The recent decision by General Motors to shut down its Cruise robotaxi division marks a major turning point in the autonomous vehicle (AV) landscape. How could GM invest over $10 billion, only to walk away? The reality is that while the technology may look “almost there,” the final 1% of edge cases is where the real challenges—and most of the costs—lie. Getting AVs to profitability requires far more investment than most people realize.
Take Waymo as an example. They currently operate about 700 vehicles with a team of over 3,000 people. That’s more than four employees per car—and these aren’t gig workers; they’re highly paid specialists ensuring smooth operations. Your $19 Waymo ride may feel futuristic, but it’s heavily subsidized and will remain so for the foreseeable future.
Meanwhile, Tesla, long bullish on its “Full Self-Driving” (FSD) system, is now effectively admitting that true autonomy is still out of reach. Tesla’s upcoming “cybertaxi” fleet will reportedly include teleoperation—remote human drivers stepping in when the system struggles. Tesla claims their FSD system only requires support every 13 miles, but in dense urban settings, drivers report interventions as frequently as every 1-2 miles. For context, Waymo’s last reported disengagement rate in 2023 was once every 17,311 miles. Even in 2015, Waymo achieved 1 disengagement per 1,250 miles—meaning Tesla could be over a decade behind.
Scaling Autonomy is Hard—and Expensive
The truth is that autonomous vehicle operations remain extraordinarily complex and costly. AV systems frequently require human intervention, though data on these interventions remain scarce. Tesla will likely lean heavily on teleoperation to support its cybertaxi rollout, starting in cities like San Francisco and Austin, where regulations are favorable and testing is already underway. Elon Musk’s growing influence in the Trump administration may also play a role in pushing for federal regulatory flexibility to speed up deployment.
Teleoperation: The Bridge to Full Autonomy
In this challenging environment, teleoperation is emerging as the practical bridge between today’s technology and the vision of full autonomy—much like hybrid cars paved the way for electric vehicles. Teleoperation enables remote human intervention when AV systems face challenges, providing a safety net and allowing gradual progress toward full autonomy.
Industry Trends: A Pragmatic Pivot
The rise of teleoperation startups over the past five years reflects this shift toward pragmatism. Numerous companies have now been founded by former engineers spun out from leading AV players like Uber ATG, Zoox, and NuTonomy. These innovators are building cutting-edge teleoperation platforms that showcase the potential of remote human intervention as a scalable, cost-effective solution.
Beyond Passenger Vehicles
Teleoperation isn’t limited to passenger cars. It’s already being deployed successfully in sectors like:
- Delivery robots: Navigating last-mile logistics challenges.
- Industrial operations: Enhancing efficiency in forestry, agriculture, and mining.
These applications are proving that teleoperation can outperform fully autonomous systems in real-world scenarios where flexibility and human oversight are essential.
Looking Ahead: A Realistic Path Forward
The AV hype cycle resembles what we saw with electric vehicles. The industry rode the wave of hype only to fall into the “trough of disillusionment,” where hybrid cars became the practical solution. Now, AVs are heading down a similar path—with teleoperation as the stepping stone to a sustainable future. Hopefully we can eventually get to autonomous, electric cars but for now the immediately viable path is teleoperated hybrid cars.
At Upshift, we see teleoperation as the next logical step toward autonomy. By focusing on lower-risk scenarios like vehicle delivery and valet parking—where no passengers are in the car—we’re able to build safe, scalable operations with safer, lower cost, and more expeditious operational design domains.
The journey to full autonomy is long and complex. But by embracing teleoperation and leveraging advanced driver assistance systems, we can deliver meaningful advancements in mobility today. The future isn’t just about replacing drivers; it’s about augmenting human capabilities with smart, adaptive technologies.
At Upshift, we’re excited to help shape this future—one step at a time.
