Seven lessons for every robotics founder from the ‘godfather of self-driving cars’

Sebastian Thrun has been described as “the godfather of self-driving cars.” In 2005, he built a robot named Stanley that won the DARPA Grand Challenge in 2005 after traversing a difficult 132-mile course in the Mojave Desert in under 10 hours. It was a stunning success in the robotics world, and Stanley the robot is now on display at the Smithsonian.

Watching the race was none other than Larry Page, co-founder of Google, who introduced himself to Sebastian. It was a sliding doors moment where the two struck up a connection, and shortly thereafter, Sebastian became the co-founder of Google X and led many moonshots for the team—including a secret self-driving car project in 2009, which would eventually become Waymo.

Beyond this, Sebastian’s list of accomplishments is long. He’s published over 380 academic papers. He launched one of the world’s first MOOCs (massive online open courses) as a co-founder of Udacity. He was a National Academy of Engineering member at age 39 and is still an adjunct professor at Stanford.

We sat down with Sebastian at our first-ever Robotics Day in San Francisco, and he shared seven insights from 20 years of experience in robotics—including some spectacular successes, and also some notable failures.

Key lessons for robotics founders from Sebastian Thrun

• Building in robotics is an uphill battle by design. The founders who get there fastest will move through mistakes, not around them. Building here means making bad decisions most of the time, and you can only gain hindsight from moving forward.
• The hardware-software divide is the robotics industry's most overlooked problem. Most builders are too specialized in one discipline to see the full picture. The strongest teams bridge both worlds, understanding both software and hardware.
• Building a wedge isn’t always necessarily easier. The incremental path feels safer but is often harder with less capital, less talent, and thinner margins. Ambitious goals attract the resources needed to actually succeed.
• Know what actually motivates you beyond financial reward. Financial reward fades as motivation. Founders who stay energized through the long middle are driven by something deeper. Getting clarity on that early enough makes all the difference.
• Be arrogant about the goal, humble about the path. Founders: hold your destination firmly, but stay genuinely open about how you get there. Ego is what fractures companies from the inside.
• Genuinely befriend the regulators — “They mean well, the same way you mean well.” Regulators want safety and innovation, too. Bring them along early, and they’ll want you to win.
• Even the savviest builders can't predict what's coming next. Stay humble about what’s coming, as the next breakthrough will blindside everyone.

1. Building in robotics is an uphill battle by design. The founders who get there fastest will move through mistakes, not around them.

Many in the industry are anticipating a “ChatGPT moment for robotics,” where mass adoption proliferates. Sebastian is optimistic it's on the horizon. But he's clear-eyed that general-purpose robotics has a longer road before they break into the mainstream.

One of the biggest differences between building software and building in robotics is the tolerance for error. “If an LLM hallucinates, you say 'that's wrong.' Then you shrug and move on," says Sebastian. In robotics, the stakes are categorically different. “You have a device that weighs a tonne travelling at 60 miles per hour in downtown San Francisco. What could possibly go wrong?”

This is true across most robotics applications: when machines operate in the physical world, the standards—and stakes—are much higher. Reliability adds another layer of difficulty, since building a machine that functions dependably for ten years is its own challenge. "Building in the space is a bit of an uphill battle," Sebastian says.

The complexity of the space makes the founder's journey in robotics longer and harder than most. When Sebastian mingles with fellow founders, he likes to ask them the following question: “Given what you know today, how fast could you rebuild your business?” The most common answer: roughly five times faster.

“If you are building a business you have never built before, there are certain things you just can't know. You can only learn through trial and error,” says Sebastian. The math is humbling. “With a five-day workweek, this means that one day of the week you make good decisions, and four days of the week you effectively make bad decisions,” he says. “You hire the wrong person. Or you have the wrong go-to-market, the wrong product, the wrong pricing."

Google Glass is his most personal example of this. Sebastian shares that the team at Google X had no real answer when asked what they were building or what its purpose was. "When people asked, 'What is this?' I said, 'It's a computer on your nose,'” recalls Sebastian. “They'd say, 'What is it for?' and we’d say, ‘We don't know.’"

The lesson is that learning through mistakes is an unavoidable aspect of building a company. While hindsight is 20/20, you can only gain hindsight by moving forward.

2. The hardware-software divide is the robotics industry's most overlooked problem

Sebastian is candid about the limitations of the industry today. "We're not at the point yet where robots are so good it's a no-brainer for people to adopt," he said. "There are still a lot of deficiencies." One reason he cites is a structural problem that starts in universities and compounds throughout entire careers.

“My biggest gripe with the field is that we distinguish the world into hardware and software people,” he says. “You're either a mechanical engineer who loves building mechatronics, or you're a software person. Most of us are unable to bridge this divide."

Teams led by hardware experts take great pride in how a robot looks, feels, and moves, but can struggle with the software and business model that would make it commercially viable. Software-led teams build elegant systems, but often in a vacuum, with a business model that depends on finding hardware partners who don't yet have businesses of their own. Either approach is missing something crucial.

Thrun traces the same pattern all the way back to the 2005 DARPA Grand Challenge. Of 196 competing teams, 195 were hardware teams. They ripped apart existing vehicles, replaced the components, and rebuilt them from scratch. But the challenge was entirely solvable with a standard rental car. The challenging part was to make the car smart, but most teams were too focused on their area of specialty to see that.

“You get blinded by what you're confident in doing,” says Sebastian. “This makes it very hard to build a strong business." The solution to this problem is simple in theory. "If I had a magic wand, I would force every software person to spend one year just building hardware, and every hardware person to spend a year just doing software,” he says.

Sebastian practices what he preaches. “I started doing lots of hardware, even though I'm really bad at it. It’s the only way to get the sense that the truth is somewhere in the middle."

3. Building a wedge isn’t always necessarily easier

A classic question for founders is whether to start narrow—pick a specific use case, solve it well, and expand from there—or pursue a more ambitious vision from the start.

In the early days of building Waymo, Sebastian leaned toward the incremental path. The team was focused on minimizing critical interventions—safety-related moments where a driver had to take over—and found themselves spending more time on highways than city streets. The logical next step seemed obvious: partner with an OEM, build the world's best highway driving system, and gradually work toward full autonomy.

But Larry Page—who Sebastian still considers a mentor today—had other ideas. "Larry said to me, ‘I want to build an end-to-end robotaxi system,’” says Sebastian. There were offers on the table from manufacturers who wanted to license Waymo's technology, but Sebastian—backed by Larry and Sergey Brin—turned them all down. "We wanted to go all the way to the end," he says.

Looking back, he's convinced it was the right call. "I think it was easier to build Waymo than it is to build a pizza restaurant," Sebastian says. He points out that a pizza restaurant is incremental by nature: thin margins, fierce competition, no outside investment. Waymo, by contrast, attracted world-class talent and billions in capital to build something entirely new. "I think people often get this wrong,” he says. “People always think the incremental path is easy. I think the incremental path is hard."

"If you want to change the world, you might as well change it in a big way," he says.

4. Know what actually motivates you beyond financial reward

Sebastian walked away from Waymo in 2014 to found Udacity, a decision he stands by. “I never regret anything,” he says. "What motivates me personally is to make the world a better place for other people. That's amazingly the most gratifying thing you can possibly do—much more than getting rich, which I don't really care that much about."

One of the world’s first MOOCs, Udacity aimed to give everyone around the world access to a calibre of education that was previously only available in Silicon Valley. "In the Middle East, there are millions of people whose lives have been transformed. We materially lifted the GDP of Egypt. I found this very gratifying."

Democratizing access to education energized him more than any financial upside. He credits that clarity of motivation as a source of sustained drive across the long, hard middle of building. The takeaway for founders: get clear on your "why" early, because it's what sustains you through the long, hard middle.

5. Be arrogant about the goal, humble about the path

The skillset required to build a great company—hiring, managing, course-correcting—is largely the same whether you're building software or robots. But Sebastian thinks the defining characteristic of a great founder comes down to a specific kind of balance.

"You have to be incredibly arrogant about the mountain you want to climb,” he says. “But then underneath, you need to be extremely open-minded and willing to adapt and learn. You need to take your ego out of it."

Getting that balance right is harder than it sounds. Sebastian admits he still wrestles with it. "To the present day, I'm struggling with this. How long do you push? When do you pivot, while still keeping the big goal in mind? That's always challenging."

When founders fail to strike the right balance of conviction and humility, the company can begin to fracture from within. "Companies tend to die most often from the inside, not the outside," says Sebastian. "Founding a company is a very raw experience. You feel like it's yours, and if you fail, the world will see it. You have to fire people and disappoint the investors who believed in you. That amplifies everything."

Sebastian’s advice is simple: Try to stay calm amid chaos or change, protect relationships with your teammates, and always maintain a healthy dose of humility.

6. Genuinely befriend the regulators

For founders building in robotics, regulation is an unavoidable reality. Sebastian has watched many treat regulators as obstacles that slow progress down. He understands the frustration, but he has a different view.

"Not bringing regulators on your side is a big mistake,” he says. “They mean well, the same way you mean well. I believe this very deeply." Sebastian points out that regulators and founders largely want the same things. "Regulators want safety, and to be part of the innovation process. It's a stupid business to build something unsafe."

Sebastian acknowledges the differences that can lead to tension. "Your tools are bits and bytes, theirs are words and regulations. Your cadence is different, too,” he says. “But if you treat them with respect, they treat you with respect."

At Kittyhawk, Sebastian brought on Mike Huerta—the former FAA Administrator—as an advisor, specifically to keep regulators close from the start. "You can get regulators to the point where they want you to win and want you to succeed, if they feel they've been taken along," he says.

His advice to founders is to invest in building genuine relationships with regulators: "The regulators can be your best friends."

7. Even the savviest builders can't predict what's coming next

Even though Sebastian is now considered the godfather of self-driving cars, he wasn’t always as confident in their future. He describes himself as a “doubter” in the early days. But this stance softened over time. "When a human first learns how to drive, they learn from their mistakes. But they don't learn from other drivers’ mistakes,” he explains. “When a self-driving car makes a mistake, all the others immediately learn from it, including all the future generations of cars.”

Sebastian figured: “It was just a matter of time until the curves crossed and autonomous vehicles were safer than human drivers. That happened in roughly 2023," he says. “Today, Waymo is roughly 100x safer than human drivers. Next year it could be 500x times safer.”

Similarly, Sebastian holds a healthy amount of skepticism and uncertainty about how innovation will unfold next. “I can't tell you the future,” he says. “It’s basically impossible for me to predict the next six months.”

Sebastian did not anticipate LLM adoption would unfold in the way that it has. "I didn't see it coming, even though they’ve been around for a long time. I didn't predict Claude—it's unbelievably amazing, but it's pretty obvious now,” he says. “Whatever you think is impossible, it becomes possible just a year later.”

"In the next three months, someone in the robotics industry will invent something I won’t see coming,” says Sebastian. “And it will be completely amazing."

Want to learn more about the future of robotics? Dive into our Bessemer Predicts report, covering what investors and founders expect next in the industry.