Liftoff

Hiring & Culture

What struck Bjelde most of all was the Coke machine in the break room. Musk had imported this innovation from Silicon Valley—unlimited free soda, to keep the workforce caffeinated at all hours. For someone from academia, and the sober environment at NASA, this was a novelty.

Suddenly, it clicked. Bjelde had not really been invited for a tour and as many Cokes as he could drink. This was a job interview.

“I think at the time it was probably eighteen degrees out,” Thompson said. “And he shows up in slacks and Neiman Marcus shoes and a skinny leather jacket.” But Musk asked good questions and listened intently.

Musk recognized what Thompson and Mueller were walking away from, so he put two years’ worth of salary for both engineers into an escrow account. That way, if Musk decided to prematurely pull the plug on the venture, they would still have a guaranteed income. This helped Thompson convince his wife he should take the job. His only regret? That he mulled it over long enough to get tagged as employee number two.

Because they were spending his money, Musk gave employees an incentive to be frugal with it.

Before he could log on to a computer at NASA, Bjelde had undergone a detailed security screening process and multiple orientations. To operate machines that would steer electron beams, Bjelde had sat through days of training courses. “At SpaceX back then, there was none of that,” Bjelde said of his first day on the job. “You show up. The door is not locked. There’s no one at the front desk. I met Hans, and he gave me a packet that had some materials about benefits and things like that. And then he told me what I needed to do.” Orientation done.

Musk kept a mostly laissez faire attitude toward his workplace. He offered just a few hard and fast rules: no strong smells, no flickering lights, and no loud noises in the cubicle farm they all shared. Often, they worked until well after midnight. Bjelde, slumped under his desk, recalls being kicked awake on more than one occasion to help finish writing a proposal.

One of Musk’s most valuable skills was his ability to determine whether someone would fit this mold. His people had to be brilliant. They had to be hardworking. And there could be no nonsense. “There are a ton of phonies out there, and not many who are the real deal,” Musk said of his approach to interviewing engineers. “I can usually tell within fifteen minutes, and I can for sure tell within a few days of working with them.” Musk made hiring a priority. He personally met with every single person the company hired through the first three thousand employees. It required late nights and weekends, but he felt it important to get the right people for his company.

Musk began to move on to his next question, but Kassouf cut him off. “Wait, there’s another place you could be.” Now Musk was interested. “If you’re north of the South Pole,” Kassouf continued, “there’s a place where the circumference of the Earth is exactly a mile. If you start a mile north of that, and go south one mile, go all the way around the Earth, and come back up a mile, you’re in the same spot.” That was true, Musk acknowledged.

Musk therefore came into his job interview with Altan prepared. About halfway through, Musk told Altan, “So I heard you don’t want to move to L.A., and one of the reasons is that your wife works for Google. Well, I just talked to Larry, and they’re going to transfer your wife down to L.A. So what are you going to do now?” To solve this problem, Musk had called his friend Larry Page, the cofounder of Google. Altan sat in stunned silence for a moment. But then he replied, given all of that, he supposed he would come work for SpaceX.

Kassouf’s friend had just a single job, finding a supplier for a bolt on the aircraft’s landing gear and ensuring that it met all quality specifications. That single bolt was the totality of his employment.

Koenigsmann didn’t have much experience with rockets. None, actually. But he’d done a lot of hands-on work with controlling the flight of small satellites in space. Surely, he thought, it couldn’t be that much different to do guidance, navigation, and control of a rocket. “I thought it would be quite natural,” he said. “It was part of the adventure. My wife really wanted to come over here. She just liked the idea of spending time in California.” They would never leave. They got their adventure. And Hans Koenigsmann has since learned an awful lot about rockets.

When Koenigsmann worked at Microcosm in the early 2000s, Koenigsmann made a number of good friends, perhaps none better than Gwynne Shotwell.

“Just come in and meet Elon,” he said. The impromptu meeting might have lasted ten minutes, but during that time Shotwell came away impressed by Musk’s knowledge of the aerospace business.

Dunn received a call saying he had not made the cut, but that he stood as the “first alternate” for one of ten internship slots. If one of those students backed out, a recruiter from Blue Origin informed Dunn, the company would call. “I told the recruiter on the phone, ‘Just let me come out and work for you for a month,’” Dunn said. “You don’t have to pay me anything, and if at the end of the month I don’t work harder than every other intern, I’ll just go home. It will be cool, and I will have appreciated the experience. But if I do come in and do an amazing job, and work harder than everyone else, then just pay me for the rest of the summer.” The recruiter rejected the suggestion. No one from the company ever called him again.

What Dunn savored most of all were the stories, the tales of life at SpaceX when there were only a few dozen employees, when everybody knew everybody, and there were Friday afternoon ice cream runs. Those times were gone. He again worried that he had missed out on the company’s formative years.

“I didn’t go to school just to go to meetings, and sit in a cubicle to try and perfect a single screw,” he said. “This was a company that wanted people to just get stuff done. I wanted to get my hands dirty, and no other company was really going to offer that besides SpaceX.” He got his hands dirty. On his very first day on the job, Altan designed a printed circuit board and sent it to manufacturing. He mused, at the time, that at most other companies he probably would not even have had an IT account set up by the end of the first day.

Zurbuchen made a list of his ten best students based on academics, leadership, and entrepreneurial performance during the previous decade, and researched where they had ended up. To his surprise, half of the students worked not for the industry’s leading companies, but at SpaceX. The results blew him away. “That was before SpaceX was successful,” said Zurbuchen, who in 2016 became the chief of science exploration at NASA. “So I interviewed these former students and asked, ‘Why did you go there?’ They went there because they believed. Many of them took pay cuts. But they believed in the mission.”

“But in the long run, talent wins over experience and an entrepreneurial culture over heritage.”

Musk’s Leadership

These kinds of trips, of which there were many, helped Musk bond with his senior leaders. He could be difficult to work for, certainly. But his early hires could immediately see the benefits of working for someone who wanted to get things done and often made decisions on the spot. When Musk decided that Spincraft could make good tanks for a fair price, that was it. No committees. No reports. Just, done.

During her first day at work, she set about formulating a strategy to sell the Falcon 1 rocket to both the U.S. government as well as small satellite customers. Seated in the cubicle farm at 1310 East Grand, Shotwell wrote a plan of action for sales. Musk took one look at it and told her that he did not care about plans. Just get on with the job. “I was like, oh, OK, this is refreshing. I don’t have to write up a damn plan,” Shotwell recalled. Here was her first real taste of Musk’s management style. Don’t talk about doing things, just do things.

“I make the spending decisions and the engineering decisions in one head,” he said. “Normally those are at least two people. There’s some engineering guy who’s trying to convince a finance guy that this money should be spent. But the finance guy doesn’t understand engineering, so he can’t tell if this is a good way to spend money or not. Whereas I’m making the engineering decisions and spending decisions. So I know, already, that my brain trusts itself.”

“He can’t stand a liar, and he hates a thief,” Reagan said. “And if you say you can do something, you’d better fucking do it.”

Musk laughed when told about Jeff Bezos’s timeline for engine development. “Bezos is not great at engineering, to be frank,” he said. “So the thing is, my ability to tell if someone is a good engineer or not is very good. And then I am very good at optimizing the engineering efficiency of a team. I’m generally supergood at engineering, personally. Most of the design decisions are mine, good or bad.” Boastful? Maybe. But SpaceX built and tested its first rocket engine in less than three years with Musk leading the way.

Musk has a rollicking wit. He will say something funny, realize it is funny, and iterate on the joke as a conversation proceeds.

At about T−30 minutes, Musk walked to Thompson’s console position and began a particularly heated conversation about why the materials had not yet been ordered. “We were right smack in the middle of a count, and he just wanted to have this deep, aggressive conversation about materials,” Thompson said. “I was absolutely dumbfounded that he was not even aware that we were trying to launch a rocket, and that I was the launch conductor, and responsible for basically calling out every single command that we’re going to run. It just blew me away.” After Musk walked off, Buzza turned to Thompson and asked, “What in the hell is going on?” In truth, this was just Musk being Musk, multitasking to the nth degree. Even in the middle of a critical countdown, he had the ability to simultaneously think about the company’s needs six months or a year into the future. The last thing on Thompson’s mind were shipping dates and aluminum costs. He had a rocket to launch. The company’s very first rocket, in fact.

“That’s the thing about Elon, he was willing to spend money to try things,” Kassouf said. “And that’s so different. Go to Boeing, and you spend money to try and figure out what your liabilities are going to be before you try anything. But Elon is like, sure, try it. If it doesn’t work we can either sell it back, or it goes into our lessons-learned pile.”

Then Musk had an idea. Perhaps, if they applied epoxy to the chambers, the sticky, glue-like material would seep into the cracks, and then cure, solving the problem. It was a Hail Mary. Mueller doubted the epoxy would stick to the ablative material, mixing about as well as oil and water. But sometimes Musk’s crazy ideas worked, and he was the boss, after all.

Musk had been wrong. But the filthy and exhausted engineers and technicians working with him all night did not begrudge Musk for keeping them at a task that proved fruitless. Rather, his willingness to jump into the fray, and get his hands dirty by their sides, won him admiration as a leader.

Musk would often visit his desk to ask detailed questions about his computer simulations for controlling the rocket in flight. And then they would make bets on some aspect of the rocket and its avionics system. Almost invariably, Musk would win. But ahead of one systems test in 2007, Davis said Musk raised the stakes. Davis bet twenty dollars he could complete some aspect of the test by a certain date. In return, Musk bet a frozen yogurt machine that Davis could not make the deadline. “The second we had a bet like that, where there was a chance of getting a yogurt machine, there was a zero percent chance I was not getting that done,” Davis said. “And if you go to SpaceX today in Hawthorne, you will see that he honored the bet, and we have a frozen yogurt machine sitting in the middle of the cafeteria, which still gives away free yogurt. So yeah, he’s very good at motivating his people.”

“I never missed that boat, ever,” Buzza said. “But sometimes, teammates would. The Army is very punctual, and would not deviate. Except for one time, they did come back to the dock for Elon.”

So Thompson swallowed his pride, and sent an email to Musk, asking for his full-time job back. And then he waited. For a day. A week. And then three weeks. There was no response from Musk, who typically made brisk replies by email. Thompson figured he had burned his bridges at SpaceX. But then, Musk called him. “Hey, I saw your note,” Musk said, as if three weeks had not slipped by in the interim. “Why don’t you just come in on Monday? You can have your old job back. See Jerry Fielder and he’ll give you a whole new salary.” Thompson recalls being stunned. He stammered something along the lines of “What???” before Musk hung up. The whole call lasted only a minute or two. When Thompson returned to SpaceX the next Monday and met with Fielder, the head of human resources, Thompson had received a raise and more options. Musk treated Thompson like nothing had ever happened.

Musk placed enormous demands on his employees. About half of those who survived the Kwaj crucible remain with SpaceX today, but the other half have moved on, often to escape the burden of toiling under Musk after it became too much.

Iterative Engineering

Musk differed from his competitors in another, important way—failure was an option. At most other aerospace companies, no employee wanted to make a mistake, lest it reflect badly on an annual performance review. Musk, by contrast, urged his team to move fast, build things, and break things. At some government labs and large aerospace firms, an engineer may devote a career to creating stacks of paperwork without ever touching hardware. The engineers designing the Falcon 1 rocket spent much of their time on the factory floor, testing ideas, rather than debating them. Talk less, do more.

The iterative approach begins with a goal and almost immediately leaps into concept designs, bench tests, and prototypes. The mantra with this approach is build and test early, find failures, and adapt.

“To adopt this method, you have to let people see you fail, and you have to push back when the critics use your early failures as an excuse to shut you down. This is why it is hard for national space agencies to adopt it. The geopolitics and domestic politics are brutal.” Failure was an option at SpaceX, partly because the boss often asked the impossible of his team.

“The bad news is that we had to change everything,” Mueller said. “The good news is that I learned everything that can go wrong with turbopumps, and really how to fix them.”

…the original pump from Barber-Nichols weighed 150 pounds, with an output of about 3,000 horsepower. Over the next fifteen years, SpaceX engineers continued to iterate, changing the design and upgrading its parts. The turbopump in the modern-day Falcon 9 rocket’s Merlin engine still weighs 150 pounds, but produces 12,000 horsepower.

For a given task, a typical aerospace company would just use whatever part had always been used before. This saved engineers from the time-consuming, difficult work of qualifying a new part for spaceflight. The SpaceX attitude was different. “True, a product may already exist,” Bjelde said. “But is it optimized for your solution? Is it from a good supplier? And what about their tier two or tier three suppliers? And if you need more of them faster, will they meet your needs? If you want to change something, are they going to be willing to change it? And if you improve that product, will they then sell it to your competitors?”

Despite these early setbacks, Koenigsmann had grown increasingly convinced SpaceX had taken the right approach to building a rocket. His experience at Microcosm in the 1990s provided a counterexample of how companies with less money and less urgency would fail.

Suppliers & Partners

Reagan had also taken a lot of orders from Boeing as it manufactured the Delta IV rocket, and other large aerospace companies. Reagan had never worked with a company that moved as quickly as SpaceX. He’d receive orders from them and, within a few days, ship parts machined out of aluminum or other materials.

The town of McGregor owned everything here and was willing to lease it all. Because local officials wanted the company as a tenant, there would be minimal interference, and no restrictions on the size of their engine. Texas, too, had a much less restrictive regulatory environment than California, with more business-friendly laws.

“Each of Kistler Aerospace’s contractors is a leader in its respective field of the aerospace industry and has significant experience in the construction of similar components,” the guide stated. Among the contractors were Lockheed Martin (LOX tanks), Northrop Grumman (structures), Aerojet (engines), Draper (avionics), and so on. After seeking to integrate all these high-priced components, it is small wonder that Kistler found its financial situation dire by 2003.

Reagan initially felt jilted by the reward Musk offered. “He gave me a ten-thousand-share bonus, and I was so pissed off because I thought that was nothing,” Reagan said, breaking into laughter. “I didn’t know the stock was going to go up to $212 a share. I guess he took care of me.”

Launch Site Battles

But her friends didn’t give up, and in early 2003 Chinnery joined SpaceX as a consultant. Soon, the time she had spent at Vandenberg as an officer a decade earlier was paying dividends for SpaceX. She knew people there and how the base operated. Chinnery hired on full-time later that year. For once, Musk’s instincts had led him astray. “Early on, Elon didn’t really understand the importance of interfacing with outside agencies, or how difficult it would be,” Chinnery explained. “There was this whole oversight the Air Force wanted over the design, development, and launch aspects of rockets. He really had no idea about that, but it was my specialty.”

The male-dominated workplace didn’t bother her. “At the time, aerospace was predominantly male, and I was very used to being in environments where I was the only female,” she said.

Air Force leadership at the base simply didn’t believe that SpaceX would succeed with its ambitious plans to build a rocket so quickly and launch it. They offered the minimum amount of support, and put their second- and third-string people on the job to look over paperwork and process approvals. But they were not obstructive. As Chinnery put it, “There was not a whole lot of scrutiny early on. They just never really believed in it until, suddenly, the static fire happened, and they woke up.”

…when SpaceX met every requirement for permission to launch its Falcon 1 rocket, checking every box, the paperwork seemed to disappear into a black hole. The Air Force simply did not sign off on the final documents. For the Air Force, it came down to a simple calculation: let the new space company fly its unproven rocket, or protect its hugely valuable national security assets from debris or other hazards should the Falcon 1 launch go awry. The decision was an easy one for the generals. The military officials would not permit a SpaceX launch until the Titan IV and its billion-dollar satellite for the N.R.O. lifted off. And they couldn’t give a firm launch date for that mission.

While the Falcon 1 waited its turn, no one would compensate SpaceX for its expenses. The company got paid when it launched. By contrast, when the military awarded a national security launch contract to an Atlas or a Delta rocket, Lockheed and Boeing signed cost-plus agreements, where any delays were billed to the government, plus a fee. “Technically, we weren’t kicked out of Vandenberg,” Musk said. “We were just put on ice. The Air Force never said no, but they never said yes. This went on for six months. The resources were draining out of the company. Effectively, it was just like being starved.”

The Air Force had not said no. If they had, SpaceX could have fought the decision. But there was nothing to protest. A lawsuit would not have brought an injunction against the military, and a favorable court ruling years later would have delivered a pointless, posthumous victory.

She marvels at the difference between the military’s attitude between then and now. To entice Firefly to launch from Vandenberg, Air Force officials have eagerly helped, and proactively found solutions to problems. “There is just no question that without SpaceX it would not have happened,” she said. “They convinced everybody else that commercial space is a real thing. And when they did that, the Department of Defense realized that they could either be a part of that, or they could be left behind.”

Life on Kwaj

At the time, she did not feel overwhelmed by the prospect of building a launch site in one of the most remote places on the planet. “I had kind of caught that weird SpaceX bug that says anything is possible,” she recalled. “I was actually thinking how cool it would be to launch from there. Kwaj really is gorgeous. I’ve never seen prettier water anywhere. I’ve never snorkeled anywhere better. It didn’t really occur to me what a challenge it would be to bring in everything that we needed.”

“That’s one thing SpaceX seemed to know how to do from the very beginning,” she explained. “We just didn’t waste any time dithering about stuff. If they knew they needed to ship stuff, they just shipped stuff.”

Just as the team on the tiny island ran low on industrial supplies, sometimes they went without food, too. On the same day as the documentation blowup, the boat designated to deliver food, beer, and cigarettes did not show up. “We had been going around the clock,” Hollman explained. “We were sick of being told to do this, or do that. At some point everybody got fed up and decided that we needed to find a way to let them know that we were a part of this team as well.” And so they went on strike.

The team on Omelek would work no longer without food and smokes, Hollman said. They had had enough. Buzza recognized the gravity of the situation, and he hastily arranged for an Army helicopter to deliver a few trays of chicken wings and some cigarettes to the island that night.

With a little food and a little nicotine, the Omelek mutiny had been stilled.

The few women on the island faced their own travails. In the early years, Anne Chinnery and Flo Li had little privacy and no running water. Using the island’s toilet necessitated filling a bucket with sea water first, so the toilet could be flushed. Showers were even more rudimentary. Initially, the SpaceXers filled a trash bucket with water to wash their hands. When she got really hot and sweaty, at the end of a long day. Li said she would put on a bathing suit and dump the rain water over her head to rinse off.

Business & Politics

From the beginning, Musk understood SpaceX could not become a sustainable, profitable business from government launch contracts alone. Although the prospect of a low-cost, launch-on-demand rocket appealed to the U.S. military, it had only so many spy and communications satellites ready to fly. To make money SpaceX needed to broaden its customer base to so-called “commercial” customers.

Musk invited Mango to dinner at a high-end Los Angeles restaurant. The meal would be well above the Army officer’s per diem, however. For the first time in his career, Mango had to call an Army lawyer with an ethics question. The Army officer was told that if he went to dinner, he would have to buy the entire, expensive meal. “I think we went to Apple-bee’s instead,” Mango said.

Musk wanted to make a splash in the nation’s capital, so his rocket was headed for Washington, D.C. It had almost arrived when, on the outskirts of the city, the tractor-trailer stopped as it crossed some railroad tracks. As it did this, the train signal’s red lights and bells activated, and the crossing arm banged down on top of the Falcon 1. The damage was minimal, and the driver pulled up just before the train came through.

dragooned

“When the government is hiring you to design, develop, build, and operate a thing, they’re the customer,” Shotwell said. “They’re paying for it. They get to have their hands in the design. The decisions. They’re covering the whole thing. But no one was paying us for design or development. They were paying us for flights.” This offered an advantage in that SpaceX could build the rocket that Musk and his engineers wanted to—but it came with a big downside. Unless Shotwell sold a multitude of launch contracts, the company would die.

“I was told by many people that we should not protest,” Musk said. “You’ve got a 90 percent chance that you’re going to lose. You’re going to make a potential customer angry. I’m like, it seems like ‘right’ is on our side here. It seems like this should go out for competition. And if we don’t fight this then I think we’re doomed, or our chances of success are dramatically lowered. NASA being one of the biggest customers of space launch would be cut off from us. I had to protest.”

SpaceX protested. More than that, the company won. After NASA learned that the U.S. Government Accountability Office would rule in favor of SpaceX on the issue of fairness, NASA pulled the award to Kistler. The space agency realized it would need to open up a new competition for cargo delivery. This became the foundation for NASA’s Commercial Orbital Transportation Services program, or COTS, that would emerge two years later and forever change SpaceX.

By 2005, the U.S. share of the global market for commercial launches, such as large satellites for television and other communications, fell to near zero. This left the two U.S. companies scrambling for Air Force contracts, alone, to make ends meet.

Almost from the beginning, then, SpaceX had to battle for its existence. One of the secrets of Musk and Shotwell’s success is they did not kowtow to the existing order of large companies and government agencies. If they had to sue the government, they would. To fight back Musk would use everything at his disposal. Within its first three years, SpaceX had sued three of its biggest rivals in the launch industry, gone against the Air Force with the proposed United Launch Alliance merger, and protested a NASA contract. Elon Musk was not walking on eggshells on the way to orbit. He was breaking a lot of eggs.

Because his price was so good, Musk wanted it front and center on the company’s website. This sort of transparency was pretty radical at the time. “It opened a curtain into a dark little corner,” said Chad Anderson, who runs an investment group, Space Angels, that closely tracks public and private investment in spaceflight. “Before this there were a handful of companies serving the government and commercial launch needs, and it was more of a cartel situation.” SpaceX changed expectations with its low prices and transparency.

These initial customers stuck with the company through its first launch failure. “I think the early customers not only needed us to be successful, they wanted us to be successful, so they were going to hang in,” Shotwell said. “Early customers don’t hire maverick companies if they don’t feel some kinship with the philosophy.”

United Launch Alliance enjoyed a monopoly on U.S. national security launch contracts, and a handful of large aerospace firms, including Boeing, Lockheed Martin, Aerojet Rocketdyne, Northrop Grumman, and ATK Aerospace carved up most of the rest of the launch business for other government contracts, including NASA. None of these companies welcomed a new competitor, especially one so potentially disruptive. In response, they started to fan the flames of political opposition. Just as these contractors had a vested interest in the status quo, so did politicians in Alabama, Florida, Texas, Utah, and a handful of other states with a disproportionate number of aerospace jobs.

SpaceX took off at a key moment in space policy history. In 2010 a battle raged between the White House and Congress over the future of human spaceflight. With its final mission set for mid-2011, everyone agreed the space shuttle would soon be retired. The big aerospace companies, which all held large contracts for the shuttle program, helped Congress devise a plan to continue on similarly lucrative contracts to build new government spacecraft and rockets. The Obama White House sought to limit funding for these expensive programs, and give emerging players like SpaceX a chance to see if they could bring down spaceflight costs. The first Falcon 9 launch, then, offered something of a referendum on President Barack Obama’s space policy.

But the reality is that a lot of NASA leaders also had ties to the legacy aerospace companies. Then, as now, senior officials at NASA often came and went between the agency and large contractors. This revolving-door effect helped the aerospace industry maintain some control over the direction taken by the agency, and reinforced skepticism of companies like SpaceX that sought to shake up the existing order.

Beal had encountered many of the same political and funding challenges that would bedevil SpaceX along its development curve. When Beal’s company went under in 2000, he cited several reasons, including an inability to secure a launch site and NASA’s favoritism toward traditional contractors. “There will never be a private launch industry as long as NASA and the U.S. government choose and subsidize launch systems,” Beal said in 2000, when he dissolved Beal Aerospace.

The Failures

Months before the Falcon 1 flight, Musk had told Jennifer Reingold, a reporter at Fast Company, that the Falcon 1 rocket had “well over” a 90 percent chance of success in its first launch.

A pleasant surprise greeted the SpaceX team at the dock the next morning. More than one hundred people, mostly civilians working for the Army on the island, had gathered there. Kwajalein island’s total population is only about one thousand people. These people weren’t bound for Meck Island, but rather had come to show their support for the small rocket company. They wanted to jump-start the investigation by helping to pick up pieces of the Falcon 1.

The loss of the inaugural Falcon 1 rocket dropped Koenigsmann into a depression. Later, his wife told him that he did not speak to her or anyone else for a month after returning to Los Angeles. He does not recall this, but said he must have just gone home in the evenings from work in a funk, lost in thought, trying to process everything. He was thinking about what had gone wrong, and what he could do to make sure it never happened again. She accepted that.

“Having experienced first-hand how hard it is to reach orbit, I have a lot of respect for those that persevered to produce the vehicles that are mainstays of space launch today,” he wrote. “SpaceX is in this for the long haul and, come hell or high water, we are going to make this work.”

After a few minutes, Hollman said, Gwynne Shotwell showed up and pulled him away from Musk’s desk. Mueller arrived to meet with Hollman. For a while, they talked through the situation. Eventually, Hollman said he was going home for the weekend to cool off. He would come back on Monday. When he did return, Hollman told Mueller he had one condition for remaining at SpaceX: he would never have to speak with Musk again.

In the end, Hollman and Thomas were not, in fact, to blame. The Flight One failure was caused by the tropical environment, not dereliction by Hollman and Thomas.

When D.A.R.P.A. published its review of the Falcon 1 flight a few months later, the defense agency concluded, “The only plausible cause of the fire was the failure of an aluminum B-nut on the fuel pump inlet pressure transducer due to inter-granular corrosion cracking.” The nut, which cost all of $5, had cracked due to corrosion from sea salt spray on Omelek the night before the first launch.

When preparing the Falcon 1 for launch on Omelek, an engineer had opened a valve during the tanking process that allowed the second-stage LOX tank to vent more easily. This valve was never closed. Had the Falcon 1 rocket’s first stage performed a nominal ascent, the second stage would not have remained sufficiently pressurized to boost FalconSAT-2 into orbit. This all came out later in the data review. Musk asked his team why a computer had not checked for something like valve closures. The answer was simple: they did not have time to install sensors like this. SpaceX’s first major failure, therefore, taught Musk there might be some limit to how fast a launch company should go.

Gradually, throughout the rest of 2006, SpaceX began adopting more traditional aerospace practices.

“In terms of our maturity and discipline, we were a completely different company coming out for Flight Two,” Anne Chinnery said. “That failure helped us.”

Despite this end result, with Flight Two SpaceX had come very close. Much closer than Flight One. Buzza said the control room on Kwaj felt pretty good about the launch. They had not achieved 100 percent mission success, but maybe 95 percent. They’d survived such a hard year since the first failure, and now they had gone through the entire first-stage burn, stage separation, second-stage ignition, and fairing separation. Less than half a decade after its founding, the company had climbed above Earth’s atmosphere. SpaceX’s next step was clear—orbit. “We had always talked about needing three flights to get to orbit,” Buzza said. “We actually kind of celebrated that night.”

The problem they had known about, discussed in detail, and ultimately dismissed as the eleventh highest avionics risk took down their rocket. “Now,” Musk says, “I ask for the eleven top risks. Always go to eleven.” It’s true. The company now does lists of the top eleven risks ahead of launches. That is one legacy of the second failure of the Falcon 1 rocket.

Following the first launch failure, Musk sought to buoy the spirits of his team by offering them a brief escape. At a cost of more than $100,000, he booked a private Zero-G flight to give employees a taste of spaceflight. Many had come to SpaceX with, if not the outright dream of becoming an astronaut, at least a hope that one day they might ride on top of one of the company’s rockets.

“It was like everybody who got a good grade got to fly on the Zero-G flight,” Mueller said. “Well, I didn’t get to fly in the Zero-G flight.” Mueller, one of SpaceX’s best engineers, had been left off the invite list.

To Mueller’s eyes, from the video feed, it appeared as though the stage separation system must have malfunctioned, causing the collision. In the heat of the moment, he expressed this opinion to Thompson. The structures engineer, who bore responsibility for this part of the rocket, reacted defensively to these recriminations, believing Mueller’s conclusion premature. “That’s bullshit,” Thompson replied. “You’ve got to look at the data before you make that accusation.” Steve Davis was already looking at the data, and after an all-nighter, he figured out what happened first. As he pored over the video, frame by frame, Davis saw the actuators working. He confirmed there had been a full separation of the first and second stage. At another console he collected printouts of data sent back by the flight computer. And he noticed one curious bit of data, a non-zero acceleration of the first stage after separation. This, he realized, vindicated Thompson. Instead, Davis deduced the problem must have been caused by Mueller’s new, regeneratively cooled engine.

At full power, the pressure inside the Merlin chamber reaches about 1,400 pounds per square inch. The transient thrust produced just after main-engine cutoff in Flight Three, by comparison, briefly produced a chamber pressure of 10 psi. This is less pressure than air exerts at sea level. So during the many engine tests in Texas, SpaceX had missed this transient thrust. “You don’t really even see it on the test stand, because the ambient air pressure is like 15 psi, and the rocket chamber pressure dropped to about 10 psi,” Musk explained. “Later, when we looked back at it, if you looked super carefully you could see a tiny, tiny trail of thrust. But a rocket engine producing 10 psi in a 15 psi ambient atmosphere, basically you don’t notice it. You don’t see it in the data.” In the vacuum of space, with rocket hardware so close, even a miniscule thrust is enough to force a catastrophic collision between stages. The solution was to change a single number in the flight software. For Flight Four, all SpaceX had to do was add four seconds to the time in between main engine cutoff and stage separation. But this assumed there would be a Flight Four.

Flight Four

“The crazy thing is that I originally budgeted for three attempts,” Musk said. “And frankly, I thought that if we couldn’t get this thing to orbit in three failures, we deserved to die. That was my going-in proposition.”

Anger and frustration, he knew, would not serve any purpose. It could cloud one’s mind during intense moments, whereas creativity or quick thinking could sometimes spark a solution. On Omelek, there was no help for the cracked skirt. But a few months later, this sanguinity would save the day for Dunn and SpaceX.

So in this dark hour, Musk chose not to play the blame game. Certainly, he could dish out brutally honest feedback, crushing feelings without regard. Instead he rallied the team with an inspiring speech. As bad as Flight Three had gone, he wanted to give his people one final swing. Outside that room, in the factory, they had the parts for a final Falcon 1 rocket. Build it, he said. And then fly it. What they did not have was much time. “He surprised me,” Koenigsmann said. “He collected everyone in the room and said we have another rocket, get your shit together, and go back to the island and launch it in six weeks.” After Musk’s staff meeting, his employees realized they were playing for everything. If their final rocket launched safely into orbit, the company would have a chance to survive.

The period that followed would be the most memorable and arguably important period of the company’s history, hardening its DNA and setting the stage for SpaceX to become the most transformative aerospace company in the world.

Because the Malaysian government did not want to risk its satellite on an unproven rocket, Musk decided to fly his remaining Falcon 1, then scattered in pieces in Hawthorne, as a demonstration mission.

Over the course of twenty-eight days, the cargo ship would offload other containers at ports in Hawaii and Guam before finally docking in Kwajalein. From there, another boat would deliver the rocket to Omelek. But SpaceX did not have a month to wait for a circuitous barge trip. It would have to send the rocket using an oversized aircraft.

“There wasn’t a time in that month that one of us was not working on the rocket, and the vast majority of the time we both were,” Dunn said. “It was always intense in those days, but this set the bar.”

“The Air Force people had been scrambled,” Bjelde said of the flight crew. “They were all shaking their heads, saying this never happens. I don’t know who the person was that said OK to the flight, but they may have saved SpaceX.”

“Hey, the rocket is crumpling, we have to go back up again,” he shouted to the pilots. Here, the pilots had a decision to make. They had a $200 million aircraft and two dozen lives to worry about. They were thinking it would be safer to simply open the plane’s large rear door, and jettison the unstable rocket into the ocean below. And in fact, had no one from SpaceX been on board the aircraft, they would have done just that. But instead, they followed Altan’s instruction. One of them replied “OK, boss,” and the C-17 immediately began to climb. Then a pilot told Altan, “By the way, we only have thirty minutes of fuel.” This gave the C-17 time for one loop around Hickam Air Force Base before they lined up for a landing. Effectively, this meant the SpaceXers had about ten minutes before the aircraft would restart its descent.

No one had anticipated needing to open up the rocket in flight, so none of the SpaceX employees had brought any tools along for the ride beyond their knives. After a frantic search for something to work with, the loadmaster produced the C-17’s meager tool chest, which contained a flat-head screwdriver and single crescent wrench. This, at least, allowed the technicians to open a couple of small lines. But to really equalize the pressure inside the rocket with that of the cargo bay, someone needed to open a large pressurization line leading into the liquid oxygen tank, which could only be accessed by climbing into the rocket’s interstage.

Before he entered, Dunn turned to his friend standing by his side, Mike Sheehan. If the rocket starts to blow, pull me out, he earnestly told his friend. To reach the pressurization port leading into the LOX tank, Dunn had to crawl all the way into the interstage. Darkness enveloped him as he moved deeper inside, along the wall. Only Sheehan’s hands, holding on to his ankles, tethered Dunn to any semblance of safety. As he went, sharp components lining the exterior structure scraped his back. And all the while, the tank continued to pop and ping ominously.

The rocket hissed as it repressurized, and just in time, as the ten minutes allotted for dealing with the rocket had passed.

”We all thought we were done,” Chinnery said. “The tank had imploded. We were devastated.”

Due to the hastily improvised nature of the transport flight, the SpaceXers lacked sleeping arrangements at the base near Pearl Harbor. They had no ride to a hotel, or even a hotel to stay at. The military side of the airport lacked accommodations, so they crashed wherever they could. Some slept in chairs. Chinnery and a few others settled into a child’s playscape, built near the main lobby of the airport.

“Elon saw that and went off the frickin’ deep end,” Thompson said. Six weeks was too long. SpaceX didn’t have six weeks. Realistically, SpaceX did not even have a month before its funding ran out.

“You need to stop talking, and shut up, and listen to what I’m about to tell you,” Thompson said. “You’re not bringing that fucking rocket back. You’re going to strip that fucking thing like a Chevy. And that rocket better be fucking disassembled by the time Buzza and I get there Monday morning.” There was dead silence in the trailer on Omelek as the words sunk in. They were going to have to fix the rocket right there, in the tropics. There was no time for quality control or meticulous records. They did not have six weeks. They had one. They were going to have to haul ass and hope for the best.

“We knew full well that if anything failed, it was game over,” Thompson said. “Believe me, it was a ballsy move. But I mean, that’s the state that we were in. It was like, we have to make this work. There’s no six weeks. This had to be done in days.”

“Of all the crazy things that we did over the years, and of all the amazing accomplishments in a short period of time, that one really stands out,” Chinnery said. “I can’t believe we disassembled an entire stage and reassembled an entire stage in the course of a week. I don’t think I could have imagined that.”

Musk had not traveled to Kwajalein for the launch, opting to remain behind in California during the summer and early fall of 2008. He needed to be on hand for both SpaceX and Tesla as they struggled for survival, simultaneously managing operations and raising funds for rockets and electric cars.

In her time zone, the launch window opened at midnight. Shotwell stayed up late that night after her husband, Robert, an engineer at NASA’s Jet Propulsion Laboratory, went to bed in the hotel room. To not disturb him, Shotwell camped out in the bathroom, sitting on the toilet, and balancing her laptop on her knees. To hide the noise, she turned on the shower.

Shortly before the rocket entered the terminal countdown phase, at about ten minutes before T−0, Buzza shared a few final thoughts with his team. They all knew the stakes, he said. Focus on your last-minute tasks. Tighten your seat belts. Then he told the team they reminded him of the early NASA flight controllers who had guided humans safely to and from the Moon. They were also mostly in their twenties inside Mission Control during the 1960s, too.

If the worst happened, Mueller knew he probably would share the blame. “You know, it’s usually propulsion that fucks up. Like it’s known that 40 percent of all rocket failures are due to propulsion. Almost half. So as soon as separation happened, we thought we had it, right?”

Like during Flight Three, SpaceX had urged employees to invite their families to the factory for the launch. As some restaurants do with a kid’s menu and crayons, the company passed out a one-sheet handout for children. They could do a word search for terms like Falcon and Kwajalein, play tic-tac-toe, or color in a mission patch. This patch, for the first time, included two green four-leaf clovers. For every launch since—remember, rocket scientists are a superstitious bunch—the mission patch design has included at least one four-leaf clover.

The SpaceX veterans had developed a simple launch tradition. If they succeeded, they were going to drink. And if they failed, they were going to drink. Until Sunday, September 28, 2008, they had never drunk to a successful launch.

After the rocket reached orbit, Shotwell ran out of her hotel room and down the hallway to find the rest of her Falcon 1 colleagues staying there for the conference. Dressed in a pajama top and yoga pants, she screamed and hollered as she went.

“I was supposed to be briefing the sad story of the Falcon 1, Flight Three to these unhappy customers,” she recalled. “I said, ‘Fuck it, I’m going to talk about Flight Four.’

Just as Musk found the lucky charm of taking his children to Space Mountain, Shotwell, too, developed a space superstition after Flight Four, what she calls her “launch juju.” On subsequent launch days since that first, sweet success, she has plastered the inside of her shoes, almost invariably high heels, with yellow sticky notes upon which she has written Scotland. “This way I am in Scotland for every launch,” she said.

Somehow, in the darkest of moments, in the most distant of tropical outposts, with the final chance before them, they had pulled together. They all knew that, very easily, they could have been drinking away their sorrows that night, making final goodbyes before they dispersed to other rocket companies, into academia, or elsewhere. Instead, they toasted to their shared experiences and a bright future. “That, maybe more than anything else, is what I’ve loved about SpaceX,” Dunn said. “Knowing that all those folks around you and next to you, in the control room or wherever, they went through it, too. They lived through being pushed as hard as they possibly could, or further, and gave it their damnedest.”

With no privately owned automobiles on Kwajalein, civilians cannot drink and drive. However, military police nonetheless patrol the island in golf carts, giving tickets to inebriated bike riders. Very late that night the SpaceX engineers sent two older merrymakers off from Vet’s Hall, wobbling on bicycles down the road, away from the lagoon. These decoys—Salty Dog and Space Mom—had not been drinking. The rest of the SpaceX team waited for the police to take the bait and then, as quietly as they could, giggling in hushed tones, rolled down to the lagoon.

“There was no jubilation or anything. I was just too stressed. It’s like the patient survived. Getting to orbit was just like, OK, we’re not going to die now. At least we’ll live a little bit longer. That’s what that launch meant. I just felt relief.” His employees did not fully realize the desperation of the situation, and Musk did not want to ruin their moment.

“I didn’t know what I didn’t know,” Dunn said of taking on this responsibility for the Falcon 1 propulsion system. “If something like that happened today, I would be super apprehensive. I would be very thoughtful about it. But at that time I was like, ‘Let’s just go get after it.’ I just wanted to do the best I could. The way I approached it was to work my ass off.”

Financial Brink

After watching the demoralizing failure of its first three rockets, no potential clients were calling Gwynne Shotwell. In fact, the company had just a single customer left for the Falcon 1: the Malaysians.

“I looked out six or eight weeks in advance, and knew we weren’t going to have enough money to make payroll.”

They were excited to tell Musk that SpaceX had won one of two contracts. Musk couldn’t believe it. He told them, “I love NASA. You guys rock.” After the call, Musk asked Shotwell to immediately sign whatever deal NASA offered. He harbored a niggling fear the space agency might take the contract back.

Musk called a meeting of the Falcon 1 team, and told them, without preamble, that the booster had flown for the last time. “That was tough for a lot of us that had worked the Falcon 1 missions,” Chinnery said. “We spent so much effort and time on making that program successful, and to move on like that was classic Elon. He was very focused on what he wanted, and Falcon 1 wasn’t in the plan beyond just learning how to do it.” Once they got beyond their initial shock, however, the Falcon 1 team accepted the wisdom of Musk’s decision. It meant less work, as they would not have to spend time developing, testing, and building the Falcon 1e. With the extra hours, they could focus on Falcon 9 and Dragon, which now represented the future.

Personal Sacrifices

Musk recognized the extraordinary demands he placed on SpaceX’s early hires. He therefore decided to reward employees who spent the majority of 2004 traveling to Texas for engine tests, and elsewhere. Anyone who spent two hundred days away from home in 2004 received an extra two weeks of time off in 2005, and an all-expenses-paid vacation wherever he or she desired to go.

When one-year-old Abby saw Hollman coming up to the door, she would react by saying, “Jeremy bad.” It made for painful goodbyes. “For years my younger daughter did not like Jeremy Hollman,” Buzza said. “Because every time she saw Jeremy I disappeared for ten days.”

He would call them, talk for a minute, and then ask his girls to find the book he had bought. Exhausted, sometimes he would wake up in the morning with the book on his face. The next night, one of his daughters would say, “Dad, you fell asleep again.”

So I told Elon I wanted to start to step down. Gwynne was there at the meeting, and Gwynne was horrified.” Through her sales experience, Shotwell understood the importance of Mueller to the SpaceX brand and, more important, its home-grown rocket engines. She and Musk persuaded Mueller to stay on for three more flights of the Falcon 9 rocket with the new Merlin 1D engine, figuring this would convince satellite operators that the supposedly new-and-improved propulsion system had actually improved. Six months later, Mueller went back to Musk and asked again. Musk realized his chief of propulsion was serious. So they came up with a plan where Mueller would get a new title, chief technology officer. “It was bullshit,” Mueller said. “But it was a great title, so it didn’t look like I was taking a step down.” As his workload eased, Mueller’s physical health improved. He had planned to have surgery on a pinched nerve in his neck, due to the stress. But after he stepped down, he canceled the surgery.

Despite the sacrifices made in his personal life, however, Mueller harbors few regrets about his time at SpaceX. Since Mueller departed the company, the Merlins have been enhanced by Musk and the propulsion team he left behind. But the fundamental design remains the same.

“I’m extremely, extremely proud of what we’ve done,” he said. “The Merlin 1D is such an awesome engine. I’m super proud of it. I can’t take much credit for Raptor. I designed the original Raptor, but it has changed so much that I can’t take much credit. I named her, and I hired and trained the team that developed Raptor. That’s the credit I’ll take. But the Merlin 1D was my baby.”

Engineers who come to SpaceX generally understand they are going to be used up in the job. The work can be all consuming. But what critics of these work schedules fail to grasp is that most new SpaceX hires willingly sign on to this bargain. They want that golden ticket for the world’s greatest thrill ride.

Dunn would do it all again, in a heartbeat. “Certainly, I put a ton in,” he said. “And I put in these, like, awesome super-productive years of my life. But it was where I wanted to spend it. I gave everything I had, to the exclusion of girlfriends or anything. I freaking gave it everything, and I wanted to give it everything. I don’t think it was a toll, it was a trade.”

“It would have been a lot easier for me to stay at JPL, or work someplace like that,” he said. “It was a toll and a sacrifice, but I wouldn’t change it for the world.”

Legacy

“It hardened the DNA of the company,” he said. “It’s still there in decisions we make today. I’ll be in a meeting, in our executive conference room, and we’ve got a picture of Omelek Island, and the Falcon 1 rocket. It recalibrates your mental model. Many of those same old-timers are still here, now in leadership positions. It created the sieve or filter by which we make decisions today. We talk about it all the time. We’re always trying to be more efficient, and more like we were in some ways.”

Shotwell herself did not believe the Mars stuff in the beginning. “I kind of ignored it,” she said. “I wasn’t even bought in.” She is today.