Valeo and Natix Network Develop WFM for the Autonomous Vehicle Revolution

A new collaboration between Valeo and Natix Network marks a significant turning point in the journey toward safe and reliable vehicle automation. Both companies have joined forces to develop the WFM (World Foundation Model), a multi-camera artificial intelligence model that surpasses the limitations of conventional perception-based approaches. WFM is designed to learn and predict real-world movements while adapting to a wide range of traffic scenarios, opening transformative opportunities in the global automotive industry.

WFM: The World Foundation Model Changing the Landscape of Autonomous Driving

The World Foundation Model is not just an evolution of previous AI technology. WFM represents a qualitative leap in how vehicle automation systems understand the physical environment. Unlike AI systems that rely solely on pattern recognition based on static text or images, WFM integrates predictive capabilities to anticipate complex traffic dynamics.

Alireza Ghods, one of the founders and CEO of Natix, places WFM in an interesting historical perspective. He sees this model as a generational moment comparable to the explosion of large language models between 2017 and 2020. According to Ghods, the team that successfully builds the first scalable world model will lay the foundation for the next wave of AI: Physical AI. This is not merely an incremental upgrade but a fundamental transformation in how machines interact with their environment.

Decentralization and Open Source Strategy in WFM Development

Valeo and Natix’s commitment to transparency sets their approach apart from competitors. Both partners pledge to release the WFM model, training datasets, and development tools openly to the global developer community. This strategy enables the ecosystem to grow faster and undergo broader testing across various real-world conditions.

This move aligns with the philosophy of DePIN (Decentralized Physical Infrastructure Network), which integrates blockchain technology with community-managed physical infrastructure. In the Solana-based ecosystem operated by Natix, participants can contribute computing resources and earn rewards in cryptocurrency. This decentralized approach allows for unprecedented scale testing of WFM, with hundreds of thousands of contributors and hundreds of millions of kilometers of driving data accumulated.

Marc Vrecko, CEO of Valeo’s Brain Division, emphasizes that the main goal is to advance mobility intelligence in a safe and responsible manner. The transparency framework, he says, facilitates faster waves of innovation while maintaining the highest safety standards—crucial factors in the automotive industry.

Wayve: Proof of Concept of WFM in Real-World Applications

Autonomous vehicle startup Wayve has been an early adopter of WFM. In an impressive trial, a WFM-powered vehicle successfully navigated parts of Las Vegas without prior training in the city. This achievement demonstrates the model’s predictive capabilities to generalize spatial understanding across different environments. This demonstration is not just a technical milestone but validation that WFM can be applied in complex real-world scenarios.

WFM vs Alpamayo: Competitive Dynamics in Physical AI

The landscape of foundational model development for vehicle automation is not without competitors. Nvidia has launched Alpamayo, an open-source vision-language-action model leveraging camera and sensor data for autonomous reasoning-based decision-making. Alpamayo shows that major tech industries are also racing in the arena of Physical AI.

However, Valeo and Natix’s approach through WFM offers a unique philosophy rooted in decentralization and more radical openness. While Nvidia provides proprietary solutions with limited access, WFM is driven by a community collaboration spirit through decentralized infrastructure.

Roadmap and Expectations: When Will WFM Be Ready for Launch?

According to a Natix spokesperson, the first version of WFM is projected to be ready in the coming months. This timeline reflects significant development momentum, although technical challenges in building a multi-modal predictive model remain substantial.

The relevance of WFM extends beyond the automotive sector. Developing a robust world foundation model opens doors for Physical AI applications across various industries—from robotics to smart infrastructure. The success of WFM will determine the acceleration of mainstream adoption of self-driving vehicles and set new standards for the next generation of AI technology.