A few years ago, Uber was widely seen as nothing more than a smartphone app for booking rides. Today, however, the company is redefining its identity. As artificial intelligence and autonomous transportation continue to evolve, the line between the digital and physical worlds is rapidly disappearing. Tech companies are no longer competing solely through software—they’re racing to power movement in the real world. Uber calls this next era Physical AI.
From refugee to the CEO of Uber
Uber CEO Dara Khosrowshahi fled Iran with his family at the age of nine following the Iranian Revolution. Forced to leave behind nearly everything they owned, the family started a new life in the United States. After successfully leading Expedia for years, Khosrowshahi made a surprising decision in 2017: he left a stable position to take over one of Silicon Valley’s most controversial companies. At the time, Uber was facing lawsuits, internal conflicts, and a severe loss of public trust.
Solving chaos like an engineering problem
Instead of trying to fix every problem at once, Khosrowshahi approached Uber’s crisis like an engineering equation. His philosophy is simple: break a complex problem into smaller components, solve each one independently, and the overall system becomes manageable. Using this framework, Uber focused on three priorities:
- ending conflicts within the board and among shareholders;
- rebuilding trust with regulators and drivers around the world;
- replacing legacy leadership with a new management team aligned around a shared vision.
A $1 trillion opportunity
Today, Uber no longer sees itself as just a ride-hailing platform. The company wants to become the infrastructure behind the next generation of autonomous transportation. According to Uber, the future market for autonomous vehicles, delivery drones, and robotics could exceed $1 trillion. The company’s AI systems have evolved dramatically as well. Uber says its search and recommendation models are now 10,000 times larger than previous generations. As a result, the app can predict where a user intends to travel with approximately 75% accuracy the moment it is opened.
The biggest challenge isn’t AI—it’s cost
Running frontier AI models is extremely expensive. In fact, Uber exhausted its annual AI budget during the first quarter alone. That experience led the company to adopt a two-stage AI strategy. For experimentation and rapid product development, Uber relies on cutting-edge models from companies like OpenAI and Anthropic. Once a feature proves successful at scale, it is migrated to more efficient open-source models that significantly reduce operating costs. This approach allows Uber to innovate quickly while maintaining long-term economic efficiency.
As autonomous transportation becomes mainstream, one question stands above the rest: Who will become the platform connecting the world’s autonomous vehicles? Uber believes the winner won’t necessarily be the company with the smartest AI model, but the one capable of connecting vehicles, logistics networks, drivers, and billions of users into a single ecosystem. That’s why Uber’s long-term vision extends far beyond ride-hailing. The company aims to connect autonomous cars, drones, and future Physical AI systems to one global Uber network. If successful, Uber could evolve from a transportation app into one of the world’s largest AI-powered mobility platforms.
