Veuron Technology Wins $4.8M Government R&D Grant to Develop and Test End-to-End Autonomous Driving Technology Globally

Veuron Technology, an AI-driven autonomous driving startup specializing in LiDAR-based perception, has been selected for the Ministry of SMEs and Startups’ 2026 Global TIPS program, securing total project funding of $4.8 million (₩6.4 billion) over four years, the company announced on August 18.

The selected project, titled “Development of World Foundation Model (WFM)-based End-to-End Autonomous Driving Technology and Global Real-Vehicle Validation,” will receive $3.75 million (₩5 billion) in government R&D funding. KAIST will participate as a joint research institution. The project will be managed by E&Venture Partners, which will also support follow-on fundraising and global commercialization strategy.

Global TIPS is a program that combines private investment and government R&D support to help promising small and medium-sized Korean technology companies enhance their capabilities and expand into international markets. The Korea Automotive Technology Institute (KATECH) will provide technical and regulatory consulting on autonomous driving standards, testing and certification.

The project aims to overcome two fundamental limitations of current autonomous driving development approaches. Traditional modular systems divide perception, prediction, decision-making, path planning and control into separate functions, each requiring independent development and validation. Integration and iterative testing of these components into a unified system demands substantial time and cost.

End-to-End (E2E) architectures reduce integration overhead by processing all driving functions within a single AI model, but face challenges in acquiring sufficient training data for complex or rare driving scenarios. Because decision-making occurs inside a unified model, understanding “why the system made a particular decision” is difficult, complicating root-cause analysis and safety certification when issues arise.

Veuron plans to address these challenges by combining its existing E2E technology base with a World Foundation Model (WFM). A WFM is an AI model trained at scale on how the physical world operates, enabling it to generate and predict driving scenarios the system has never directly encountered. The understanding and reasoning capabilities derived from the WFM will be transferred to the onboard driving AI deployed in vehicles.

The company will also broaden the input data streams. Veuron will integrate LiDAR data—its core technical strength—with camera imagery and vehicle control signals to improve perception and decision-making performance in complex road environments.

The training and validation framework will be redesigned around a hybrid learning approach that combines synthetic driving scenarios generated by the WFM with real-world driving data collected on public roads. Vision-language models (VLM) and vision-language-action (VLA) techniques will be employed to describe camera-perceived scenes in natural language and link those descriptions to corresponding driving actions, structuring the rationale behind each decision into analyzable data.

The resulting model will be optimized to run in real time on edge devices installed in vehicles. Validation will follow a closed-loop methodology in which the vehicle’s decisions and control outputs directly influence the next state of the simulated driving environment, mirroring real-world cause-and-effect dynamics. This iterative closed-loop testing is designed to repeatedly verify driving performance and safety under conditions where each decision alters the subsequent scenario.

The research will deploy Veuron’s existing technology assets: Copilot, the company’s ADAS and autonomous perception solution; X Factory, an AI LiDAR platform covering data collection, processing, development and deployment; Prospector, a high-precision sensor-fusion data construction system; and Titan, Veuron’s E2E autonomous driving model. KAIST will contribute validation and testing infrastructure from its Mobility AX Lab and Eco-Friendly Smart Vehicle Research Center, as well as remote driving, ultra-low-latency communication and cooperative vehicle-to-everything (C-V2X) autonomous vehicle control technologies.

Real-vehicle validation will begin in South Korea and expand to major cities in the United States, Germany and Australia. The company plans to test its E2E autonomous driving technology in a total of ten cities worldwide, verifying technical maturity, product readiness and market viability across diverse national regulations and road environments.

Leveraging the outcomes of the Global TIPS project, Veuron intends to expand its E2E autonomous driving software licensing business targeting original equipment manufacturers (OEMs) and broaden deployment in robotaxi services. The driving data and edge-case database accumulated during validation will also be commercialized separately.

“This Global TIPS selection marks a critical inflection point for Veuron as we scale our accumulated perception technology, data and onboard AI capabilities into a unified E2E autonomous driving system,” said Jaekwang Kim, CEO of Veuron Technology. “We will prove on real roads in ten cities across different countries and driving environments that our technology operates reliably under diverse conditions, and connect that validation directly to commercialization with robotaxi operators and global OEMs.”

Veuron Technology has built an end-to-end technology portfolio spanning perception solutions, data infrastructure, AI model development and vehicle deployment for autonomous driving. With this project, the company plans to move beyond supplying individual perception solutions and position itself as an autonomous AI company capable of self-learning and scaling in global operating environments.

South Korea’s autonomous mobility sector has seen accelerating technical and commercial development across multiple fronts. In August 2026, WEFLO demonstrated its automated navigation-system testing equipment for commercial drones at the Commercial UAV Expo in Las Vegas, showcasing infrastructure for GPS-denied autonomous navigation validation. Earlier in the year, electric mobility subscription platform Cycloid raised $360,000 from AXE Ventures to expand its lifecycle-management service for delivery riders, digitizing vehicle operations from purchase through insurance and maintenance.