Lyft has always been the second-fiddle in the ride-hailing world. Its presence has always been shadowed by that of Uber—and that’s despite it technically being first to market in the 2000s under the Zimride branding. And as the old saying goes: if you’re not first, you’re last.
That’s why Lyft has decided to take a page out of Uber’s book. The ride-hailing app wants to secure its place in the future by launching its own autonomous robotaxi offering into the increasingly populated AV market. That’s been heating up for a while, but now autonomous test rides are coming to Atlanta.
Welcome back to Critical Materials, your daily roundup for all things electric and tech in the automotive space. Also on deck: China is cracking down on EV trolls and false advertising, and Tesla reveals how many owners are opting for Full Self-Driving. Let’s jump in.
30%: Lyft Readies Up In Robotaxi War
Photo by: Lyft
Autonomous ride-hailing isn’t new. Companies have been throwing money at the tech for years—some have proven to be fairly successful thus far, while others have thrown in the towel early on. Then there are the newcomers that seek to use variations of already-commercialized tech and a massive fleet of cars to collect training data to get ahead quickly.
Lyft just happens to be the latest company hopping onto the robotaxi train before its business model becomes trampled by robots on wheels. Its bid to take on the likes of Waymo (which has partnered with Uber in Austin, Texas), Tesla, Amazon’s Zoox and the like will soon dispatch a fleet of Lyft-branded Toyota Sienna hybrids to the streets of Atlanta to kick things off.
Lyft has been adjacent to the AV game for a while. It tried this before with Hyundai’s Motional venture and Ford-backed Argo AI; both of which didn’t exactly pan out. Its newest effort relies on a lesser-known company called May Mobility, which has an interesting take on how its vehicles make decisions on the road.
The company calls itself an Autonomy-as-a-Service (AaaS) platform, meaning that it enables other providers to launch AVs using the tech developed and maintained by May that can quickly integrate and scale based on a partner’s needs. May Mobility says that its “Multi-Policy Decision Making” rationale is what makes its software the ideal choice for an AV provider focused on safety at its core.
It justifies this by explaining the Minimum Risk Maneuver decision-making process that simulates multiple “what-if” scenarios—will that pedestrian cross the street? Will that car merge into my lane?
Photo by: Lyft
It will then instruct the host vehicle to take the safest possible course of action based on the variables in the world around it and how it predicts how things could play out in the environment.
TechCrunch sheds a bit more light on the partnership:
Lyft’s deployment with May Mobility comes a month after the company announced a deal with China’s Baidu to launch robotaxis in Europe next year. Lyft CEO David Risher has also said the company would work with Mobileye to deploy Mobileye-powered vehicles on the Lyft app in Dallas “as soon as 2026,” with “thousands more AVs/other cities to follow.”
Not all of Lyft’s AV partnerships have panned out. The company previously launched a robotaxi service—always with a human safety driver behind the wheel—in Las Vegas via a partnership with Motional. It had a similar agreement in Austin and Miami with Argo AI.
Interestingly, May Mobility is also one of Uber’s partners. The duo is set to launch robotaxis in Arlington, Texas later this year.
The bigger question is whether or not Lyft has the resources to compete with all of the other providers currently preparing for battle in what could shape up to be a crowded industry. Lyft has struggled with inconsistent profitability and operational challenges for its human drivers over the last few years. It’s not clear how long Lyft can afford to shred capital to fund its robotaxi experiment.
But on the flip side, if Lyft doesn’t get ahead of competitors doing the same thing, it could easily sink in the long run.
60%: China Takes Aim At False Advertising and Internet Trolls Causing EV ‘Chaos’
Photo by: BYD
On Wednesday, China drew an abundantly clear line in the digital sand that’s intended to squash automakers with deceptive marketing practices and internet trolls alike.
The country’s Ministry of Industry and Information Technology has begun a three-month “special rectification” crackdown that takes aim at automakers and online actors who are spreading false advertising, fabricated content, as well as the organization and manipulation of online content that misrepresents cars built by domestic brands. In short, they’ve had enough with the lies in an uber-competitive industry and are committed to taking action.
Here’s the scoop from Bloomberg:
Chinese regulators have launched a three-month campaign to crack down on advertising practices carried out by automakers and media companies that they describe as illegal and which may have contributed to “online chaos.”
The offenses include carmakers exaggerating the specifications and quality of their vehicles and batteries, according to a document released Sept. 10 by six agencies including the Ministry of Industry and Information Technology and the Cyberspace Affairs Commission.
Some media companies have also used reviews and model testing to blackmail manufacturers for favorable feedback, it said.
China’s EV industry is in a bit of a panic right now. Margins have been gutted and the numbers game that many brands have been accused of playing has finally reached critical mass. Regulators have watched this take place in real-time and realize that many of the tactics that brands have been relying on to advertise is now undermining both safety and quality.
MIIT and other agencies involved in the special crackdown will require brands (both automakers and online platforms) to “rectify irregularities” related to misleading information and advertising. This even includes media companies that stand accused of blackmailing manufacturers.
China hasn’t said what the punishment will be. But a statement from the China Advertising Association warned in a statement that the maximum penalty for the offenses is up to two years of imprisonment, plus fines. Yeesh.
90%: Tesla Says Half Of Model S and X Owners Are Buying FSD
Photo by: Tesla
Tesla’s Full Self-Driving has long been the automaker’s crown jewel up-sell. An $8,000 software option that promises your car will one day be able to chauffeur you around (maybe). But how many people are actually paying for it?
That number has long been a big fat question mark. But in an interview with Jay Leno, Tesla’s VP of Vehicle Engineering, Lars Moravy, let an internal estimate of how many people are actually buying FSD into the open. It turns out that the automaker’s luxury segment is pulling a lot of weight with more than half of all buyers opting for FSD, according to Moravy.
“The take rate has gone up from single-digit percentages on all our cars. We’re in the teens,” said Moravy. “On [Model] S and X, it’s over 50% [to] 60%.”
Historically, Tesla has never really been open about its take-rate for FSD. One survey estimated the overall percentage of folks who purchased FSD after trialing it was around 2%. Musk said that the rate was actually “much higher,” but didn’t specify what that number actually is.
Over the last year, Tesla sold 69,830 units of the Model S, Model X, and Cybertruck. This means the estimated take-rate of FSD in Tesla’s luxury models would be somewhere in the neighborhood of 35,000 and 42,000 vehicles. Now, we know that Tesla’s entire fleet is “in the teens,” according to Moravy.
Tesla also sold 1,679,263 Model 3 and Model Y in the same time period. This would estimate that between 185,000 and 275,000 Model 3 and Y buyers either purchased or subscribed to FSD, or between 11% and 17%.
Let’s talk cash. If all of Tesla’s new buyers were subscribing to FSD, that’s anywhere between $21.8 million and $31.4 million in annual recurring revenue. More than likely, the majority of Model S and X buyers are working the cost of FSD into their vehicle purchase—hence the introduction of the new “Luxe Package,” which could account for as much as $336 million in additional one-time revenue for Model S and X buyers. Realistically, it was likely much lower than that since before the Luxe Package bundled FSD in as a mandatory add-on.
The problem for automakers is that vehicle autonomy is currently a cash-burn. General Motors knew this with Cruise, and Waymo has been sinking money into it for longer than most. But Tesla has been living on the promise of “eventually” for a while now, offering owners just a taste of partial autonomy to hold them off. It has also been able to fund its autonomy ventures by convincing customers to pay for an incomplete feature that, in some cases, is not expected to actually be fully autonomous using hardware currently installed in the vehicle.
It’s also only natural that Tesla’s lower-priced models have a lower adoption rate—when the software feature is nearly 20% of the vehicle’s cost, it might be hard to justify buying outright.
Tesla does still pitch FSD as the future for both driver safety and company profit. Elon Musk has long said that the real value of Tesla is in its ability to deliver on autonomy, but that vision has slipped down the road for nearly a decade. If Tesla can deliver on that promise, it can only hope that the take-rate will increase, too.
At the same time, hands-free driving systems are catching on quickly as a must-have feature. GM and Ford’s respective Super Cruise and BlueCruise are seeing higher and higher take rates, use during road trips and overall demand among consumers. If Tesla can’t stay competitive there, it doesn’t bode well for the autonomous future Elon Musk wants.
100%: Are Robotaxis Headed For A Bubble?
Photo by: InsideEVs
AI-this, AI-that. It’s all we hear about lately. We’ve been warned that the AI bubble could burst at any time, but that hasn’t stopped companies from implementing it wherever the buzzword fits.
When it comes to driving, however, it’s not quite the same as throwing written slop at an email to your colleague. We have to remember that software is being used to control 4,000-pound cars that need to make safe decisions. That’s part of the reason why there’s just so much money being invested into the compute needed to train these models.
Lyft’s latest move shows that more and more companies are pivoting their business models to slot into the AV race as quickly as possible, which is great for MaaS providers like May Mobility, but could spell disaster for employees, shareholders and folks who rely on the company’s services if there’s ever a time when the company decides to pull the plug.
What do you think—are robotaxis at risk for the AI bubble? Let me know your thoughts in the comments.
