Perhaps the most revealing moment in the PayPal saga happened at its outset. In March, 2000, as the merger was under way, Musk was driving his new McLaren, with Thiel in the passenger seat. The two were on Sand Hill Road, an artery that cuts through Silicon Valley. Thiel asked Musk, “So what can this do?” Musk replied, “Watch this,” then floored the gas pedal, hit an embankment, and sent the car airborne and spinning before it slammed back onto the pavement, blowing out its suspension and its windows. “This isn’t insured,” Musk told Thiel. Musk’s critics have used the story to illustrate his reckless showboating, but it also underscores how often Musk has been rewarded for that behavior: he repaired the McLaren, drove it for several more years, then reportedly sold it at a profit. Musk delights in telling the story, lingering on the risk to his life. In one interview, asked whether there were parallels with his approach to building companies, Musk said, “I hope not.” Appearing to consider the idea, he added, “Watch this. Yeah, that could be awkward with a rocket launch.”
Of all Musk’s enterprises, SpaceX may be the one that most fundamentally reflects his appetite for risk. Staff at SpaceX’s Starship facility, in Boca Chica, Texas, spent December of 2020 preparing for the launch of a rocket known as SN8, then the newest prototype in the company’s Starship program, which it hopes will eventually transport humans to orbit, to the moon, and, in the mission Musk speaks about with the most passion, to Mars. The F.A.A. had approved an initial launch date for the rocket. But an engine issue forced SpaceX to delay by a day. By then, the weather had shifted. On the new day, the F.A.A. told SpaceX that, according to its model of the wind’s speed and direction, if the rocket exploded it could create a blast wave that risked damaging the windows of nearby houses. A series of tense meetings followed, with SpaceX presenting its own modelling to establish that the launch was safe, and the F.A.A. refusing to grant permission. Wayne Monteith, then the head of the agency’s space division, was leaving an event at the Cape Canaveral Space Force Station when he received a frustrated call from Musk. “Look, you cannot launch,” Monteith told him. “You’re not cleared to launch.” Musk acknowledged the order.
Musk was on site in Boca Chica when SpaceX launched anyway. The rocket achieved liftoff and successfully performed several maneuvers intended to rehearse those of an eventual manned Starship. But, on landing, the SN8 came in too fast, and exploded on impact. (No windows were damaged.) The next day, Musk visited the crash site. In a picture taken that day, Musk stands next to the twisted steel of the rocket, dressed in a black T-shirt and jeans, looking determined, his arms crossed and his eyes narrowed. His tweets about the explosion were celebratory, not apologetic. “He has a long history of launching and blowing up rockets. And then he puts out videos of all the rockets that he’s blown up. And like half of America thinks it’s really cool,” the former NASA administrator Jim Bridenstine told me. “He has a different set of rules.”
Hans Koenigsmann, then SpaceX’s vice-president for flight reliability, started working on a customary report to the F.A.A. about the launch. Koenigsmann told me that he felt pressure to minimize focus on the launch process and Musk’s role in it. “I sensed that he wanted it taken out,” Koenigsmann said. “I disagreed, and in the end we wound up with a very different version from what was originally intended.” Eventually, Koenigsmann was told not to write a report at all, and a letter was sent to the F.A.A. instead. The agency, meanwhile, opened its own investigation. Monteith told me that he agreed with Musk that the F.A.A. had been conservative about a situation that presented little statistical risk of casualties, but he was nevertheless troubled. “We had safety folks who were very upset about it,” Monteith recalled. In a series of letters to SpaceX, Monteith accused the company of relying on data “hastily developed to meet a launch window,” launching “based on ‘impressions’ and ‘assumptions,’ ” and exhibiting “a concerning lack of operational control and process discipline that is inconsistent with a strong safety culture.” In its responses, SpaceX proposed various safety reforms, but also pushed back, complaining that the F.A.A.’s weather model was unreliable and suggesting that the agency had been resistant to discussions about improving it. (SpaceX did not respond to requests for comment.)
The following March, Steve Dickson, then the F.A.A.’s administrator, called Musk. The two men spoke for thirty minutes. Like Kahl, Dickson was deferential, thanking Musk for his role in transforming the commercial space sector and acknowledging that SpaceX was taking steps to make its launches less risky. But Dickson, an F.A.A. spokesperson said in a statement, “made it clear that the FAA expects SpaceX to develop and foster a robust safety culture that stresses adherence to FAA rules.” Dickson had navigated such conversations before, including with Boeing after two 737 max aircraft crashed. But this situation presented a thornier challenge. “It’s not every day that the F.A.A. administrator releases a statement about a phone call that they have with the C.E.O. or the head of an aerospace company,” an official at the agency told me. “That kind of gets into the soft pressure, public pressure that you don’t do unless you are trying to change the incentive structure.”
The F.A.A. issued no fine, though it grounded SpaceX for two months. “I didn’t see that a fine would make any difference,” Monteith told me. “He could pull that out of his pocket. However, not allowing launches, that would get the attention of a company that prides itself on being able to iterate and go fast.” Musk has continued to complain about the agency. After it postponed another launch, he tweeted, “The FAA space division has a fundamentally broken regulatory structure.” He added, “Under those rules, humanity will never get to Mars.”
Musk has been fixated on space since his childhood. The idea for SpaceX came about after his exile from PayPal. “I went to the NASA website so I could see the schedule of when we’re supposed to go” to Mars, Musk told Wired, in 2012. “At first I thought, jeez, maybe I’m just looking in the wrong place! Why was there no plan, no schedule? There was nothing.” In 2001, he connected with space-exploration enthusiasts, and even travelled to Russia in an unsuccessful bid to buy missiles to use as rockets. The next year, he moved to Los Angeles, closer to California’s aerospace industry, and ultimately he pulled together a team of engineers and entrepreneurs and founded SpaceX, to make his own rockets. Private rocket launches date back to the eighties, but no one had attempted anything on the scale that Musk envisioned, and it proved to be more difficult and expensive than he had anticipated. Musk has said that, by 2008, the company was nearly bankrupt, and that, after putting much of his wealth into SpaceX and Tesla, he wasn’t far behind. “That was definitely the worst year of my life,” he said in an interview on “60 Minutes.” SpaceX’s first three launches had failed, and there was no budget for another. “I had no more money left,” Musk told Bridenstine, the NASA administrator, years later. “We managed to put together enough spare parts to do a fourth launch.” Had that failed, he added, “SpaceX would have died.” The launch was successful, and NASA soon awarded SpaceX a $1.6-billion contract to resupply the International Space Station. In 2020, the company flew its first manned mission there—ending nearly a decade of American reliance on Russian craft for the task. SpaceX now launches more satellites than any other private company, with four thousand five hundred and nineteen in orbit as of July, occupying many of Earth’s orbital routes. “Once the carrying capacity of an orbit is maxed out, you’ve basically blocked everyone from trying to compete in that market,” Bridenstine told me.
There are competitors in the field, including Jeff Bezos’s Blue Origin and Richard Branson’s Virgin Galactic, but none yet rival SpaceX. The new space race has the potential to shape the global balance of power. Satellites enable the navigation of drones and missiles and generate imagery used for intelligence, and they are mostly under the control of private companies. “The U.S. government is in massive catch-up to build a more resilient space architecture,” Kahl, the former Pentagon Under-Secretary, told me. “And that only works if you can leverage the explosion of commercial space.” Several officials told me that they were alarmed by NASA’s reliance on SpaceX for essential services. “There is only one thing worse than a government monopoly. And that is a private monopoly that the government is dependent on,” Bridenstine said. “I do worry that we have put all of our eggs into one basket, and it’s the SpaceX basket.”
Even Musk’s critics concede that his tendency to push against constraints has helped catalyze SpaceX’s success. A number of officials suggested to me that, despite the tensions related to the company, it has made government bureaucracies nimbler. “When SpaceX and NASA work together, we work closer to optimal speed,” Kenneth Bowersox, NASA’s associate administrator for space operations, told me. Still, some figures in the aerospace world, even ones who think that Musk’s rockets are basically safe, fear that concentrating so much power in private companies, with so few restraints, invites tragedy. “At some point, with new competitors emerging, progress will be thwarted when there’s an accident, and people won’t be confident in the capabilities commercial companies have,” Bridenstine said. “I mean, we just saw this submersible going down to visit the Titanic implode. I think we have to think about the non-regulatory environment as sometimes hurting the industry more than the regulatory environment.”
In early 2022, Steven Cliff, then the deputy administrator of the Department of Transportation’s National Highway Traffic Safety Administration, learned that potentially tens of thousands of Tesla vehicles had a feature that he found concerning. For years, Tesla has been working to create a totally self-driving car, a long-standing ambition of Musk’s. Now Cliff was told that a version of Tesla’s Full Self-Driving software, an experimental feature that lets the cars navigate with little intervention from a driver, permitted cars to roll through stop signs, at up to about six miles an hour. This was clearly illegal. Cliff’s enforcement team contacted Tesla, and, in several meetings, a surprising conversation about safety and artificial intelligence played out. Representatives for Tesla seemed confused. Their response, as Cliff recalled, was “That’s what humans do all the time. Show us the data, why it’s unsafe.” N.H.T.S.A. officials told Tesla that, regardless of human compliance, “you should not be able to program a computer to break the law for you.” They demanded that Tesla update all the affected cars, removing the feature—a recall, in industry terms, albeit a digital one. “There was a lot of back-and-forth,” Cliff told me. “Like, at midnight on the very last day, they blinked and ended up recalling the rolling-stop feature.” (Tesla did not respond to requests for comment.)
Musk joined Tesla as an investor in 2004, a year after it was incorporated. (He has spent years defending the formative nature of his role and was eventually, in a legal settlement, one of several people granted permission to use the term “co-founder.”) Musk was again entering a market bound by entrenched private interests and stringent regulation, which opened him up to more clashes with regulators. Some of the skirmishes were trivial. Tesla for a time included in its vehicles the ability to replace the humming noises that electric cars must emit—since their engines make little sound—with goat bleats, farting, or a sound of the owner’s choice. “We’re, like, ‘No, that’s not compliant with the regulations, don’t be stupid,’ ” Cliff told me. Tesla argued with regulators for more than a year, according to an N.H.T.S.A. safety report. Nine days after the rolling-stop recall, the company pulled the noises, too. On Twitter, Musk wrote, “The fun police made us do it (sigh).”
“It’s a little like Mom and Dad and children. Like, How far can I push Mom and Dad until they push back?” Cliff said. “And that’s not a recipe for a strong safety culture.”
The fart debate had low stakes; the over-all safety of the cars is a far greater matter. Tesla has repeatedly said that Autopilot, a more limited technology than Full Self-Driving, is safer than a human driver. Last year, Musk added that he would be “shocked” if Full Self-Driving didn’t become safer than human drivers by the end of the year. But he has never made public the data needed to fully corroborate those claims. In recent months, new crash numbers from the N.H.T.S.A., which were first reported by the Washington Post, have shown an uptick in accidents—and fatalities—involving Autopilot and Full Self-Driving. Tesla has been secretive about the specifics. A person at the N.H.T.S.A. told me that the company instructed the agency to redact specifics about whether driver-assistance software was in use during crashes. (By law, regulators must abide by such requests for confidentiality, unless they decide to contest them in court.) Pete Buttigieg, the Secretary of Transportation, recently said that there were “concerns” about the marketing of Autopilot. Cliff told me he had seen data that showed Teslas were involved in “a disproportionate number of crashes involving emergency vehicles,” though he said that the agency had not yet determined whether the technology or the human drivers was the cause. In a statement, a spokesperson for the agency said, “Multiple investigations remain open.”