Natix and Valeo are building a decentralized, self-driving camera model to offer a transparent foundation for the safe mainstream deployment of physical AIs.
Automotive technology supplier Valeo and Natix Network, a Solana-based decentralized physical infrastructure network (DePIN) provider, have partnered to build an open-source, artificial intelligence (AI) multi-camera model aimed at improving autonomous driving systems.
The two companies said Thursday they are building what they call the World Foundation Model (WFM), which they say will be able to learn and predict real-world motion, while adapting to traffic situations.
The multi-camera model aims to push the boundaries of AI models from text-based understanding to real-world scenarios in physical environments, according to an announcement shared with Cointelegraph. It also aims to advance research into autonomous driving.
Valeo and Natix pledged to openly release their models, datasets and training tools to enable developers to fine-tune these capabilities. The first version of the WFM model is expected to be ready within the next couple of months, a spokesperson for Natix told Cointelegraph.
WFM to accelerate advent of self-driving vehicles
Autonomous driving startup Wayve is already using WFMs in its vehicles, including a test in which a vehicle navigated parts of Las Vegas with no prior training in the city, according to material shared by the company’s CEO, Alex Kendall, on Friday.
WFM is part of the wider DePIN sector, which merges blockchain technology with community-owned physical infrastructure to create decentralized networks where participants can contribute resources, such as computing power, in exchange for cryptocurrency.
The ultimate goal of the WFM self-driving camera model is to safely and responsibly “advance mobility intelligence” and the advent of autonomous vehicles, said Marc Vrecko, CEO of Valeo’s Brain Division.
Alireza Ghods, co-founder and CEO of Natix, sees WFMs as a generational opportunity similar to the rise of large language models (LLMs) from 2017 to 2020.
“The teams that build the first scalable world models will define the foundation of the next AI wave: Physical AIs.”
Unlike current perception-only AI models, the multi-camera world model seeks to build predictive capabilities to accelerate the mainstream deployment of self-driving vehicles.
Natix said decentralizing and open-sourcing the WFM could allow physical AI systems to be trained and tested across a wider range of real-world conditions before deployment. “Transparent frameworks allow the ecosystem to move faster,” a company spokesperson said, adding that extensive testing is critical for safety.
Competitive landscape and scale of Valeo-Natix bet
One of Valeo’s and Natix’s main competitors is Alpamayo, a family of open-source vision-language-action models launched by chipmaking giant Nvidia. The solution uses camera and sensor data for decision-making through reasoning-based autonomy.
Natix was founded in 2020 and operates a decentralized multi-camera data network which, according to industry research firm Messari, includes hundreds of thousands of contributors and hundreds of millions of kilometers of recorded driving data.
