Volkswagen Group's AI Development System Works

Volkswagen Group is building an AI development system that links data collected during real-world driving with simulation, model training, and testing. The group's vehicle fleet generates data equivalent to the daily driving of approximately 11,000 average drivers, providing a substantial real-world information base for developing automated driving technology. The shared framework is intended to make that data more useful across the development process, helping teams prepare training material, recreate driving situations in simulation, and assess how driving systems respond. The partners expect this integrated approach to support faster development and more efficient testing across participating brands.

Roles of Voxel51, Porsche, NVIDIA, and Microsoft

Each participant contributes a distinct capability to the shared system. Volkswagen Group supplies connected vehicles, real-world driving data, vehicle platforms, and system requirements. Voxel51 organizes driving data into useful training and simulation material and helps identify scenarios in which AI models require improvement. NVIDIA provides tools for preparing training data, recreating real-world driving situations in simulations, and testing driving systems. Microsoft, through Microsoft Azure, supplies the cloud computing infrastructure needed to operate the system securely at scale. Porsche is helping develop the shared framework across Volkswagen Group brands, using NVIDIA technologies through Voxel51 on Microsoft Azure.

Targeted Reductions in Physical Testing and Regression Cycles

The collaboration aims to reduce repeated physical test drives by up to 90% and shorten AI regression testing cycles by up to 85%. These figures are stated targets for the shared development approach, rather than confirmation that the reductions have already been achieved across production programs. Reducing the need for repeated road testing could allow development teams to use simulation and prepared data more extensively when evaluating automated driving systems. Shorter regression cycles could also help teams repeat tests more efficiently as AI models change. The announcement does not provide a detailed measurement methodology, implementation schedule, or results demonstrating that the targeted reductions have been realized.

A Shared Framework for Automated Driving Development

The partners intend to create a common technology framework that helps automakers develop and test advanced driving systems faster and more efficiently. Its approach combines vehicle-generated data, data preparation, simulation, AI training, testing, and cloud infrastructure rather than treating these activities as separate tasks. Porsche is leading the initial implementation with NVIDIA and Microsoft, while the wider collaboration includes Volkswagen Group brands, CARIAD, and Voxel51. The announcement describes the intended structure and benefits of the project but does not specify a final deployment date, the complete set of participating vehicle programs, or whether the framework will eventually be extended beyond the current collaboration.

Industry Impact & Outlook

The collaboration illustrates how automakers and technology providers are combining real-world vehicle data, simulation tools, AI development, and cloud infrastructure to improve automated driving workflows. If the targeted reductions in physical testing and regression cycles are achieved, participating teams could spend less time repeating road tests and move through evaluation cycles more quickly. The shared framework may also help Volkswagen Group brands coordinate development activities, while Porsche's initial implementation provides the starting point described in the announcement. The next important evidence will be whether the system delivers the stated efficiency gains in practice; the announcement does not yet establish achieved results or confirm a broader rollout timeline.

Frequently Asked Questions

What did Voxel51 announce on October 21?
Voxel51 announced a collaboration with Porsche, Volkswagen Group brands, and CARIAD to develop a shared artificial intelligence data system for automated driving. The announcement was made at GTC Berlin in Germany. Porsche is leading the initial implementation alongside NVIDIA and Microsoft. The project connects real-world driving data with simulation, AI training, and testing in a shared development framework. Its stated objective is to help automakers develop and evaluate advanced driving systems more efficiently, with fewer repeated physical test drives and shorter AI regression testing cycles.

What role does Volkswagen Group play in the project?
Volkswagen Group contributes connected vehicles, real-world driving data, vehicle platforms, and system requirements to the shared AI development system. Its vehicle fleet generates data equivalent to the daily driving of around 11,000 average drivers. That information forms part of the real-world data base used in the development workflow. The system is intended to connect vehicle data with simulation, training, and testing so that participating teams can prepare and evaluate automated driving systems more efficiently. The announcement does not specify a separate deployment schedule for each Volkswagen Group brand.

What does Voxel51 contribute to the shared AI system?
Voxel51 organizes driving data into training and simulation material and helps identify scenarios where AI models need improvement. This role supports the process of turning real-world driving information into resources that development teams can use when training and evaluating automated driving systems. The shared framework also includes Volkswagen Group vehicle data, NVIDIA tools for data preparation and simulation, and Microsoft Azure cloud infrastructure. Voxel51 is therefore one of several contributors, with the overall system designed to connect data management and testing activities rather than relying on a single technology provider.

How much could the collaboration reduce physical test drives?
The partners aim to reduce repeated physical test drives by up to 90% through the shared AI development and testing framework. This is a stated target, not a verified result demonstrating that all participating programs have already achieved that reduction. The system connects real-world driving data with simulation, training, and testing, potentially allowing teams to evaluate more scenarios without repeating the same physical drives as often. The announcement does not provide the baseline, measurement method, or program-specific results needed to assess how the target will be calculated in practice.

How much could AI regression testing cycles be shortened?
The collaboration targets shortening AI regression testing cycles by up to 85%. Regression testing helps teams check whether changes to an AI model or driving system affect its previously evaluated behavior. The planned framework brings together driving data, simulation tools, testing capabilities, and cloud infrastructure to support this development process. The 85% figure is an intended reduction, and the announcement does not establish that it has already been achieved. It also does not provide a detailed testing methodology, a comparison period, or results for specific vehicle programs.

What are NVIDIA and Microsoft contributing to the project?
NVIDIA supplies tools for preparing training data, recreating real-world driving situations in simulations, and testing driving systems. Microsoft Azure provides the cloud computing infrastructure required to operate the system securely at scale. Porsche is helping develop the shared framework across Volkswagen Group brands using NVIDIA technologies through Voxel51 on Microsoft Azure. These roles combine simulation and testing capabilities with cloud infrastructure, alongside the vehicle data supplied by Volkswagen Group and data organization performed by Voxel51. The announcement does not specify the complete technical architecture or the full list of tools to be deployed.