Zachary Greenberger has spent his career in connected mobility, from working on Tesla’s Model 3 to six years at Lyft, and now as CEO of Nexar and Nauto.
Today, their network includes roughly 350,000 connected cameras capturing around 300 million miles of road every month.
But this conversation is not really about dash cams.
It is about what happens when AI leaves the screen and enters the physical world.
In this episode of Aggaeus, Haggai Klorman-Eraqi sits down with Zachary Greenberger, CEO of Nexar and Nauto, for a conversation about autonomous vehicles, physical AI, real-world data, leadership, and what machines still need to learn before they can safely operate around us.
For years, autonomous vehicles were trained heavily through simulation.
But eventually simulation reaches a limit.
A traffic cone moves. A child rolls a ball into the street. Debris flies across a highway. A road changes overnight.
The real world does not behave like a clean simulation.
That insight changed how Zach and the Nexar team saw the company.
What started as a network of connected cameras became something much bigger: a way to capture the rare, messy edge cases that autonomous systems need to understand.
That led to BADAS, Nexar’s Beyond ADAS world model.
And something surprising happened.
The same model trained around driving could identify collision risk in drones, forklifts, delivery robots, humanoid robots, and other physical environments.
The lesson was bigger than cars.
If AI is going to move through the physical world, it needs to understand how the physical world actually behaves.
In this episode, we cover:
• Why real-world data becomes critical as autonomous systems improve
• Why edge cases may be more valuable than millions of ordinary miles
• How BADAS moved beyond cars into robots, drones, and other physical systems
• What Zach learned taking over a founder-led company and later bringing teams together
• Why “debate hard, but be kind” is central to how he builds culture
This is a conversation about autonomy, physical AI, data, leadership, culture, and why teaching machines about reality may require something simulation alone can never provide: reality itself.
Robots Need the Real World with Zachary…
Zachary Greenberger has spent his career in connected mobility, from working on Tesla’s Model 3 to six years at Lyft, and now as CEO of Nexar and Nauto. Today, their network includes roughly 350,000 connected cameras capturing around 300 million miles of road every month. But this conversation is not really about dash cams. It is about what happens when AI leaves the screen and enters the physical world. In this episode of Aggaeus, Haggai Klorman-Eraqi sits down with Zachary Greenberger, CEO of Nexar and Nauto, for a conversation about autonomous vehicles, physical AI, real-world data, leadership, and what machines still need to learn before they can safely operate around us. For years, autonomous vehicles were trained heavily through simulation. But eventually simulation reaches a limit. A traffic cone moves. A child rolls a ball into the street. Debris flies across a highway. A road changes overnight. The real world does not behave like a clean simulation. That insight changed how Zach and the Nexar team saw the company. What started as a network of connected cameras became something much bigger: a way to capture the rare, messy edge cases that autonomous systems need to understand. That led to BADAS, Nexar’s Beyond ADAS world model. And something surprising happened. The same model trained around driving could identify collision risk in drones, forklifts, delivery robots, humanoid robots, and other physical environments. The lesson was bigger than cars. If AI is going to move through the physical world, it needs to understand how the physical world actually behaves. In this episode, we cover: • Why real-world data becomes critical as autonomous systems improve • Why edge cases may be more valuable than millions of ordinary miles • How BADAS moved beyond cars into robots, drones, and other physical systems • What Zach learned taking over a founder-led company and later bringing teams together • Why “debate hard, but be kind” is central to how he builds culture This is a conversation about autonomy, physical AI, data, leadership, culture, and why teaching machines about reality may require something simulation alone can never provide: reality itself.