DeepRoute.ai Unveils Scalable DeepRoute IO 2.0 at IAA Mobility 2025

DeepRoute.ai, a pioneer in autonomous driving technology, has presented its next-generation platform, DeepRoute IO 2.0, at the IAA Mobility 2025 event in Munich. This launch marks a significant milestone in the company’s ambition to bring safe and adaptable intelligent driving solutions to the global market. The platform is powered by a cutting-edge Vision-Language-Action (VLA) model designed for mass production.

With over 100,000 passenger vehicles already equipped with its autonomous driving system, DeepRoute.ai is also aiming to expand its robotaxi services using mass-produced vehicles. Such accomplishments reflect the company's growing capabilities in large-scale production and engineering maturity when it comes to adaptation efficiency, system stability, and safety standards.

DeepRoute.ai is actively pursuing expansion into key international markets, including Europe, Japan, and South Korea, driven by a dual globalization strategy. On one hand, the company utilizes its extensive mass production experience and proven technology to assist domestic partners, such as smart, in accelerating their offerings of intelligent vehicles abroad. On the other hand, DeepRoute.ai delivers localized and highly adaptable assisted driving solutions to foreign automakers and joint ventures, aiding partners in penetrating broader market segments.

Maxwell Zhou, the CEO of DeepRoute.ai, emphasized that their VLA-based platform combines intelligence with adaptability, making it particularly suitable for European roads. The DeepRoute IO 2.0 platform offers exceptional flexibility, compatible with configurations using LiDAR, pure vision, and multiple automotive-grade chips. Its modular architecture allows automakers to implement intelligent driving functions across various models and vehicle segments, meeting diverse market demands and facilitating global adoption.

At the heart of the platform, the VLA model incorporates multimodal vision, language, and action, supported by chain-of-thought reasoning for advanced temporal and causal inference. Compared to traditional models, VLA provides high interpretability and traceable decision pathways, fostering trust through transparency.

Capitalizing on the extensive capabilities of the language model, VLA integrates a comprehensive knowledge base along with continuous learning capacity, enabling quick adaptation to different road systems, signage, and driving cultures. This model showcases rapid generalization and resilience, offering practical usability in Europe’s most challenging driving environments.

Additionally, VLA enhances safety with spatial awareness, logical explanation, optical character recognition (OCR), and voice control. These features enable proactive responses to high-risk scenarios, such as blind curves, truck blockages, or complex intersections, executing defensive driving maneuvers reminiscent of human behavior, including early deceleration and intelligent diversion strategies.

By merging intelligence, adaptability, and safety, DeepRoute.ai is propelling the industry toward a future of smarter and more reliable global mobility, defining a new era of intelligent transport.

About DeepRoute.ai

DeepRoute.ai is an AI-focused company dedicated to developing and applying intelligent driving solutions. The company is a trailblazer in the utilization of end-to-end and VLA models for passenger vehicles. During IAA Mobility 2025, DeepRoute.ai showcases its innovations at Hall A2, Booth D25, and its partner will deliver a keynote speech on September 9 at the B2 Summit stage. For more information, visit deeproute.ai.