Tekedra Mawakana won’t take the bait, no matter how many times I try. The co-CEO of Waymo, which operates the biggest fleet of driverless cars in the country, is rolling through the streets of downtown San Francisco with me in a modified Jaguar I-Pace. Elon Musk has claimed he’s coming to take down Waymo and its more than 1,500 robotaxis with competing Teslas that can operate in “Full Self-Driving” mode—beginning with a 20-vehicle invite-only autonomous taxi service that’s started testing in parts of Austin. In 2019 Musk pledged to have a million Tesla robotaxis in service by 2020. So far, the Austin experiment hasn’t exactly been flawless. Can you believe he’s trying this again? I pepper Mawakana in the back of the Jag. When his “self-driving” cars need a human in the seat to keep from crashing? “I don’t know,” she half-whispers, looking out the window. She lets out a quiet challenge to Musk, even as she makes it a point not to say his name. “There’s zero evidence of anything. And there’s so much talk about it. We have a whole service across many cities, right? It’s not a play to have fanboys. It’s a play to, like, actually change people’s lives.”
Waymo and Tesla are in the same game, at least from this angle. Mawakana’s company has built an autonomous driver, made of a suite of sensors and software, that can be used in vehicles from multiple car companies; Musk has said that if his cars can’t drive themselves, Tesla is “worth basically zero.” Each is using robotaxis as proof of concept. The similarities end there.
Mawakana, like the business she helps lead, is deliberate, strategic, focused, cautious. Waymo is her one and only company, and she shares the leadership duties. Musk is Musk. Mawakana is a safety obsessive, as is Waymo. Musk is emblematic of the move-fast-and-break-things ideology. While Full Self-Driving Teslas have killed multiple people, Waymo vehicles have not killed a single person. Musk is apparently convinced that all Teslas need to drive themselves are advanced cameras and the right amount of AI. His cars start at $42,000; hers, outfitted with 29 of those cameras plus advanced radar, laser range finders, and acoustic sensors, can cost seven times as much. She’s a lawyer by training and spent much of her career at the nexus of government and technology. Musk, in fairness, also spent time at the nexus of government and technology—130 rocky days. She’s one of a handful of Black female (or Black, or female) chief executives in Silicon Valley; he goes on his social media platform to whine about how “teachers in California spend their time indoctrinating kids in DEI racism & sexism & communism.” It’s hard to imagine Musk on skates (though he appears to be making good on his promise to build a “roller skates & rock restaurant” in Los Angeles). Mawakana says, “I can be found with roller skates and a beach tent in my trunk at all times.”
It sounds like a cute study in contrast. Except the future of the global automotive industry—and how we move through cities—could be at stake, depending on which CEO’s vision wins out. Tesla is by far the bigger company, with annual revenue of almost $98 billion and about $7 billion in earnings, compared to an estimated $75 million in revenue and an estimated $1.12 billion in losses last year for Waymo. Mawakana doesn’t have great answers for how many jobs her autonomous vehicles might eventually cost (opponents say it could be millions), or how many people it currently takes to supervise her self-driving fleet (her deputies will only talk about the “minutes of human time” needed for each hour on the road), or the resilience of that fleet when protesters start lighting her robocars on fire (in Los Angeles last June). But when it comes to autonomy, Mawakana is ahead. Waymo is logging at least 250,000 driverless paid rides per week in Austin, Phoenix, Los Angeles, and Silicon Valley, more than 10 million such rides overall. Waymo had expanded its operations to Atlanta on the day Mawakana and I rolled through San Francisco together. Miami and DC are next, and Waymo has begun preparation for New York City.
“She believes the technology should speak for itself, in customers’ hands, as opposed to telling customers what to believe about the future,” says Alex Roy, a general partner at New Industry Venture Capital and a former executive at Argo, the Ford-backed autonomy company. “That’s why I call her the un-Elon.”
Mawakana lives a low-key life, at least by Silicon Valley mogul standards. She wasn’t invited to the Gilded Age Bezos wedding in Venice. She doesn’t bowhunt with Zuck, and she’s not joining Sam Altman in his doomsday structures. She lives in the same Bay Area home she bought when she was briefly a VP at eBay, in 2016 and 2017. Her art collection is modest, largely centered around Black and female artists, and is a “celebration of goddess energy.” She spends a chunk of her weekends going to her son’s basketball games; sometimes she’ll take him to Coachella or Rolling Loud. Vacations often involve a group of girlfriends sailing in the Caribbean or going to EDM festivals. Her big splurge last year was a quick trip to Paris for Vogue World.
“I’ve been, for a long while, toiling away at this and pretty under the radar,” she says. “I genuinely have come to this through a very different path than most people and have arrived in it in a different capacity. And so, like, you know, it looks different” from the tech-CEO cliché.
Yet she and Waymo are now central to what is arguably the most important conversation happening in tech, and maybe our broader society too: how far can automated systems go and how much human labor will they replace. And it’s all happening while Mawakana fends off a challenge from the world’s richest man, leading the planet’s best-capitalized carmaker. What happens next?
“She believes the technology should speak for itself, as opposed to telling customers what to believe about the future. That’s why I call her the un-Elon.”
I meet Mawakana at Waymo’s nondescript offices on Market Street, in a bland tower that used to be the local headquarters of Chevron Oil. She’s dressed in black jeans and gold Air Jordans. Minutes into our chat, she recalls a moment and lets out a combination of a laugh and a groan. Her only kid recently turned 16 and just got his learner’s permit. So now the driverless-car CEO is teaching her son to drive, and “it’s really stressful to me. Waymos won’t be everywhere before he’s out in the world,” she says. About 40,000 Americans die every year in traffic accidents. Her son may be “super careful” on the road but that makes him kind of an anomaly. “People aren’t that great driving. People are speeding. People are talking on the phone. Their baby’s crying in the back seat. There’s a lot happening.”
Every erratic driver multiplies the complexity. The streetscape dials up the confusion: pop-up construction sites, traffic cops in the middle of the road, bikers, swerving scooter riders, potholes, buses, stop signs hidden by trees, club kids pouring out into the night, children running out to grab a lost ball. Making sense of it all—and then reacting to every chaotic element in fractions of a second—has proven to be one of the most challenging problems in computer vision and machine learning. The smartest minds in the profession thought they might be on the verge of solving it 20 years ago. They were wrong.
I was there in the Mojave Desert in 2004 when the Pentagon’s research arm, the Defense Advanced Research Projects Agency, held a million-dollar contest to see which autonomous car could complete a 150-mile course the fastest. Only one made it just past seven miles. A Stanford professor named Sebastian Thrun tricked out a Volkswagen Touareg enough to complete the course the next year. Google tapped Thrun to help develop Google Street View and build an autonomous car for the masses, and that company eventually became Waymo. Every big automaker poured billions into similar projects, even as Waymo pivoted to making sensors and software that would go into other cars. By 2012, Google cofounder Sergey Brin said they’d be ready for the general public in less than five years. Tesla allegedly faked a video of Tesla’s self-driving capabilities to keep up. But Musk still confidently made his predictions, like autonomous cross-country Tesla trips by the end of 2017.
Mawakana would shake her head when she heard Musk talk like this. “Not a snowball’s chance in hell,” she remembers thinking back then. Despite the hype, Tesla’s Full Self-Driving program (and its predecessor, Autopilot) still required a human to take charge in times of trouble, making it closer to cruise control than true autonomy. Nor did Teslas have the array of sensors Waymos did: radar, to make sense of their surroundings in foul weather, and lidar, which relies on laser pulses, to “see” in the dark. And while Musk wanted his Teslas to be able to drive themselves anywhere, Waymo tried to constrain the problem by spending years mapping the streets of individual cities, starting in the Phoenix area. The company spent years more with Waymo-equipped vehicles driving themselves around those streets with humans behind the wheels to watch when and how the robocars screwed up. “It was when we removed the safety drivers that we realized in 2020 how much further we had to go,” she says. Every edge case and odd situation led to 10 more to figure out. She picks out one: “You hail the car. We know nothing about you. It’s a four-lane road that you’re standing in front of. We stop in the middle of the block across the street from you. And you’re vision-impaired. What do we do? How do you get to us?”
It was too many variables for most of the big carmakers at too high a cost. One by one, they dropped their efforts to build their own mass-market autonomous vehicles. The few that remained began to face a human toll. Teslas running the company’s Autopilot system were involved in 736 crashes over four years, according to federal statistics. Seventeen were fatal. General Motors invested more than $10 billion in Cruise, which had been operating the largest robotaxi fleet in San Francisco. Then one of Cruise’s cars struck a pedestrian and dragged her underneath the vehicle for 20 feet. The outcry was deafening and only grew louder when the company was caught trying to hide details of the incident from federal regulators. GM shut down Cruise’s robotaxi operations in December.
Mawakana and Waymo have managed to avoid those nightmare scenarios so far. Part of that is by staying away from riskier rides: Mawakana and her co-CEO, Dmitri Dolgov, announced an end to an autonomous trucking program in the summer of 2023 to focus on commercial autonomous ride hailing, and two years later Waymo is close to allowing these self-driving cars on highways. Part of it is cultural: Waymo executives told me Mawakana has created an environment where they’re free to hit pause on major developments if they’re not sufficiently safe.
But a big part is technical. David Margines, Waymo’s director of product development, shows me a video of a recent almost-incident in December. A young girl falls off her scooter and into a two-lane road. The self-driving car smoothly moves clockwise around her. “What do humans do in situations like this? They either turn or they brake,” he tells me, drawing diagrams on a whiteboard. “When we were avoiding this, we had a half of a g of deceleration combined with a third of a g of turning at the same time.” It’s a different—and maybe better—way of driving.
On my trip with Mawakana around San Francisco, an older man in a floppy hat stepped out in front of the car right as we were rounding a corner near Telegraph Hill; the Jaguar stopped itself with plenty of space to spare. Downtown, when a fire truck was coming up fast in the left lane, the Jaguar quickly dipped from the middle to the far-right lane to give it maximum room.
How safe are Waymos, really? Depends on whom you ask, and how. Waymo boasts of having driven 71 million miles without a fatal accident. But the average for human drivers is about 1.2 fatalities per 100 million miles. The company claims its autonomous driver, compared to humans, has 88 percent fewer accidents that result in serious injuries. Research out of the University of California, Berkeley, suggests the robotaxis are about as safe as the average flesh-and-blood Uber driver. Outside safety researchers give Waymo credit for publishing peer-reviewed studies of its data; Tesla, for example, doesn’t do anything like it, and the data Musk’s company does release is debatable at best, these experts say. But the safety researchers I spoke with also contend that Waymo isn’t providing enough information to back up its safety claims. “They are seen as the most trustworthy player in the industry that is notorious for opacity,” says Carnegie Mellon University engineering professor Phil Koopman. “Their marketing, PR people have been claiming ‘saving lives’ for years now, and they do not have enough data to know whether that’s true.”
I’m the target market for a robocab. I didn’t get my license until I was 26, and that was after failing my first two driver’s tests. My skills behind the wheel, or lack thereof, are legendary in my family. With extensive practice, I can now parallel park—sometimes, if the space has been vacated by a school bus. So yes, I would love a robot chauffeur to take me around. But that comes with a cost.
Driving is one of the most common jobs in the country, especially for folks without a college degree. There’s something like 3 million full-time truck and delivery drivers, plus millions of more people driving Lyfts, Ubers, and old-school cabs. And while Waymo and other autonomous vehicle companies are concentrating on robotaxis right now, it’s clear that’s not the endgame. The endgame is replacing as many of those humans as possible with a Waymo driver—and to make your car robo-friendly too. The taxis are just a way to train the system. “Today, we’re deploying that in a Jaguar,” says Margines. “We’re going to do it with trucks someday, and we’ll do it on personally owned vehicles.” Which means whatever benefits autonomous driving brings—and I’m convinced there will be many—it will likely cost millions of people their livelihoods. “There’s not going to be driving jobs, and companies like Waymo don’t give a shit,” says Peter Finn, who heads Teamsters Local 856 in Northern California. The state Assembly has passed bills trying to keep a person in commercial vehicles, Finn adds, but Big Tech’s managed to prevent those bills from going anywhere. Waymo alone spent nearly $5 million on lobbying for various issues in California’s last legislative session.
Mawakana says she understands the anxiety. She grew up working-class, the daughter of an airman first class in the US Air Force, in Sandy Springs, north of Atlanta. Her uncle Joe Nathan was a truck driver in Mississippi. “It was a hard job. He ate in that truck, and he lived in the truck, and he died in that truck,” she says. But the best she can do to alleviate Finn’s concerns is to note that there has been a truck-driver shortage in recent years—and to riff about how the horse-and-buggy economy eventually grew into a more vibrant one based around the automobile.
But there are downsides to that vision. “If you give everyone’s car like a robot chauffeur, it would actually have a horrible impact,” says Edward Niedermeyer, author of The Autopilot Effect, a forthcoming book on the Tesla-Waymo competition. “Most of traffic is caused by single-occupant vehicles. Private AVs [autonomous vehicles] would give people zero occupant vehicles. You’d send your car off to deliver stuff. Why would you park the car when it could just circle blocks endlessly and just create traffic.”
And while the company may say it is “optimistic about how autonomous driving technology will create demand for many new businesses and jobs over time,” for now maintaining Waymo’s current fleet takes fewer people than you might think. The company claims in a statement it “directly and indirectly supports 2,000+ jobs” in the city and county of San Francisco, but that figure includes jobs “that benefit from increased economic activity (e.g., retail, hospitality, entertainment).” I was expecting to see a crew of mechanics at the company’s depot on Toland Street. Instead there were just a handful of contractors, swapping out tires and plugging the vehicles into their electric charging stations. The cars mostly drove themselves around the facility and parked themselves when their human caretakers were done with them.
Waymo also has people working in call centers who can respond to riders’ questions and teams that monitor and can provide guidance to the robotaxis if they get stuck. Exactly how many of these “remote assistants” are doing this work, where they’re located, and how many cars each person supervises neither Mawakana nor anyone else at Waymo will say exactly. All they’ll note is that there are people both in the United States and “outside,” and that, for every hour of vehicle time in San Francisco, Waymo needs “only a few minutes of human time on average.” Tesla may be advertising for “teleoperation” jobs to “access and control” its supposedly self-driving cars. (An inherently unsafe proposition, experts tell me, because of the lag time involved.) But for Waymos in the United States, Mawakana and her team insist, there’s no off-site steering wheel or joystick for a human to drive the cars from afar.
In an earlier job, Mawakana was forced to sit on a far more important secret. US intelligence agencies were sucking up huge amounts of data from commercial internet services—and ordered the companies not to say a thing about it. Mawakana was about a month into a new job as deputy general counsel of Yahoo when Edward Snowden’s revelations about the security state’s subornment of Silicon Valley started to come out. “On that day, we couldn’t go out and say to our users, ‘This is what’s happening.’ Was against the law,” she recalls. “There was an entire ecosystem that was happening on the back end that consumers didn’t even know about.” The disconnect made her seriously uncomfortable. Eventually, she left for eBay, and then for Waymo in 2017.
She became the head of global policy. At many tech companies that can be an ancillary role, but policy is critical to Waymo’s expansion. If it couldn’t persuade local governments to let its autonomous cars onto their roads, if it couldn’t build some kind of community comfort with the idea of robots on the streets, Waymo had no business, no commercial future. “I remember going home from the interview and being like, Okay, this is ‘put your money where your mouth is,’ mama,” she tells me. By 2021 she was named co-CEO.
That put her in rarified company. According to a 2024 report from the US Equal Employment Opportunity Commission, less than a quarter of American tech workers are women and less than 8 percent are Black. That hasn’t stopped a slew of MAGA-friendly executives from demanding more “masculine energy” in the workplace and trying to roll back the diversity, equity, and inclusion efforts of the Black Lives Matter era. “At every level of my career, I have been one of very few,” Mawakana says. So when Musk and his mates shout that DEI is so dangerous, “it’s like railing against a machine that didn’t yield a lot of results. What it did yield is more voices at the table, more opinions. More is great. It’s been great. It’s been wonderful. But it hasn’t been transformative. And it definitely hasn’t been status quo–threatening.”
Under Mawakana, Waymo stood up to a trade organization to make laws more robocar-friendly in state capitals. She got together with groups like Mothers Against Drunk Driving and the Foundation for Blind Children so they could help plead the case for Waymos. “They couldn’t take their seeing-eye dogs into Ubers and Lyfts without being discriminated against,” she recalls. “They didn’t feel comfortable with their 22-year-old daughter after a club.” Still, there was major pushback, especially in San Francisco, where the city attempted to sue state regulators for granting licenses for the robotaxis to operate (a state court rejected San Francisco’s legal challenge in January). But even the city attorney acknowledged that Waymo was a relatively “good actor” in the field.
This is the kind of tiptoeing approach that Tesla fans would ordinarily make fun of. Except Musk’s robotaxi rollout in Austin is more cautious still. The cars may not go out in inclement weather and they won’t operate downtown, at least not at first. Finally, these driverless cars will have, in addition to the teleoperators, a “Tesla safety monitor sitting in the right front passenger seat.” Musk is once again promising those million robocabs are coming soon. Perhaps this time he’ll be proven right. But for now, Alex Roy notes, “Tesla is literally deploying, like verbatim, the deployment strategy used by Waymo.” With one key difference: Mawakana has been doing this for years, and Musk is millions of miles behind her.
More Great Stories From Vanity Fair
-
How a Death Row Murderer Exposed One of America’s Most Prolific Serial Killers
-
Zen and the Art of Being Jennifer Aniston
-
Eliot Spitzer Speaks His Piece
-
On Set for The Pitt Season Two
-
The Singular Style of Princess Anne
-
Where Today’s NFL Players Become Tomorrow’s Pundits
-
A Mission Divided at the Chan Zuckerberg Initiative
-
24 TV Shows We Can’t Wait to See This Fall
-
Inside Coco Chanel’s French Riviera Getaway
-
Weapons Ending Explained
-
From the Archive: Dating Jeffrey Epstein
The post Elon Musk Has His Vision. Waymo Chief Tekedra Mawakana Says She’s Got a Better One appeared first on Vanity Fair.
