12

“Space Is Hard”

THE ROCKET DETONATED suddenly into a giant orange fireball, sending a mushroom cloud floating ominously in an otherwise pristine, sun-drenched Texas sky. Bits of debris scattered, trailing smoke and fire on the way down like fireworks in a display as beautiful as it was violent.

As it climbed a few hundred feet over the company’s McGregor test site, the rocket had suddenly spun out of control and started falling back down. Before it could get too far off course, its “flight termination system” kicked in, blowing the rocket up a few hundred feet above the prairie. No one was injured. And it was only a test, one that SpaceX stressed “was particularly complex, pushing the limits of the vehicle further than any previous test.” Elon Musk even had coined an acronym for such spectacular failures: RUD, or rapid unscheduled disassembly.

But it was also a reminder that for all the advancements in rocket science, launches were really just controlled explosions of a combustible mix of propellants. As Musk knew as well as anyone, one small error, even something as small as a corroded nut, could send the whole thing up in flames. “Still so damn intense,” Musk had tweeted after a recent launch. “Looking fwd to it feeling normal one day.”

After the rocket test explosion, he tweeted: “Rockets are tricky.”

Still, SpaceX had rattled off an improbable string of successful launches of the Falcon 9 without a single failure. It was an amazing streak stretching over four years that was beginning to make the exercise feel routine. But Musk still sweated every one, shooting an e-mail to the entire company, urging employees to step forward if anyone had a reason to call off the flight. He was the CEO-cum-wedding officiant—speak now or forever hold your peace.

SpaceX’s success had raised expectations, but its brashness had attracted criticism. Its fan base was huge and growing. The SpaceX page at Reddit, the social media site, had ten thousand subscribers in June 2014. More and more people were buying the $22 “Occupy Mars” T-shirts from the SpaceX online store. And Musk had become something more than a business executive; he was now a cult figure, whose legend was growing well beyond Silicon Valley.

With Tesla and Solar City, the solar energy company, he had set out to transform American transportation and energy use. SpaceX was an improbable success story that was not only disrupting the industry but singlehandedly reigniting interest in space. CBS’s 60 Minutes declared that Musk had built an “industrial empire.” Time magazine put him on the cover of its “100 Most Influential People” issue. The Atlantic canonized him as “perhaps this era’s most ambitious innovator.”

“In the spirit of inveterate and wide-ranging tinkerers like Leonardo da Vinci and Benjamin Franklin, Musk has transformed virtually every field he’s taken an interest in, from electronic payments to commercial spaceflight to electric cars,” the magazine wrote. “The range and scale of Musk’s ambitions have attracted skepticism, but over time, he has proved himself to be not only an ideas man but an astute business thinker.”

But within the somewhat clubby space community, SpaceX was becoming the company people loved to hate. At a space industry party, there was a photo Musk taped to the inside of the toilet so that his competitors could take turns pissing on him.

Once dismissed as an “ankle biter,” SpaceX was now a formidable competitor, one to reckon with. It was also gunning for the United Launch Alliance’s breadbasket—the lucrative launches for the Pentagon and intelligence agencies.

For a decade, the company had a monopoly on the contracts, worth hundreds of millions of dollars. A decade earlier, Musk had sued, arguing that SpaceX should be allowed to compete. But without a rocket capable of flying, the suit was dismissed.

Now he had a rocket. He just didn’t yet have the US Air Force’s certification required for the launch—and the Pentagon was about to award another big batch of contracts to the Alliance, effectively locking SpaceX out for years. Filing another lawsuit would be risky—it’s usually not the best business practice to sue the agency you’re trying desperately to get to hire you.

The list of cons was long. But the pros were substantial as well. National security launches paid big money—the multiyear program could be worth as much as $70 billion—and SpaceX knew it could undercut Lockheed’s and Boeing’s prices, disrupting the market, giving it a stream of revenue that could sustain it for years and help it get to Mars. But the clock was ticking. If the company was going to protest the contract, it had to act quickly.

“Suing the military industrial complex is something you do not take lightly,” Musk recalled.

During a visit to Washington, DC, while sitting in the back of a sedan after a speech, a pair of his advisors asked him what he wanted to do.

Musk went quiet, closed his eyes, and put his head back. He stayed that way for two minutes, then three. A long time. He had several quirks, and his sudden retreat into his own mind was an eccentricity the people at SpaceX were used to. People coming in to interview with Musk were sometimes warned that when he goes silent, it was because he’s thinking and it’s best not to interrupt him. The advisors knew not to say a word. Six minutes passed. Then eight. An eternity.

“I’d seen him go Zen before, but I’d never seen him go this Zen,” one of the advisors recalled.

Then Musk opened his eyes. “File the lawsuit,” he said. He got out of the car and went to the next event.

The advisors looked at each other and one of them said, “He just teleported himself into the future!”

ONCE THE LAWSUIT was filed, Musk continued to attack throughout the spring and summer of 2014, delighting the press corps in the nation’s capital, who were unaccustomed to such a bombastic character. “Musk is,” one defense reporter wrote, “a good interview.”

Which meant he was unfiltered, a refreshing change of tone in a buttoned-up town where officials rarely deviated from the script. At an event at the National Press Club, Musk defended the lawsuit, saying SpaceX should be given the opportunity to compete, and he derided the air force’s certification process, calling it a “paperwork exercise.” If his rockets were good enough for NASA, he said, they should be good enough for the Pentagon.

He purposefully picked a fight with the United Launch Alliance (ULA). It was the dominant player, but it had a big weakness. The RD-180 engines it used for its Atlas V rocket were made in Russia, and this was coming at a time of increased tension between the United States and Russia over the latter’s annexation of Crimea. Musk went after the Alliance relentlessly.

At a reception at the Newseum near the Capitol that spring, as he showed off the crew version of Dragon, Musk stood in the middle of a massive media scrum to deliver an indictment of the entire process.

“Our toughest competitor on the international launch market is the Russians, and the US Air Force sends them hundreds of millions of dollars every year for Russian engines,” he said. “It’s super messed up. I mean, what the fuck, you know?…

“Can you imagine if you went back forty years ago and told people that in 2014 the United States would be at the mercy of Russia for access to low Earth orbit, let alone the moon or anything else? People would have thought you were insane. It’s just incredible that we’re in this position. Something needs to be done to get us out of this.”

In response to a question about the wisdom of going up against the Alliance, he said that “Eisenhower warned about the military-industrial complex, and he ought to know. Has it gotten better or worse since Eisenhower? It hasn’t gotten better.… Lockheed and Boeing are used to stomping on new companies, and they certainly tried to stomp on us. I think we’ve got a shot at prevailing. We’re certainly a small up-and-comer going against giants.”

The bombast, the lawsuit, the media attention began to tick off the Pentagon. The head of the air force’s space command at the time told a reporter, “Generally, the person you are doing business with you don’t sue.”

For the first time, the Alliance began to fight back publicly, highlighting its long history over SpaceX’s inexperience in a marketing campaign—what it called “results over rhetoric.”

“The whole tenor of the campaign is to make perfectly clear that there is a lot at stake when it comes to successful space launches—literally lives are at stake,” Mike Gass, the CEO of the Alliance, said at a press conference. “We also want to make clear that there is a big distinction between a company that has a hundred-year combined heritage in successfully delivering satellites into orbit and a company that is not yet even certified to conduct one [national security] launch.…

“SpaceX is trying to cut corners and just wants the USAF to rubber stamp it,” Gass said. “SpaceX’s view is just ‘trust us.’ We obviously think that’s a dangerous approach and, thankfully, so do most people.”

SpaceX didn’t have anything to lose, and was ready to get into a fight.

“ULA doesn’t believe in competition. Monopolists never do,” spokesman John Taylor said in a statement. “In ULA’s case, it would rather call a press conference to announce an inside-the-Beltway lobbying campaign aimed at distracting lawmakers from the benefits competition brings to the marketplace: better technology, improved reliability and affordable prices.”

Two months later, Gass was ousted. Tory Bruno, the Alliance’s new chief executive, was brought in to make the company leaner, more efficient—able to compete with SpaceX, which was now threatening its business. Bruno vowed to “literally transform the company” by cutting the price of launch in half and developing a new rocket.

In addition to streamlining the business, the Alliance had a secret weapon in the war against SpaceX: Jeff Bezos.

FOR YEARS, BLUE Origin had been building a monster of a new rocket engine, one that stood 12 feet tall and had 550,000 pounds of thrust—even more than the engines that powered the space shuttle. The BE-4, as it was known, was not as robust as Bezos’s beloved F-1s, the most powerful rocket engines ever built. But it was designed to be a reliable workhorse, one that could fly again and again at a relatively low cost.

The fact that Blue Origin was developing its own engine, and building the infrastructure in West Texas to test it, was yet another sign that Bezos was dead serious about space, and that he had poured a vast amount of resources—perhaps even as much as $1 billion—into the development of the engine alone.

At a press conference at the National Press Club, Bruno and Bezos sat side by side before a banner with the phrase “Igniting the Future” to announce they were joining forces—Blue Origin would sell the BE-4 to the Alliance. That would allow the Alliance to avoid using the Russian-made RD-180—and just as important, take away Musk’s line of attack.

It was a shocking and unlikely marriage—the Lockheed Martin–Boeing conglomerate, which together had a century of experience in space, with Blue Origin, the quiet upstart that had plodded along carefully in the shadows. But now for one of the first times, Blue Origin was standing squarely in the spotlight—and it was doing so with SpaceX’s archenemy.

It’s kind of the best of both worlds,” Bruno beamed. “We have their innovative, entrepreneurship together with ULA’s solid track record of success, certainty and reliability.”

Bezos praised his new partner and its long heritage, noting that the Alliance “has for the last eight years put a satellite into orbit almost once a month. It’s an unmatched record of success and an incredible tribute to detail orientation and operational excellence.”

He geeked out over the technical details of the engine, discussing how its “oxygen-enriched stage combustion cycle” was better than a “gas generator” and how the engine has only a “single turbo pump” and had just “one shaft, so it’s as simple as it can be while still being high performing and highly reliable.”

Later that day, when Musk was asked about the United Launch Alliance–Blue Origin partnership, he was, as always, blunt: “If all your competitors are banding together to attack you, that’s, like, a good compliment,” he said. “I think a very sincere compliment.”

It also increased the pressure on SpaceX. Musk couldn’t afford a misstep. Not now. Not with his rivals gunning for him, and the Obama administration investing heavily in SpaceX, and Musk now a celebrity with the ability to move markets and make the media swoon with a single tweet. All of it was building to a Hollywood-like crescendo that was propelling Musk and his space company higher and higher, to a rarefied altitude where it finally had something to lose.

SEPTEMBER 16, 2014, was the 1,167th day since NASA had launched an astronaut from American soil, an embarrassing streak that stretched back to the last shuttle flight in 2011. Every day without a crewed launch brought NASA closer to breaking an ignominious record: the 2,098-day hiatus in human spaceflight between the last of the Apollo launches and the first shuttle flight in 1981.

But on the 1,167th day, the space agency had good news: its plan of how it would fly astronauts once again, an announcement that NASA administrator Charlie Bolden said set “the stage for what promises to be the most ambitious and exciting chapter in the history of NASA and human spaceflight.”

Two companies, SpaceX and Boeing, had won the contracts as part of NASA’s “commercial crew” program to fly the next generation of astronauts to the International Space Station, the agency announced. The companies would fly the same number of flights, and be required to hit the same milestones. But SpaceX had simply bid less, and as a result laid out in stark contrast the difference between itself and its rivals.

Boeing’s award was $4.2 billion. SpaceX would receive $2.6 billion.

Musk had been saying for years that SpaceX could fly cheaper and more efficiently than the traditional contractors, and NASA was taking him up on it—while also hiring Boeing, the more expensive, and experienced, company.

By now, SpaceX had flown the Falcon 9 and the Dragon spacecraft to the station multiple times. But Musk was eager to move to the next phase in his quest to get to Mars—flying actual people in his new version of the Dragon spacecraft. It looked like a sleeker, sexier version of the capsules that had taken the Apollo astronauts to space. Outfitted with reclined seats, giant screens, and a shiny interior, it could have passed for the VIP, bottle-service section of a nightclub. (In addition to being SpaceX’s founder and chief executive officer, he had the title of lead designer.) But unlike traditional spacecraft that splashed under parachutes into the ocean, the Dragon had its own engines, giving it the ability to land propulsively—using engine thrust to slow itself down—virtually anywhere on Earth.

“That is how a 21st century spaceship should land,” he said.

The White House’s risky bet to rely on the commercial sector, now being led by SpaceX, seemed to be playing out just as the Obama administration had hoped.

Four days after winning the commercial crew contract, SpaceX had flown yet another successful cargo flight to the station, and the Dragon was about to return home. Orbital Sciences, which along with SpaceX had been hired by NASA to fly passenger-less cargo missions to the station—had launched its Cygnus spacecraft to the station three times.

And now, on October 28, they were about to fly again. Leading up to the launch, Frank Culbertson, Orbital’s executive vice president and a former NASA astronaut, joked that the astronauts aboard the station might need “some of those red and green wands they use on the deck of an aircraft carrier” to direct all the spacecraft traffic coming and going.

Like SpaceX, Orbital was ready to fly to the station “more and more frequently and do it for many years to come,” he said. The goal was to make access to the station routine—“a stepping stone to what we’re going to do next, which is to go beyond low Earth orbit, go out to the moon, continue to explore that, and eventually go to Mars and to asteroids and to continue to explore our solar system.”

The weather was perfect for the launch, and the mood was upbeat. NASA—and the commercial sector—had momentum now and were eager to keep it going, even if some were concerned about allowing the still-young industry to fly NASA’s most precious resource: its astronauts.

FOR THE EVENING launch, crowds gathered along the Virginia shoreline, several people deep. Kids sat on parents’ shoulders for a better view. Some even crept on top of cars. The air was full of the illuminated screens of cell phone cameras ready to record the rocket blasting off. They counted down in unison: “Five. Four. Three. Two. One!” And cheered as Orbital’s Antares rocket lifted off over a ball of yellow-orange fire and smoke at 6:22 p.m., 15 minutes after sunset.

But within seconds, the majesty of the liftoff morphed into a menacing flash as the rocket exploded into a shrapnel-spewing fireball. A massive mushroom cloud filled the sky, scattering bits of debris like fireworks. A few miles away, the spectators could see it—and feel the heat—before they could hear it. Then, the blast arrived like a cannonball, knocking some off their feet and sending others running for cover.

The explosion incinerated the rocket and the 5,000 pounds of cargo it was to ferry to the International Space Station. It devastated the launchpad, leaving a crater 30 feet deep and 60 feet across that would cost $15 million to repair.

The hole it left in NASA’s plans to rely so heavily on the commercial sector, however, was even bigger.

THREE DAYS LATER, on October 31, 2014, Branson was home on Necker Island, his retreat in the Caribbean, talking on the phone to his son, Sam, who had just completed training in a centrifuge near Philadelphia, preparing for his spaceflight.

For years, Branson had been promising that the “world’s first spaceline” would soon be flying tourists to space. The first flights were supposed to start in 2009, but the date slipped again and again, until the company stopped providing dates. Instead it said that while it “expects to be the first company to provide sub-orbital flights to the general public (and certainly the best!)” it would launch only “when we are happy with the results of the exhaustive WhiteKnightTwo and SpaceShipTwo flight programme.”

It didn’t, however, tone down its promotional materials. For $250,000, Virgin promised quite a thrill: “Astronauts tell us that nothing can really prepare you for your first experience of space, but we will ensure that you are fully equipped to savour every second of an experience which will be intense, wonderful and truly unforgettable,” it declared.

Now, after years of delays, Virgin Galactic was getting ready to finally fly. Although company executives had warned him not to reveal a timetable for flights, Branson couldn’t contain himself. The company was so close that he’d told the media the first human test flights to space would take place by Christmas. Then, he and his son would go together in the early part of 2015, followed by paying customers.

The spacecraft had completed more than fifty test flights, but most of them were just unpowered “glide flights.” It had tested the “feather system,” the Rutan invention that would help it gently return to Earth, ten times. But even though it had conducted only four powered test flights by firing the engine, Branson and Virgin Galactic had gone into full marketing mode.

Virgin Galactic had signed up sponsorships with Grey Goose, the vodka maker, and Land Rover, which had sponsored a contest to send four winners to space. It had inked a deal with NBC to broadcast the first flight “in a primetime special airing on NBC the night before the launch, and a 3-hour live event on ‘TODAY,’ hosted by Matt Lauer and Savannah Guthrie,” the companies said in a press release. And it had signed a deal to fly its paying customers out of Spaceport America, the futuristic spaceport in New Mexico that cost taxpayers $220 million to build.

The first flights were nearly here—just weeks away—and Branson was already looking ahead to the future.

“I think we can in the years to come bring the price down so that a lot of people will have the chance to become astronauts,” he told an interviewer.

Like his son, Branson had also been training to go to space. Less than two weeks earlier, he had taken an acrobatic flight in a stunt plane to get his body used to the additional g-forces.

Flying above the Mojave Desert, he was ever the showman, asking his pilot, “Are we able to do a loop? Why don’t we do a little show-off over the runway.”

“You feel okay?” the pilot asked as they spun upside down, Earth swirling beneath them.

“Absolutely perfect,” Branson enthused. “That was brilliant!”

But now at Necker Island, he was talking with his son about his experience in the centrifuge, when he got an urgent note. It was from George Whitesides, the CEO of Virgin Galactic. There had been a catastrophic accident. Branson had to go.

PETER SIEBOLD WAS in the cockpit again. The forty-three-year-old test pilot for Burt Rutan’s Scaled Composites, which had built and designed SpaceShipTwo for Virgin Galactic, was ready to fly this time.

A decade earlier, during the Ansari X Prize, he’d had misgivings about the safety of SpaceShipOne, and dropped out of the program. But he had stuck with the SpaceShipTwo development, and on the morning of October 31 strapped himself into the copilot’s seat, alongside Michael Alsbury, the other test pilot.

They were close friends whose children played together on weekends. Both were self-taught flying obsessives who wanted their pilot’s licenses far more than their driver’s permits. They had gone to the same university, and jumped at the chance to become a test pilot for Rutan, trying out his latest inventions. Alsbury had been picked to fly the last vehicle Rutan built—a flying car—before he retired in 2011. Siebold had grown up flying in his father’s plane, from being perched on a stack of pillows when he was five years old so that he could experience the thrill of flight as if he were the pilot.

Now it was a bright October morning, and SpaceShipTwo was attached to the belly of its mothership, WhiteKnightTwo, which was taking the spacecraft up higher and higher. When it came to 50,000 feet, it was released and the pilots ignited the engines.

Siebold had considered the mission “high risk.” When asked why, he said they were “doing a significant envelope expansion that day. Flying an unproven rocket motor in an unexplored aerodynamic regime… classic test hazard assessments would categorize that as a high-risk flight.”

Also, the crew was “using a propulsion system that history has shown can be unreliable, or much less reliable than a turbine or reciprocating engine.”

Given how risky the flight was, Siebold took a moment to collect himself just before the spaceship was to be released from the mothership. He ran his hands over the rip cord of his parachute and his oxygen mask and his seatbelts, as if rehearsing the steps he would have to take in an emergency and to improve his “muscle memory.”

After WhiteKnightTwo released the spacecraft, Siebold and Alsbury fired the engines and soon they were soaring toward the heavens, breaking the sound barrier, 10 miles high.

“Ignition! #SpaceShipTwo is flying under rocket power again. Stay tuned for updates,” the company tweeted at 10:07 a.m.

The update six minutes later wasn’t good: “#SpaceShipTwo has experienced an in-flight anomaly. Additional info and statement forthcoming.”

THE LAST THING Siebold remembered of being in the spacecraft was a sickening jolt, grunting noises, a loud bang, and then cabin depressurization. The aircraft pitched up violently, he would tell investigators from the National Transportation Safety Board. The sound of the spacecraft coming apart was odd, delicate even, like “paper fluttering in the wind.” Then the crushing g-forces caused cerebral hypoxia, a lack of oxygen to the brain, and he blacked out.

When he came to, Siebold was outside the spaceplane in a free fall, his helmet askew, his oxygen mask shifted. The wind howled in his ears. The extreme cold air engulfed his body. There was something bothering his eyes, and when he opened them, he saw the wide expanse of the desert below.

The Mojave Desert floor rising fast.

Just as he had rehearsed a few minutes before, Siebold used his hands, Braille-like, to find the buckles of his seatbelt and release them. Falling through wispy cirrus clouds, his training kicked in, and he went into the free-fall position, opening his arms and legs out spread-eagle-style, to create drag.

The next thing he remembered was another jolt that surprised and possibly woke him up. He told investigators he wasn’t sure whether he had gone unconscious again. But if he had, the snap of the bright red parachute deploying automatically brought him back. His shoulder was killing him—he thought it was dislocated. As he floated under the parachute, he tried unsuccessfully to shove it back in place so that he could use it to steer.

He braced for a hard landing, right into a windswept creosote bush in the middle of the desert. As he waited for help, he noticed his chest was covered in blood. His arm was broken in four places and his right hand was numb, “as if he were out throwing snowballs without gloves.” He had corneal scratches in his eyes and he later had a piece of fiberglass removed from his left eye when at the hospital. But he was alive.

Emergency crews would find Alsbury’s lifeless body not far away in the wreckage, still in his seat. The coroner determined the cause of death was “blunt force trauma to the head, neck, chest, abdomen, pelvic area and all extremities and internal organs.”

He was thirty-nine years old, with two children, aged ten and seven.

AFTER HANGING UP on the call with his son, Branson hopped in a plane and headed to the crash site. He knew he had to get there as soon as possible.

This was the second fatal accident for the program. In 2007, three employees of Scaled Composites had been killed during a ground test of the engine’s nitrous oxide system. The explosion had charred the desert floor, which had looked like a war zone, with debris scattered everywhere and multiple people injured.

The California Occupational Safety and Health Administration fined Scaled Composites $28,870, later dropped to $18,560 after appeal.

The explosion was “obviously horrendous for the families and a big setback for our program as well,” Branson recalled later. “And after that I think we decided to take the testing in house and have our own team do it.”

That same year, his train company had had a fatal derailment in northeast England. He had gotten to the scene as soon as possible, racing over from a vacation in Zermatt, Switzerland.

“I knew the importance of getting there as fast as possible and being there as fast as possible and confronting it head on whether it’s your fault or not—actually particularly if it’s your fault,” he recalled.

Once he arrived, he spoke to the Virgin Galactic team before talking to the media.

“I addressed them all and reassured them as best I could that they had built a beautiful craft,” he said. “We had the biggest hug in history, and I made it clear we would continue knowing the spaceship was fundamentally fine.”

But the investigation by the National Transportation Safety Board was only just beginning. And questions from the press were mounting.

On the Today show, three days after the crash, Matt Lauer pressed Branson on the future of the company.

“People are already wondering whether this accident and the death of this copilot is a crippling blow,” Lauer said. “There have been delays; there have been setbacks in the past with this program. Can Virgin Galactic survive the image that has been seen around the world of that vehicle coming apart at forty-five thousand feet?”

Was it, Lauer wanted to know, worth the risks?

Branson had thought about this very question on the plane ride to the Mojave Desert. He’d been happy to risk his own life in hot-air balloons, in speedboats, and in all manner of stunts, that were daring and dangerous and good for business. But this was none of that. This was a sobering stop to a carousel that had perhaps spun too fast and too loudly for too long.

Maybe he should give it up. Maybe space was too hard. The company had spent $500 million or more on the venture and had still not flown a single person to space. But as Branson landed and met the team, they’d urged him on. He began to think he owed it not just to Alsbury who wouldn’t have wanted them to quit, but to them. Hadn’t every explorer faced such difficulty? This was the moment Virgin Galactic and the industry had feared—and prepared for. This was their crucible, their Apollo 1 moment. The time to decide whether they would retreat or reassemble and attack again stronger, if scarred.

“Absolutely, it’s worth the risks,” Branson said. “It’s a grand program, which has had a horrible setback. But I don’t think anybody watching this program would want us to abandon it at this stage.”

He’d made his decision. They’d carry on.

UNTARNISHED AND SEEMINGLY invincible, SpaceX ended 2014, one of its most successful years, with a massive holiday party. Why not celebrate? It had proved it could fly reliably to the International Space Station. It had won the right to fly astronauts, which combined with the cargo contract amounted to more than a $4.2 billion investment from NASA. It was signing up new commercial satellite customers, had added 39A to its mantle, was winning the war in Washington over the national security launches. And it was getting closer to landing a rocket safely so that it could be reused.

It was on a roll.

The holiday party was so big that a map detailed the various venues—including an indoor beach with hammocks and trucked-in sand, a casino, an “unclean room” where employees put on a full-body white suit and painted away. There was a dance hall with swirling lights and circuslike acrobats who swung from giant hula hoops from the ceiling. Rickshaws and a mini-train, the “SpaceXpress,” ferried partygoers across the great bacchanal, from the stocked bars to the game room with glowing foosball table and the “Candy” room with treats to be nibbled from the wall and a giant desert display. There was also a wall of donuts laid out in the form of the SpaceX logo and an adult-size ball pit that was, of course, right next to a bar.

“My favorite part of the @SpaceX X-Mas party—definitely the ball pit!” tweeted Garrett Resiman, a SpaceX executive and former NASA astronaut, who flew on two space shuttle missions and spent three months on the International Space Station.

SIX MONTHS LATER, they filed into SpaceX’s headquarters around dawn on a Sunday, ready to party again, ready to celebrate yet another milestone. They packed in several deep around mission control for the launch of the Falcon 9 scheduled for just after seven a.m. Pacific time on June 28, 2015. It was a lovely morning in Florida: mid-80s, light wind. There was only a 1 percent possibility of calling off the launch because of weather.

The drama leading up to the flight wasn’t the launch, but the landing. Or rather, the landing attempt. For months, SpaceX had been practicing an unprecedented maneuver, flying its first stages back to land on what it called its “autonomous spaceport drone ship.” Although each of the previous two attempts had ended in fireballs—“rapid unscheduled diassamblies”—the company was getting very close.

On both tries, the rocket had actually hit the ship, an incredible achievement considering the booster had to fly back to Earth after screaming into space. But each time, something had gone wrong at the last minute, giving the company enough material for a fiery blooper reel.

Now, Musk felt that SpaceX had finally figured out how to do it, and was confident about its chances. He had invited a crew from National Geographic to film at SpaceX’s headquarters to capture what would be a significant moment in the history of spaceflight.

If SpaceX could successfully land the first stage, it would be the first time anyone had done it—a huge leap forward for the company and the industry as a whole. And it would send a defiant message to the Alliance and Blue Origin. It was also Musk’s forty-third birthday—what better way to celebrate than to make history, while defying your critics?

The launch, too, was critical.

Seven months after Orbital Sciences’ rocket blew up on its way to deliver supplies to the space station, a Russian spacecraft, laden with thousands of pounds of cargo, supplies, and food, started spinning wildly, as if it were an out-of-control amusement park ride in zero gravity.

Another launch to the space station. Another failure.

Now, it was SpaceX’s turn.

After the back-to-back failures, the increased competition from the Alliance and Blue Origin, and the heightened expectations that came with winning the award to fly astronauts to the station, the pressure was on for SpaceX. Another disaster would undermine the Obama administration’s bold experiment to contract out missions to the space station so that NASA could pursue the more grandiose mission of flying to Mars. And it would fuel questions about a still-nascent industry where failure is common, expensive, and measured in mushroom-cloud plumes of smoke.

Then there was the question of what two failed missions had cost the space station. In the summer of 2015, NASA officials maintained that the astronauts aboard the orbiting lab were in no danger. But a NASA slide showed that with current food levels, the space station would reach what NASA calls “reserve level” by the end of July and run out by September 5.

AT T-MINUS 13 minutes, the launch director conducted the go/no-go poll of the thirteen members of the launch readiness team, a last check of all the systems and stations to make sure the rocket was ready. It was a call-and-response ritual passed down from NASA, and at SpaceX they did it now with ease and confidence.

“All stations verify ready for launch,” said the launch conductor into his headset, as he began the poll, surveying the teams in a call and response, each ending with an affirmative “Go!”

Propulsion was go. So was Avionics. It was a go from Guidance, Navigation and Control, and the chief engineer, go all the way down the line until the mission director said, “MD is go!” and the launch director said that the “LD is go to initiate terminal count.”

THE COUNTDOWN WENT smoothly, and soon: “We have liftoff of the Falcon 9. The Falcon 9 has cleared the tower.”

The employees gathered at SpaceX’s headquarters cheered wildly, as if at a football game, pumping their fists at another successful liftoff. Ready to party once that thing landed.

As the rocket climbed, it looked excellent—or “nominal,” which in rocketry means everything is going fine.

“Stage one propulsion is nominal,” the propulsion engineer said shortly after liftoff. “Power and telemetry nominal,” said the avionics engineer.

At T-plus 1 minute and 30 seconds, the Falcon 9 passed through what’s known as Max-Q, or maximum dynamic pressure, the moment when the rocket is under the most stress as it screams skyward. Still, the status was good: “Stage one propulsion is still nominal.”

At T-plus 2 minutes, the Falcon 9 was at an altitude just shy of 20 miles, racing to space at 0.6 mile per second. The fiery plume of smoke and fire behind it had expanded, which was normal given the reduced air pressure at that altitude. Everything was going smoothly.

Until it wasn’t.

Just over two minutes after liftoff, the rocket exploded, engulfed by a white, wispy cloud. After a little while, the smoke and debris dissipated, leaving only the pale blue sky. It was as if a magician had somehow made the rocket, and the 4,000 pounds of cargo it was carrying, disappear.

On the space station, NASA astronaut Scott Kelly tweeted that he had watched the launch from the station. “Sadly, failed Space is hard.”

The crowd at SpaceX headquarters went silent, some with their hands over their mouth. The National Geographic team kept the cameras rolling, capturing the devastating silence in what was supposed to be a birthday victory celebration but now felt more like a funeral.

A FEW WEEKS after the launch, SpaceX would identify the cause: a single faulty steel strut, 2 feet long and 1 inch wide at its thickest point. It was supposed to be able to withstand 10,000 pounds of pressure, but had buckled under 2,000, causing helium to overpressurize an oxygen tank in the second stage, which had led to the explosion.

In a call with reporters a month later, Musk was every bit the rocket scientist he had become, giving a preliminary but detailed analysis of the failure. But it was also as if he were delivering a business-school lecture on how a successful startup can retain its innovative culture and edge as it grows into a corporate behemoth.

He would point to another possible cause, saying that as the company continued to grow, it may have lost some of the inherent paranoia that fueled SpaceX in its early days, when it was unclear whether it would ever be able to launch rockets reliably.

After a string of successes, the explosion was the “first time we’ve had a failure in seven years” outside of the test flight, Musk said. “To some degree I think the company as a whole maybe became a little complacent.”

When the company lost a string of rockets in its early days, only a few hundred people were working at SpaceX. Now, there were four thousand. “The vast majority of people at the company today have only ever seen success,” he said. “You don’t fear failure as much.”

And so, when he sent out his wedding officiant e-mail before each launch, asking people to come forward, it didn’t “resonate with the same force” as it had when the company was small and scrappy and feared going out of business.

The reaction became “There’s Elon being paranoid again,” he said.

But now even the uninitiated knew the driving power of failure—and fear—“and we’ll be the stronger for it,” he said.