More than 100 million road accidents occur globally every year, imposing an estimated $871 billion annual cost through lost productivity, medical care, and secondary economic impact. Those number are widely understood to be an undercount, as road crashes are systematically under-reported worldwide, a reality long acknowledged by the United Nations and the World Health Organization.
To Omar Mukhtar, 100 million road accidents a year is both a diagnosis and a design brief.
As founder and chief executive of Omar Autonomous, a Silicon Valley–based Physical AI and robotics company developing a scalable modular autonomous platform combining LiDAR and camera-based intelligence for all vehicles, existing and new, he believes the persistence of road trauma is less a mystery than a structural mismatch between human capability and machine speed.
“Usain Bolt, the fastest human ever, reached a top speed of 27.8 miles per hour (44.72 kilometers per hour),” Omar says. “Yet everyday driving routinely happens at double that speed. In rain, fog, glare, or even a split-second distraction, the limits of human perception and reaction are quickly exposed.”
94% of crashes can be traced back to human error, according to the U.S. National Highway Traffic Safety Administration (NHTSA). Safety improvements therefore must focus on systems that can sense, predict, and intervene before impact, shifting safety from reaction to anticipation.
“I want to be at a point in five to ten years, and if I had a magic wand, ideally today, where every car is safe, and instead of 100 million crashes a year, we are aiming for zero.” For Omar, this is where Physical AI and robotics developed by his company come in as a practical upgrade to perception and decision-making.
The claim is not that autonomy makes vehicles infallible. It is that a well-engineered, deployable system can narrow the gap between what the road demands and what people can reliably do under fatigue, weather, and distraction. In his view, the real constraints now sit outside the algorithm, centered on whether the technology can be made scalable, trustworthy, and widely available, rather than confined to a handful of premium models.
Why a Century of Cars Still Produces 50 Million Injuries and $871 Billion in Losses Every Year
Automobiles have been on the road for more than a century, and for much of that time, progress prioritized making cars faster and more widely available, only later adding safety measures designed to reduce harm after a crash rather than prevent one in the first place.
“Seat belts were revolutionary,” Omar says, referencing Volvo’s decision in the 1960s to make its patent freely available, a move that helped save tens of millions of lives by embedding safety directly into the physical system of the vehicle.
But passive protection has limits. Every year, there are more than 100 million road accidents globally, including over 14 million in the United States alone. More than 50 million people are injured annually, many dealing with long-term physical, emotional, and financial consequences. The economic damage is staggering, around $871 billion a year in the U.S., once lost productivity, medical costs, and secondary impacts are accounted for.
The most frustrating part is that 94% of these crashes are caused by human error, according to NHTSA: distraction, fatigue, slow reaction times, glare, low light, and the fundamental limits that come with being human. Studies by U.S. safety authorities show that between 2011 and 2021, fatal accidents increased by around 30%, despite the growing presence of so-called driver assistance features. Many of these systems function as luxury additions rather than genuine safety mechanisms.
“The intention is there,” Omar says. “But the results are not.”
Human Reaction Fails at Road Speed, It Would Fail Even for Usain Bolt
At the core of the issue is physics. Humans evolved to move at walking or running speeds. Cars operate far beyond that envelope. At highway speeds, even milliseconds matter, and no amount of training can overcome biological limits.
Omar began building AI products in 2004, initially in AI-based stock market prediction and large-scale scheduling systems. His work has spanned startups as well as world-scale production environments at Amazon and Microsoft, where reliability, latency, and real-world constraints were non-negotiable. That experience now defines his approach to autonomy.
Modern autonomous safety, in his view, depends on combining complementary forms of intelligence in physical AI systems operating in the real world. Omar Autonomous integrates LiDAR for precise physical measurement with cameras that interpret context such as traffic signals, signage, and pedestrian behavior.
“LiDAR gives you physics,” Omar says. “Cameras give you semantics. Together, they allow the system to understand not just what is there, but what it means.”
Unlike traditional radar-based systems with narrow fields of vision, AI-driven perception enables near 360-degree awareness. More importantly, it allows prediction. “Two cars approaching an intersection may not be a danger right now,” Omar says. “But an AI system can calculate that in five seconds they will occupy the same trajectory.”
Road Safety Is a Scale Problem, Not a Feature Problem
One of Omar’s sharpest critiques of the autonomous sector is its historical focus on bespoke vehicles and controlled environments. “You can build something beautiful in a lab,” he says. “If it cannot be deployed widely, it does not solve the problem.”
Omar Autonomous takes a modular platform approach, enabling autonomy to be deployed on both existing and new vehicles in minutes rather than years. By targeting the global vehicle fleet, the company aims to meet customers where they already are.
“Safety cannot be a luxury,” Omar says. “It has to be accessible in price, availability, and installation.”
This goes beyond engineering. Organizations working on safety-critical AI must begin with genuine care for human outcomes. “This is not an intellectual exercise,” he says. “Human life is the most precious thing we have. You have to demand more from yourself than any regulator ever could. That obsession is what closes the gap between intention and impact.”
Why the Next Breakthrough in AI Will Be Physical AI and Robotics
While much attention has focused on generative AI, software tools, and digital productivity, Omar believes the next major breakthrough will come from Physical AI and robotics, systems that act autonomously in the physical world, where failure carries real human cost.
“This is where AI can save lives,” he says, citing applications not only on roads but also in hospitals, factories, and other high-risk environments.
For Omar, transportation is ultimately a means rather than an end. Safer mobility enables commerce, connection, freedom, and empowerment. It allows people to reach loved ones, access healthcare, and participate fully in society.
Eliminating 100 million crashes a year is deliberately bold, not because it is speculative, but because the problem is already understood and the solution already exists. Ninety-four percent of crashes are caused by human error, a failure mode that Physical AI was built to replace. At Omar Autonomous, that intelligence is operating in the real world, systems that can see, predict, and act faster than humans, designed to scale across existing and new vehicles alike. Ending these failures does not just prevent tens of millions of injuries; it removes an $871 billion annual drag on the economy and restores safety as a baseline, not a privilege. This is not a question of invention or belief. The outcome is settled: Omar Autonomous scales proven systems to replace preventable failure everywhere it still exists.
Follow Omar Mukhtar on LinkedIn for ongoing work at the intersection of Physical AI, robotics, and real-world safety.
