🔒 Elon Musk’s Tesla robotaxi needs to come with a time machine: Liam Denning

08 Oct 2024, 10:17 by Editor BizNews · BizNews

In advance of Tesla’s highly anticipated “robotaxi unveil,” Elon Musk faces pressure to maintain control over the narrative he crafted five years ago. Initially promising self-driving robotaxis by 2020, Musk must now address significant delays amid declining EV sales and a fluctuating market cap. As competitors like Waymo make steady progress, Musk’s upcoming presentation on October 10 aims to rejuvenate interest in Tesla’s autonomy claims, potentially relying on spectacle over transparency.

Sign up for your early morning brew of the BizNews Insider to keep you up to speed with the content that matters. The newsletter will land in your inbox at 5:30am weekdays. Register here.

By Liam Denning

When Elon Musk hosted Tesla Inc.’s “autonomy day” in March 2019, he was crafting a narrative. When he takes the stage at Thursday’s “robotaxi unveil,” he will fight to keep control of that narrative. ___STEADY_PAYWALL___

Five years ago, Musk touted robotaxis — self-driving, self-funding wonders earning money for their owners — that would be ready by the end of 2020. Even allowing for the pandemic, they are well past due. In the meantime, Tesla’s sales of electric vehicles surged fivefold but then stalled. Its market cap ballooned from less than $50 billion to a peak of $1.2 trillion before giving back about $400 billion of that.

As the EV growth story faded, artificial intelligence has filled the gap. Bullish analysts, and Musk himself, routinely ascribe a fraction of Tesla’s value to selling EVs, with the majority now tied to visions of robotaxis and robots instead. Musk touts Oct. 10 as a day “for the history books” under the tagline “we, robot.”

It’s worth remembering that the robotaxi unveil was precipitated by, what else, a spat on X. About a year after Tesla hosted an investor day in March 2023 to present Musk’s “Master Plan 3,” which focused heavily on creating a low-cost EV, Reuters reported that project had been canned. Tesla’s stock, weakened already by the company having abandoned its EV growth target, was hit hard. Musk accused Reuters of lying but then effectively confirmed the story by suddenly announcing a “robotaxi reveal” for Aug. 8. He later teased a new master plan.

The sound of glass breaking was unmistakable, especially as the reveal then got pushed back by two months. Assuming Musk’s new blueprint gets debuted this week, the lifespan of his master plans will have been shortened mightily.

Expect Tesla’s Optimus humanoid robot to feature prominently as a cameo from the future on Thursday. Yet no one should expect Tesla to announce that it has solved the problem it set out five years ago: Namely, creating a robotaxi that drives under general AI to a level where Tesla assumes liability; relies on cameras and radar rather than expensive, laser-targeting LiDAR technology; and can be switched on across millions of existing Teslas with an over-the-air software update to drive anywhere in the US.

Even Adam Jonas, the very bullish Morgan Stanley analyst, isn’t expecting that. Rather, according to a recent report, he expects a “cybercab” able to drive itself around a closed course on the Los Angeles studio lot Tesla has booked. He also foresees Tesla revealing a “‘supervised’ autonomous/FSD rideshare service,” whereby Tesla owners with the Full Self Driving driver-assistance package offer rideshare services; like catching an Uber but with the driver letting the car control itself at times. Tesla would gain more data for its autonomy machine learning, no doubt, but it’s hard to see why passengers would choose this — unless, of course, Tesla heavily discounts those rides to take market share (not unlike its efforts to shore up EV sales at the expense of profits).

All of this would represent a climbdown for the company, but also a means of repositioning in an autonomous vehicle sector that has seen some notable shifts of late; especially concerning the leading operator of actual, existing robotaxis, Alphabet Inc.’s Waymo LLC.

Waymo is, in a sense, the tortoise to Tesla’s hare, building its capabilities incrementally with small numbers of robotaxis deployed in limited settings and backed by remote teleoperators able to intervene. That methodical approach is safer but also inherently slower. Alex Roy, an expert on vehicle autonomy and founder of New Industry Management LLC, a venture fund, is something of a Waymo evangelist. (“I don’t have to choose a setting for conversation and there’s no way it will cancel.”) He notes, however, that if Tesla does eventually crack generalized autonomy with its lighter hardware approach then “Waymo is in mortal danger.”

In the meantime, however, Waymo is free to scale at its own pace — and is doing so cleverly. It just announced it will expand to Austin (Tesla’s backyard) and Atlanta exclusively via Uber Technologies Inc.’s app, building on an existing partnership in Phoenix. Waymo gains access to a giant, established customer network managed on its behalf and can size its fleet for optimal utilization, since Uber’s variable pool of human drivers can handle surges, which also reduces Waymo’s unit costs. Uber, meanwhile, gets a new, differentiated source of vehicles to meet regular demand and can establish itself as a credible partner for autonomy after its in-house effort went awry.

Cracking generalized autonomy, à la Tesla, remains a formidable and time-and-money-consuming task. Philip Koopman, an associate professor and automated driving specialist at Carnegie Mellon University, believes the industry underestimates the sheer number of potential surprises out there on the roads that machines must learn to deal with, yet only rarely encounter, in order to be truly safe. “It’s not how often a thing happens that’s rare; it’s how many rare things there are to happen to you,” he sums up.

What Tesla needs, therefore, is even more time. Roy recently penned a fascinating post on Tesla’s “narrative command.” This describes how the company defines how society thinks, and speaks, about EVs and autonomy by making tangible advancements in things like batteries and vehicle design while also disseminating terms like “Autopilot,” “robotaxi” and “supercharger” into the vernacular. This reality-distortion field has been maintained despite numerous missed timelines and has, above all, kept Tesla’s stock priced for breakneck growth even as its core business has stagnated.

The festivities in Los Angeles are all about bolstering that command in the face of undeniable progress by competitors. One useful way to do this might involve, say, Tesla releasing the data behind its claims that its driver assistance features already outdo humans significantly on safety. Another would be to reveal how many Tesla buyers are actually paying extra for FSD.

It is odd that a company which proclaims safe vehicle autonomy to be its overriding investment case won’t enhance that with a few simple numbers. Somehow, I suspect Thursday will see Musk lean more heavily into spectacle rather than disclosure. As so often, the question will be whether investors are dazzled enough to forget all that was pledged previously.

Read also:

Cyril Ramaphosa: The Audio Biography

Listen to the story of Cyril Ramaphosa's rise to presidential power, narrated by our very own Alec Hogg.

Get the Audiobook
Narration by Alec Hogg