This article was originally published on Fool.com. All figures quoted in US dollars unless otherwise stated.
Autonomous driving is emerging as one of the most exciting opportunities in the artificial intelligence (AI) landscape. Developing self-driving vehicles stitches together semiconductors, software development, and robotics. Hence, there are several different ways to invest in the technology.
For now, the most mainstream opportunities in the autonomous vehicle market seem to be through Tesla (NASDAQ: TSLA) and Alphabet, which owns self-driving car business Waymo. What most investors likely overlook is that Waymo and Tesla have approached building fleets of self-driving cars through different lenses.
While Waymo has a first-mover advantage in scaling self-driving taxi fleets, some findings from the company's recent academic research paper titled "Scaling Laws of Motion Forecasting and Planning" could suggest that Tesla may have a technological edge.
Let's explore Waymo's progress so far and assess why Tesla could be the superior autonomous driving opportunity in the long run.
Waymo beat Tesla to the punch
According to Alphabet's first-quarter earnings report, Waymo completes more than a quarter of a million paid rides on a weekly basis. Not only is this a fivefold increase compared to last year, but Waymo's serviceable markets are still quite limited. For now, Waymo primarily operates in Austin, Texas, and has planned expansions in Washington, D.C., and Miami over the next year.
By contrast, Tesla just launched its long-anticipated robotaxi service in Austin a couple of weeks ago. Considering Waymo's successful early adoption rates, investors may be wondering how Tesla plans to close the gap against competitive forces.
Did Waymo just quietly admit Tesla has an edge?
There are several differences between how Waymo and Tesla have approached building self-driving car fleets. From a technical standpoint, the two companies have opposing views on variables such as mapping, sensors, and developing the compute power needed to train and hone AI models.
Waymo's approach is grounded in simulating real-world environments and driving behaviors. By contrast, Tesla's general approach has been to use a data-heavy feedback loop from its actual drivers.
Tesla vehicles are constantly collecting loads of sample data from drivers such as speed and braking patterns or interventions. Subsequently, Tesla trains and iterates its models on this large and expanding data set to improve its autonomous driving software platform and budding robotaxi fleet. Tesla collected more than 3.5 billion miles' worth of driver data from its Full Self-Driving (FSD) software platform.
In Waymo's research report, the company suggests that "collecting more data" could be advantageous when building and scaling autonomous vehicle platforms. Waymo goes on to say that "every 10 observed miles are equivalent to 2 to 3 demonstrated miles."
To me, these statements seem to suggest that building a sophisticated model alone is not enough to perfect autonomous vehicle software. Rather, collecting data at scale is a critical part of the equation. Furthermore, it appears that Waymo is saying that using large volumes of real driver behavior (i.e., non-autonomous vehicle platforms) is important for training these models over time.
In essence, Waymo's study seems to endorse many of the pillars supporting Tesla's approach to developing self-driving cars -- learning from actual driver behaviors, collecting billions of various data points, and ultimately iterating and scaling the technology based on these takeaways.
Is Tesla stock a buy right now?
At the moment, Tesla stock appears to be pricing in a lot of upside from the robotaxi launch. I wouldn't chase momentum at these valuation levels, per se. Tesla stock often trades on narratives, which smart investors know can change quickly.
As a long-term investor, I encourage readers to think about the bigger picture here. Self-driving cars are still a new, evolving technology and I do not personally think there is a single correct way to build and scale these platforms.
Candidly, I don't see Waymo's study as some sort of veiled admission that Tesla's approach to autonomous driving is right and theirs is wrong. I simply think Waymo's takeaways subtly imply that Tesla's technological approach has some merit.
What will determine who wins the autonomous driving opportunity will boil down to which company can acquire more customers, expand to new markets, strategically partner with existing ridehailing applications, and scale its fleets more rapidly and profitably. For these reasons, I do not think Tesla necessarily has Waymo is a checkmate position -- at least not yet.
This article was originally published on Fool.com. All figures quoted in US dollars unless otherwise stated.
