Decentralized Data Fuels Open-Source Self-Driving Revolution

HAMBURG, Germany – April 13, 2026 – In a significant move to accelerate the development of open-source autonomous driving, NATIX Network has joined the Autoware Foundation as a Premium Member. The Hamburg-based company will supply its vast, decentralized collection of multi-camera driving data to power the creation of Autoware’s first open-source end-to-end (E2E) autonomous driving model, tackling one of the most persistent bottlenecks in the industry: the scarcity of diverse, real-world data.

This collaboration brings together NATIX, which operates the world's largest decentralized camera network for Physical AI, and the Autoware Foundation, a non-profit organization that marshals industry giants like AMD, AWS, Arm, and TIER IV to build a leading open-source software stack for autonomous vehicles. The partnership aims to democratize access to the critical data resources needed to train, test, and validate the next generation of self-driving systems, potentially leveling the playing field in a sector long dominated by proprietary, closed-off ecosystems.

Confronting the Great Data Bottleneck

For years, the path to autonomous driving was paved with modular systems, where distinct components for perception, planning, and control were engineered separately. While foundational, this approach has shown its limits when faced with the immense complexity of real-world driving, struggling to scale across new environments or react to the countless rare “edge cases” that defy pre-programmed rules. In response, the industry is pivoting toward a more holistic, data-centric paradigm: end-to-end (E2E) AI.

E2E models function more like the human brain, learning the entire driving task—from seeing the road to controlling the vehicle—as a single, unified policy trained directly on experience. This method promises more adaptable and scalable systems, but it comes with a voracious appetite for data. To achieve a level of safety that is significantly better than a human driver, these AI systems must be exposed to an astronomical amount of diverse driving scenarios.

This requirement has crystallized into what industry leaders call the “data problem.” As Tesla CEO Elon Musk has noted, the hardest challenge is not getting an autonomous system to “sort of work,” but closing the final gap to superhuman safety, a feat that requires immense volumes of real-world data. Access to such large-scale visual datasets has become one of the biggest barriers to progress, particularly for the open-source community, which often lacks the resources to deploy massive, proprietary data-collection fleets like those operated by Waymo or Cruise.

A New Model for Data Acquisition

NATIX is addressing this data challenge with an innovative and disruptive approach: a Decentralized Physical Infrastructure Network (DePIN). Instead of operating its own fleet, the company leverages blockchain technology and token incentives to crowdsource data collection from a global community of participants. Through its smartphone app “Drive&” and its VX360 device for Tesla vehicles, NATIX has enlisted over 270,000 drivers who have collectively covered more than 250 million kilometers.

This decentralized model provides two key advantages: scale and diversity. The network gathers multi-camera, 360-degree footage from vehicles operating across multiple continents, capturing a rich tapestry of different countries, road types, weather conditions, and traffic patterns. This variety is precisely what E2E models need to become robust and generalizable, allowing them to learn from a global spectrum of driving realities rather than a geographically limited set of test routes.

By distributing the cost and effort of data collection across a wide network, the DePIN model offers a more scalable and cost-effective alternative to the capital-intensive, centralized data acquisition strategies that have defined the industry to date. It represents a fundamental shift in how the foundational resources for AI can be gathered and shared.

Powering Open-Source Autonomy

The Autoware Foundation stands to be a primary beneficiary of this new data paradigm. Autoware is already making significant strides in its transition toward an E2E architecture, with ongoing development of learned planning algorithms and machine-learning-based perception. The foundation is pursuing a pragmatic hybrid approach, combining the adaptability of E2E models with a rule-based safety layer to ensure reliability for real-world deployment.

NATIX’s contribution will be instrumental in this effort. “Open-source autonomy needs access to serious real-world data as the industry moves toward end-to-end AI,” said Alireza Ghods, CEO and co-founder of NATIX. “By joining the Autoware Foundation and contributing global multi-camera driving data, we are helping the community train and validate the next generation of autonomous driving systems.”

The data will serve multiple purposes within the Autoware ecosystem. It will be used to directly train E2E driving models, validate system performance across different regions, and, crucially, develop sophisticated “world models.” These models help an AI system build an intuitive understanding of how the world works, enabling it to simulate and predict future events.

“NATIX brings a large and diverse open-source data set which covers many long-tail edge case scenarios,” said Muhammad Zain Khawaja, Managing Director of Product at the Autoware Foundation. “This allows us to find the data that matters most on roads across the world to build safe, introspectable, and generalizable End-to-End AI models.”

Building Worlds for a Smarter Future

The impact of this data extends beyond direct model training into the realm of advanced simulation. The development of world models, a key focus for Autoware, represents a frontier in AI research. These generative systems can use real-world data as a foundation to create millions of realistic virtual scenarios, allowing developers to test their autonomous systems against rare and dangerous situations—like a child chasing a ball into the street—without ever putting a real vehicle at risk.

Reinforcing this strategy, NATIX has also entered a partnership with automotive supplier Valeo to build one of the largest open-source, multi-camera World Foundation Models (WFMs). By feeding NATIX’s global 360° driving data into Valeo's generative AI expertise, the project aims to create a powerful simulation engine for the entire Physical AI ecosystem, benefiting not only autonomous driving but also robotics and other applications that require an AI to interact with the physical world.

By supplying the foundational data for these next-generation tools, the collaboration between NATIX and Autoware is not just about building a better self-driving car; it’s about constructing a more robust, accessible, and collaborative foundation for the future of intelligent machines.