The major players in the autonomous trucking space all agree on one thing: driverless operations will revolutionize the American trucking industry. What they don’t agree on, however, is exactly how that revolution becomes a viable commercial product. 

In San Francisco, voices from Bot Auto, Torc Robotics, and Kodiak came together to pitch their vision for the autonomous future. The panel at AUTONOMOUS: The Future of Robotics & Physical AI, was made up of Bot Auto Founder & CEO Xiaodi Hou, Kodiak Head of AI Shubbham Shrivastava, Torc CCO Andrew Cullhane, Anyscale Field CTO Christian Stano, and was moderated by Kishor Veerashekar, investor at Plug and Play Ventures. 

The panelists debated software frameworks, OEM integration friction, and edge-case collection, but Dr. Hou presented a sharp, pragmatic framework for real-world profitability and true autonomy.

Bot Auto Uses Transportation-as-a-Service to Max Efficiency

One of the central debates in autonomous trucking is whether to operate strictly as a software vendor licensing technology to third parties or to run as a true trucking carrier. Dr. Hou unequivocally advocated for the integrated carrier model as the fastest, most direct path to commercial profitability.

"Think of an autonomous truck like a high-end espresso machine. If you buy a $5,000 espresso machine at home and only use it for five minutes every morning, it's a terrible asset. The only way to capitalize on that level of technological utilization is to run it 24 hours a day. By operating our own fleet at Bot Auto, we ensure maximum asset utilization." 

Bot Auto currently operates a fleet of trucks upfit with its physical AI stack and internally developed software. Customers seek them out to move freight on lanes across Texas.

Hou highlighted that licensing software to traditional fleets often leads to autonomous trucks being treated as secondary, backup capacity. To unlock true economic value, Bot Auto owns and operates its trucks, co-engineering software and fleet operations to guarantee continuous 24/7 deployment on commercial lanes. 

The Goal is to Make Money on Every Mile Today

Freight is an unemotional industry strictly governed by total cost of ownership (TCO). Bot Auto is consistently aiming to be below the industry average for cost per mile of moving goods.

By eliminating human labor constraints and maximizing vehicle uptime on active commercial corridors like Houston-to-Dallas, Bot Auto generates positive unit economics right now. As Hou noted during the discussion, "we are making money in every mile rather than losing money in every mile. If you scale without being ready, it leads to terrifying financial results."

"Data is the Past, Compute is the Future"

For years, autonomous vehicle startups have claimed that accumulating petabytes of driving data builds an unassailable competitive moat. Dr. Hou directly challenged this outdated narrative, placing Bot Auto at the forefront of modern generative and foundation AI architecture.

In 2026, raw data collection and manual annotation are no longer differentiators. Operating autonomous trucks 20 hours a day produces a functionally infinite stream of routine driving data. The true competitive advantage lies in compute power and automated processing—using advanced AI models to automatically extract, filter, and learn from rare edge cases without human intervention.

"Data is the past; compute is the future. Manually annotating data gave companies an advantage five years ago. Today, deep learning thrives on unstructured data. What matters is leveraging compute to turn raw data into actionable models with minimal human involvement—turning a workshop into a next-generation factory."

True Driverless Scale Accounts for Safety Response Time

A critical portion of the panel focused on defining what actually constitutes "driverless" operation. Dr. Hou urged the industry to abandon misleading definitions, such as keeping safety observers in passenger seats or relying on high ratios of remote fallback operators.

Hou advocates for the industry to establish a true commercial standard of “driverless” and to adhere to measuring things like the “Safety Response Window”. 

A Safety Response Window is the amount of time an AV can safely navigate a complex, unexpected scenario without human intervention. Most driverless systems have a response window of less than 10 seconds; the Bot Auto team has intentionally given the system a 60-second window to extend autonomy.

This extended time allows a single remote assistance team to oversee large fleets effortlessly, operating as a unit akin to air traffic control instead of one-to-one human to vehicle monitoring. 

The Road Ahead with Bot Auto

As the race to scale autonomous freight accelerates, Bot Auto stands out by combining integrated carrier operations, positive unit economics, cutting-edge compute infrastructure, and rigorous safety standards on active commercial lanes. 

While the other routes of OEM partnerships and software licensing could lead to viable autonomous scale, real fleet utilization and genuine driverless independence, sets a shorter timeline from deployment to profitability. 

Bot Auto is setting the definitive benchmark for the future of transportation.

This article is a summary of “The Race To Driverless Freight”, a session of the AUTONOMOUS: The Future of Robotics & Physical AI conference.