Tesla (TSLA 3.29%) has gained a reputation for smashing records, shattering expectations, and making constant technological breakthroughs. The company is on the verge of cementing that reputation with its biggest accomplishment to date.
On Aug. 25, CEO Elon Musk posted a 45-minute video of himself in the driver's seat of a moving Tesla Model 3, except he wasn't driving. To showcase Tesla's new version 12 (widely abbreviated as "v12") of its full self-driving (FSD) software, Musk showed off the vehicle parallel-parking, starting and stopping at traffic lights, and navigating roundabouts, all while his hands held his phone to record the video.
Image source: Getty Images.
Tesla is one step closer to full autonomy in its vehicles
Although v12 still has limitations and needs considerable refinement, the software represents a monumental leap in Tesla's pursuit of obtaining Level 4 or Level 5 autonomy -- levels in which drivers no longer need to monitor the car while it's underway.
What distinguishes v12 from previous versions of FSD is its sole reliance on neural networks powered by artificial intelligence (AI). Past FSD software used neural networks to map vehicle surroundings, but the course of action the vehicle would take still required hard coding by Tesla programmers. With v12, humans are no longer needed to define specific responses.
Thanks to advances in AI development, v12 can interpret and discern the best response for a vehicle on its own. By analyzing a metric ton of video and driver behavior collected from its 4 million vehicles around the world, Tesla's supercomputer Dojo and a host of Nvidia graphics processing units (GPUs) were able to train neural networks to learn the proper decisions to make in any situation.
Why v12 is a big deal for Tesla
Developing responses for self-driving cars was a difficult task for Tesla's programmers. The unpredictable nature of driving meant they had to account for a wide range of variations, limiting the capabilities of previous versions of Tesla's FSD.
However, advances in v12 and self-training neural networks will make human programmers unnecessary, increasing autonomy and the speed at which the neural networks learn. Now, Tesla's biggest challenge in achieving Levels 4 and 5 of autonomous driving will lie in its ability to collect and process video data. Fortunately, this task should be much easier than the initial challenge of developing and training neural networks, since Tesla already has a large fleet of vehicles collecting video and has its own computing infrastructure to handle the heavy lifting.
Taking Tesla to new heights
While Level 4 or 5 autonomy will make roads safer and is a noble pursuit in itself, development of more advanced FSD will help Tesla reach its ultimate goal of creating a robotaxi business. Developing such a service is currently Tesla's No. 1 priority, mainly because Elon Musk said he believes that there is "quasi-infinite demand" for it, and because it could transform Tesla's sources of income.
Although Musk is known for exaggerating timelines and flaunting his confidence that Tesla shares could 10x from today's prices, he isn't alone in recognizing that a robotaxi fleet could revolutionize Tesla's finances, and potentially catapult it to being one of the most valuable companies in the world.
In an attempt to quantify just how impactful a robotaxi business would be for Tesla, the investing firm Ark Invest (which has some of Tesla's biggest fans) released a report in April 2023 highlighting findings from a Monte Carlo simulation.
The results were grouped into a bearish scenario, an average scenario, and a bullish scenario. Based on the outcomes, Ark Invest hypothesized that successful implementation of robotaxis could generate revenue anywhere from $200 billion in the bearish scenario to nearly $613 billion in the bullish scenario. Furthermore, it estimated that robotaxis could contribute more than two-thirds of the company's future earnings before interest, taxes, depreciation, and amortization (EBITDA).
Like any simulation or model, Ark Invest's analysis has its limitations. Still, even in its bear case, in which robotaxis produce just $200 billion in annual revenue, that would be a jump of 700% from Tesla's current revenue of just under $25 billion.
Moving forward
Although simulations like the one Ark conducted remain speculative, they shed some light on just how transformative the achievement of autonomous driving and the development of a robotaxi business could be for Tesla. With its shares down more than 30% from their all-time high, now could be an opportunistic moment for investors.
Macroeconomic factors such as inflation and rising interest rates have cut into Tesla's profit margins and are the primary reason share prices took a hit. But the company is confident in its ability to weather short-term market fluctuations, thanks to its robust financial standing and continued record-breaking production and revenue. With the likelihood of a highly lucrative robotaxi business and FSD on the horizon, Tesla could lead AI, electric vehicles (EVs), and portfolios for years to come.
