Cruise doing things that don't scale

In the race to develop autonomous vehicles, many companies have pursued ambitious visions of fully self-driving cars capable of navigating complex urban environments. But when Kyle Vogt founded Cruise Automation in 2013, he took a decidedly different approach—one that exemplified the startup philosophy of “doing things that don’t scale.” Instead of attempting to solve all the challenges of autonomous driving at once, Vogt and his team focused on a much narrower problem: creating a system that could handle highway driving only. This strategic decision to start small would ultimately lead to one of the most significant acquisitions in the autonomous vehicle industry and position Cruise as a major player in the self-driving car revolution.

The Battlebot Origins

Kyle Vogt’s journey toward autonomous vehicles began in an unlikely place—a Las Vegas arena where his battlebot “Decimator” was being destroyed in competition. As a 13-year-old watching his robot take a beating, Vogt had an epiphany during the long drive home to Kansas City. Looking at the monotonous highway stretching before him, he thought, “This seems like something a robot could do.”

This early insight would stay with Vogt through his education and early career. Before founding Cruise, Vogt had already established himself as a successful entrepreneur in the tech world. He was a co-founder of Twitch (originally Justin.tv), which would later be acquired by Amazon for nearly $1 billion, and Socialcam, which was sold to Autodesk for $60 million. Despite his success in video streaming technology, Vogt’s interest in automation and robotics remained strong.

The Highway-First Approach

When Vogt founded Cruise Automation in 2013, he didn’t set out to build a fully autonomous vehicle that could navigate the complex streets of San Francisco. Instead, he focused on a much more manageable problem: highway driving. This approach made perfect sense from both a technical and business perspective.

Highways present a relatively controlled environment for autonomous systems. They have clearly marked lanes, predictable traffic patterns, and fewer variables like pedestrians, cyclists, or complex intersections. By focusing exclusively on highway driving, Cruise could develop a working product much faster than competitors attempting to solve the full spectrum of autonomous driving challenges.

“Our focus is to get a simple version of this product in people’s hands quickly and slowly build on that,” Vogt explained in a 2014 interview with The Verge. This philosophy of starting small and iterating quickly is a classic startup strategy, but it was relatively novel in the autonomous vehicle space, where many competitors were taking a more comprehensive approach.

The RP-1: A $10,000 Retrofit Kit

Cruise’s first product was the RP-1 (Retrofit Pilot), a $10,000 kit that could be installed on specific Audi models—the 2012 or newer A4 or S4. The system consisted of sensors mounted on the roof of the car, actuators for the steering wheel, brakes, and accelerator, and a computer system installed in the trunk. A special button in the center console allowed drivers to switch between human and computer control.

Unlike more ambitious autonomous systems, the RP-1 was designed specifically for highway use in California. It could maintain the vehicle’s position within a lane, adjust speed based on surrounding traffic, and avoid obstacles—but only on highways that had been previously mapped by Cruise.

The system had significant limitations. It wouldn’t work in rain, fog, or darkness. It required pre-mapping of routes before they could be driven autonomously. And at $10,000—nearly 29 percent of the cost of an Audi A4 itself—it was expensive for what was essentially an advanced cruise control system. But these limitations were by design. Rather than trying to solve every problem at once, Cruise was focusing on delivering a working product that could handle a specific use case well.

“This is a little different than what you’ve seen from Google, for example, because that’s like a $100,000 research car, with tons of fancy sensors,” Vogt explained during a demonstration. “Everything on this car is priced for a consumer product. It’s, like, very low-cost everything, and the performance is still very good.”

The Beta Testing Approach

When Cruise announced the RP-1 in June 2014, it limited pre-orders to just 50 units, with deliveries expected to begin in 2015. This limited release allowed the company to treat its early customers as beta testers, gathering valuable data and feedback while continuing to refine the technology.

The company’s approach to mapping highways was similarly incremental. “We have to drive the route once, and sort of map it, and then it can do the route on its own after that,” Vogt explained. The company planned to gradually build a network of mapped highways, starting with major freeways in the Bay Area and expanding over time.

This data-gathering strategy was crucial for Cruise’s long-term vision. Each vehicle would contribute to a growing database of road information, creating a network effect that would make the system more valuable as more users adopted it. This approach mirrored strategies used by other successful tech companies, where user data becomes a competitive advantage over time.

The Pivot to Full Autonomy

Less than a year after announcing the RP-1, Cruise made a significant pivot. The company went silent on its retrofit kit and shifted its focus to developing fully autonomous vehicles. Vogt had decided that highway autonomy wasn’t enough—he wanted to challenge Google and other competitors in building cars that never need humans.

This pivot might seem like an abandonment of the “do things that don’t scale” philosophy, but it was actually an evolution of it. The highway-only approach had allowed Cruise to develop core technologies, gather real-world data, and build a team with expertise in autonomous systems. Now, they could leverage that foundation to tackle the more complex challenge of urban driving.

The company began adding LiDAR sensors to its vehicles in 2015, enabling them to create more detailed 3D maps of their surroundings—a crucial capability for navigating urban environments. They also joined Y Combinator, the prestigious startup accelerator, which provided additional resources and connections to help fuel their growth.

The GM Acquisition

Cruise’s strategic approach paid off in March 2016, when General Motors acquired the company for approximately $1 billion (reports vary between $600 million and $1 billion). At the time, Cruise had fewer than 40 employees, making this one of the most valuable per-employee acquisitions in the tech industry.

For GM, the acquisition was a way to accelerate its autonomous vehicle program. The traditional automaker recognized that it needed to partner with tech companies to compete in the rapidly evolving transportation landscape. By acquiring Cruise, GM gained access not just to technology but to a team with a proven ability to innovate quickly in the autonomous vehicle space.

For Cruise, the acquisition provided the resources and manufacturing expertise needed to scale their technology. As Vogt explained after the acquisition, “What motivates us is the chance to achieve massive social impact, and that takes lots and lots of vehicles.”

Lessons from Cruise’s Highway-First Approach

Cruise’s journey from a highway-only retrofit kit to a billion-dollar acquisition offers several valuable lessons for entrepreneurs:

1. Start with a manageable problem: By focusing initially on highway driving, Cruise was able to develop a working product much faster than if they had attempted to solve all autonomous driving challenges at once.
2. Build a foundation for future growth: The technologies and expertise developed for highway driving provided a foundation for tackling more complex urban environments later.
3. Use constraints as an advantage: By limiting their system to specific vehicles and highways, Cruise could control variables and deliver a more reliable product within those constraints.
4. Gather data strategically: Cruise’s approach to mapping highways and gathering data from users created a valuable asset that would become increasingly difficult for competitors to replicate.
5. Be willing to pivot: When Vogt recognized that the bigger opportunity was in full autonomy, he was willing to shift the company’s focus, building on what they had learned rather than sticking rigidly to their initial plan.

The Legacy of Cruise’s Approach

Today, Cruise is one of the leading companies in the autonomous vehicle industry. After the GM acquisition, the company expanded rapidly, growing to over 1,600 engineers working on urban autonomous driving. They’ve deployed fleets of self-driving Chevrolet Bolts in San Francisco and other cities, gradually expanding the capabilities and reach of their technology.

While the company has faced challenges and setbacks—including a significant incident in 2023 that led to a temporary suspension of operations and the resignation of Vogt as CEO—its impact on the industry is undeniable. The approach of starting with a narrowly defined problem, developing expertise in that area, and then expanding to more complex challenges has influenced how many companies approach autonomous vehicle development.

The $10,000 highway autopilot kit that Cruise announced in 2014 never made it to market in its original form. But it served its purpose as a stepping stone, allowing a small startup to establish itself in an industry dominated by tech giants and automotive behemoths. By doing things that didn’t scale—focusing on a specific use case, limiting their initial product to certain vehicles and highways, and treating early customers as beta testers—Cruise built the foundation for what would become a billion-dollar company at the forefront of the autonomous vehicle revolution.

In the end, Cruise’s story demonstrates that sometimes the path to solving big problems starts with tackling smaller ones well. As Kyle Vogt reflected in a 2016 interview, “Everything’s a matter of time and work.” By breaking down the enormous challenge of autonomous driving into manageable pieces, Cruise found a way to make meaningful progress in an industry where the ultimate goal—fully autonomous vehicles—remains one of the most difficult technical challenges of our time.