Self-Driving Car — The Mid. Game
Mar. 2022 is a busy month of Self-Driving capital games. A few $B are changing hands, and the mid. game of Self-Driving consumer adoption unfolds. Which is hitting what the beginning of the end can be.
Self-Driving capital reload
This spring seems to be a great time for cashing out from Self-Driving. A new chapter of Self-Driving is unfolding because a venture capital has a better appetite on risk than a “bank loan officer”.
- On Mar. 7, Intel filed an IPO for Mobileye. Intel acquired Mobileye for $15B in 2017 and now may exit with a $50B valuation. [Reuters]
- On Mar. 18, SoftBank cashed out $2.1B from its Cruise $900M investment in 2018. Cruise is fully under GM’s will.[Robotics247]
- On Mar. 17, TuSimple is looking to sell its China business for $1B to reduce regulatory risk & focus on the U.S. market. Two top regions of SD are on their separated ways. [Reuters]
So what? tl;dr each type of capital has its expectation on return of capital. Which will guide a venture toward certain choices. So, founders choose wisely on who’s money to take 😉.
Waterfall — the same old love
To be fair, carmakers have been making good progress on self-driving. Arguably, some of them even “beat” Tesla by getting their Level 3 self-driving approved. Whereas, Tesla FSD is still “Level 2”.
- On Mar. 5, 2021, Honda announced the leasing plan in Japan of Honda Legend EX with its SENSING Elite, a Level 3 self-driving system. [Reuters]
- In the first half of 2022, Mercedes-Benz customers may buy an S-Class with DRIVE PILOT, a Level 3 system/conditionally automated mode on suitable motorways in Germany in heavy traffic at a speed under 60 km/h (37 ml/h). [Mercedes-Benz]
- BMW plans to launch a Level 3 in the second half of 2022 powered by Mobileye. In the meanwhile, BMW, Qualcomm & Arriver are working on a new system. Which should be ready in 2025. [Forbes]
They may be good achievements. But, there is a fundamental problem: carmakers still treat Self-Driving as a fixed feature to sell cars. Just like the cruise control feature, carmakers may take ~2 years to develop & test a new model & features. Before the launch of the new model, the development is properly done & validated. This is fine for a well-defined feature & working in predictable environments. Only, self-driving is much more complicated, and the real world is very dynamic.
Iterate fast at scale — the only path to success
From late 2020, Tesla has been building FSD capability & adaptivity by frequent SW updates. FSD beta users grows from ~2k in Mar. 2021 to 60k in Q4 2021. Tesla Safety Score is a criterion to qualify beta users, which reduces the risk & encourages good driving behaviors. Will Tesla FSD be safer than human drivers in 2022? The answer may still be debatable later. But, Tesla is pretty ahead of others a few years. So, if Apple Cars will hit the road in 2025, Apple needs to put a new dream team together soon. Only, Tesla will not standstill.
The opinions stated here are my own, not those of my company. They are mostly extrapolations from public information. I don’t have insider knowledge of those companies, nor a whatever expert.