Tesla Inc. (NASDAQ:TSLA) held its Autonomy Investor Day on April 22, 2019, announcing a Full Self-Driving computer and introducing a robotaxi product called Tesla Network. Below are the comments from analysts on their takeaway from the event.
Analysts Comment On Tesla’s Autonomy Investor Day
At its Autonomy Investor Day, CEO Musk targeted the deployment of “feature complete self-driving” capabilities (defined in Q&A as Level 5) this year & the launch of its robotaxi in 2020. Pete Bannon (started 2016), Director & Chip Architect said the in-house designed TSLA chip has more power (72W vs. 57W prior), uses 80% less silicon & has the ability to handle a frame rate of 2,300/second (21x more than prior gen NVDA chip). NVDA responded with a statement noting that TSLA was inaccurate in its comparison. Andrej Karpathy (2017), Director of AI & Stuart Bowers (2018), VP Engineering discussed the AI & software and highlighted the value of TSLA’s data. The event ended with test drives. Our ~20 minute test-drive went smoothly but still had 2 driver interventions. The ride was also not a complex urban environment or during increment weather. Based our AV experts interviews (& short test drive), we believe TSLA’s AV timeline is too aggressive, & therefore near term challenges are more pressing.
We attended Tesla’s autonomy investor day in Palo Alto on Monday. Management offered a few highly detailed presentations from executives responsible for the various components of autonomous driving development: Hardware, Vision (Neural Net), and Software. Additionally, CEO Elon Musk finished the presentation by formally introducing a robotaxi product – Tesla Network – which the company claims is set to go live some time in 2020, pending regulatory approval. Tesla finished the event by offering demonstration drives in vehicles updated with Tesla’s latest hardware and Autopilot software beta.
Starting in 2020, we believe Level 2 Advanced vehicles, not to be confused with self-driving or L4/L5, will reach the mass-market and offer similar capabilities as Tesla, such as AutoPilot, lane merge, and highway exit. According to ABI (per NVIDIA's 2019 Analyst Day), the number of L2A vehicles could reach 35 million by 2025. Major semiconductor companies, namely NVIDIA and Intel/Mobile Eye, have a broad platform of hardware, software, training/development of neural networks, and validation (simulation, data analytics) that will be made available to the entire automotive market. Moreover, an ADAS / autonomous ecosystem has been developing over the last several years across the Tier-1 auto suppliers, software and chip companies. Lastly, we are skeptical of the robotaxi initiative as we believe the company will face significant regulation hurdles.
Timing of the autonomous event two days ahead of the 1Q result is not insignificant, in our view. We believe the autonomy investor day was meant to help the stock market focus on the multiple (strategic value) rather than the earnings (fundamental value). The tone of the event seemed as much geared to a venture capital audience as it was to a public market equity audience, in our opinion. Many other firms in the private and even public AV space have been receiving some modicum of credit from investors and strategic partners for their AV capabilities. Tesla appears to be making a move to get its share of value here.