Giving up on investing in self driving technologies is a huge mistake
The majority of investors have given up on investing additional capital to fund self-driving technologies. Ford/VW shut down Argo in November. Aurora’s market cap is $1.4B, despite having $1.2B of cash on the balance sheet. This is a huge mistake.
Self driving is not only about developing robot cars, but also about funding advanced technologies that will transform many industries in other ways. Here are 4 examples of successful examples:
Scale AI (~$7.3B val) | Started building data annotation tools for self driving. Rapidly expanded to serving robotics, AR/VR, and content/language use cases, becoming the backbone to accelerate the development of all AI applications.
Applied Intuition (~$3.6B val) | Started building simulation tech for self driving cars. Rapidly expanded to serving ADAS, trucking, defense, construction and mining use cases, accelerating the world’s adoption of safe and intelligent machines.
Deepmap (acquired by NVIDIA) | Started building HD maps for self driving cars. Expanded to create HD maps of 500k of highways globally, creating the infrastructure for automakers building the future of transportation.
rideOS (acquired by Gopuff) | Started building dispatch and routing technology for self driving fleets. Made a hard pivot during the pandemic to serving on-demand delivery, and has been a crucial piece of transforming Gopuff’s unit economics.
Self driving investments started in 2007 when it was a technology searching for a problem to solve. The start-ups I met as we were starting Uber’s self driving division were struggling to raise $2M rounds. When Amazon and Uber experienced their hyper-growth phases and driver costs escalated, the business case became clear. Start-ups raised billions of dollars. This was too frothy.
We need investors to balance longer term thinking along with the pressure to maximize short term profits. Humans tend to process technology advancements linearly, when they are actually developing exponentially. There are many ways to win!
We also need the self-driving companies to take a hard look at the valuable technologies built, and urgently assess if there are other industries or use cases that could benefit. Investors are not wrong to demand a path to sustainability. There are other ways to change the world in addition to supply cars or trucks to serve the ride-hail and trucking market.
I could not be more excited to see how this industry continues to develop as we look towards the next decade. :)
Hey Justin,
Found your article to be pretty interesting. Currently working on a project similar to Applied Intuition - creating synthetic datasets to train AI models for self-driving cars. I've got some questions about the space:
- How much does creating synthetic data cost?
- From your experience at Uber, how helpful is simulation tech? Does it really improve an AI model's accuracy or is real world training data better?