8

A Four-Leaf Clover

ON JULY 29, 2003, the Defense Advanced Research Projects Agency (DARPA) issued an announcement that it was looking for a “transformational capability.” Which was not unusual since the mysterious Pentagon agency was always looking for transformational capabilities. But even by DARPA standards, this solicitation—for a space weapon—stood out.

The Pentagon wanted to develop “a means of delivering a substantial payload from within the continental United States (CONUS) to anywhere on Earth in less than two hours,” the announcement read.

This was no benign delivery service, however. The “substantial payload” would be a deadly arsenal—missiles, bombs, and a mysterious spacecraft capable of traveling at hypersonic speeds, or at least Mach 5, five times the speed of sound, some 3,800 mph. Launched from the East Coast, the strike would be able to hit Baghdad in just over an hour and a half.

Like many Pentagon programs, this one was given a clunky acronym, FALCON—which stood for “Force Application and Launch from CONUS.” And it was born from necessity, or at least a wish list coming out of the top reaches of the Pentagon.

Although United States forces were able to hit Afghanistan with a punishing wave of ordnance in the wake of the September 11, 2001, terrorist attacks, and then light up Baghdad in 2003 with a devastating “shock and awe campaign,” those attacks required a substantial buildup of forces in the region, which consumed precious time in the heat of war.

“While advancements in target identification and precision strike have been abundantly demonstrated, deficiencies in engaging and defeating time-critical and high value, hard and deeply buried targets (HDBT) have also been revealed,” the FALCON solicitation read.

In other words, the Pentagon needed to act fast if it got word that someone such as Osama bin Laden was hunkered down in a bunker somewhere.

Now the Pentagon was looking for a way to strike targets thousands of miles away without having to rely on forward operating bases or aircraft carriers. The way to do that was to go into space.

“It was clear to leadership at the Pentagon that we had no quick way to reach out and touch somebody, so to speak—a Saddam Hussein or somebody that we needed to take care of quickly,” said Steven Walker, then FALCON’s program manager. “If you didn’t have bases already prepositioned close to where action needed to occur, there’s no quick way for the US to respond.”

Developing a system that could be up and running quickly and hit anywhere in the world within a couple of hours was just the sort of near-impossible challenge that DARPA took on all the time. It was, as they liked to say at the agency, “DARPA-hard.” Since its creation in 1958, when Bezos’s grandfather, Lawrence Gise, was hired, DARPA had come a long way. Although Gise was worried that the agency would go “up in blue smoke” because of the threatening political pressure, it instead solidified itself as an indispensable, if mysterious, part of the defense establishment. With a relatively small budget, it advanced all sorts of military technology, with the goal of staying a step or two ahead of the enemy and preventing another Sputnik-like surprise.

DARPA was tasked with looking into the future to envision what sorts of technologies the United States would need for the future of war: “To cast a javelin into the infinite spaces of the future” was its motto, a quote from Hungarian composer Franz Liszt. Walled off from the rest of the giant Pentagon bureaucracy so that it could innovate freely, the agency strove for nothing short of revolutionary advancement and “engineering alchemy” that would pierce the realm of science fiction. It had been given the authority to hire as it needed, as it sought “extraordinary individuals who are at the top of their fields and are hungry for the opportunity to push the limits of their disciplines.”

During Gise’s time, DARPA, then known as ARPA, was focused on preventing nuclear war and winning the space race. It even helped develop NASA’s Saturn V rocket, which took the Apollo astronauts to the moon. Since then, its reach and influence had broadened. In the late 1960s it started work on what would become ARPANET (Advanced Research Projects Agency Network), a network of computers in different geographic locations that became a precursor to the Internet.

Over the years, it helped develop all sorts of technological advancements that have transformed war, and, in some cases, everyday life. DARPA helped give birth to the Global Positioning System (GPS), stealth technology, cloud computing, early versions of artificial intelligence, and autonomous aerial vehicles. As early as the late 1970s, it was working on a “surrogate travel system” that created something like a Google Street View map of Aspen, Colorado. More recently, its work was focused on underwater drones, geckolike gloves designed to enable soldiers to climb walls, humanoid robots, bullets that can change direction, and a blood-cleansing “artificial spleen” to help treat sepsis.

FALCON, if it could be achieved, would join the pantheon of “breakthrough technologies” fostered by the agency. The program came in two parts. The first was to develop the hypersonic vehicle that eventually would be able to take off and land from a military runway and then hit anywhere in the world within a couple of hours.

In the near term, though, the hypersonic spacecraft would be launched from a rocket. And that presented a problem, as the FALCON solicitation pointed out: “Existing booster systems are costly and in limited supply.”

So, the second part of the program sought a new kind of rocket, one that was inexpensive and could launch quickly—“after authorization from an alert status within 24 hours.” It would have to be able to carry not only the hypersonic weapon, but also small satellites that could be used for spying. And it would have to be cheap—less than $5 million a launch.

ONE OF THE proposals that DARPA, and its partners in the US Air Force, received was intriguing, not just because the name of the rocket, Falcon, mimicked the name of the program, but because it was already well under development. Walker, the program manager, had never heard of the company SpaceX, or its founder, Elon Musk. But once he saw Musk’s plan to develop a rocket that could launch extremely inexpensively, he wanted to know more about this Internet tycoon turned space entrepreneur.

“We were interested in bringing him into the program because his development was obviously in the early stages, but it looked like he was going down a path that would enable affordable launch,” recalled Walker, who would go on to be the acting director of the agency. “Our goal was five million dollars. He was saying six million, which was much lower than anyone was doing it.”

When Walker visited SpaceX’s headquarters, he liked what he saw: a dedicated and passionate band of rocketeers that “had basically developed the first American-made engine in a long time,” he said. “I was impressed with what they had done technically, and with Elon’s business model and plan. I was really impressed with his ability to bring in key people on his team and give them the resources early on to go make this thing happen.”

He was also struck by how they were building the rocket almost entirely on their own, without subcontractors. “Elon cared so much about the quality of rocket,” that he’d pay for it to be made in-house, Walker said.

Walker was convinced. In 2004, DARPA agreed to invest a few million dollars into SpaceX, helping fund its first launch attempt. After years of relying on Musk’s fortune, the company finally had some outside backing, if a minimal amount.

Now all it had to do was fly.

THE LAUNCH WAS supposed to happen at Vandenberg Air Force Base. SpaceX had spent $7 million on a pad there, adapting it for its Falcon 1 rocket. But in 2005, some of the traditional contractors, including Lockheed Martin, complained about SpaceX’s presence at Vandenberg. Lockheed was worried that if SpaceX’s rocket exploded, it could damage the facilities at a time when Lockheed was working to launch sensitive military satellites worth millions.

The air force had promised to help SpaceX find another suitable accommodation, but it became clear to Musk that the offer was mere lip service. The US military had a long cozy relationship with Lockheed and was comfortable with the defense contractor’s Atlas V rocket. SpaceX was new to the neighborhood, an outsider with an unproven rocket built in what was essentially a garage in El Segundo.

Instead of staying on the range, SpaceX moved to an outpost in the Marshall Islands in the middle of the Pacific Ocean, thousands of miles away. DARPA helped with the transition, but Musk was fuming once again at the large contractors—and the air force, which he felt was screwing him.

“It’s like you build your house.… somebody else builds a house next to you and tells you to get out of your house,” he said at the time. “Like, what the hell… after we’ve just made that big investment and everything. We’re going to fight that issue because it is just fundamentally unfair.”

Just because the Atlas V was needed for national security launches, “that doesn’t mean they can completely shaft us,” he said.

The distance wasn’t the only problem with the site in the Marshall Islands. Musk worried about the conditions wreaking havoc on his new rocket. “I don’t think there’s a place in the world with more corrosion,” he said. “It’s the perfect environment of right temperature, humidity, and salt spray.”

There were, however, some upsides. While the range, known as the Ronald Reagan Ballistic Missile Test Site, was a government facility, SpaceX and DARPA had almost free rein with limited interference. Surrounded by turquoise water, sandy beaches, and palm trees on an island known as Kwajalein Atoll, or Kwaj, the setting felt like an island vacation spot.

“It was fun; we had the run of the place,” Walker said. It was a location where the young company “could learn and practice” the precise art of the launch.

The teams slept in army-style barracks, often two to a room. There was outstanding snorkeling. Fishing boats were available to rent for ten dollars an hour. And there was lots of grilling out on the beach at sunset. “They say the only things to do on Kwaj are: work, sleep, work out, fish, drink, and screw,” said an air force official who was there. “One friend who worked there for two years told me that he knew every square inch of every available woman on the island under sixty.”

Musk’s team worked to get the site ready as if they were on a mission. With little else to do on Kwaj, they kept extraordinarily long hours that both impressed and worried the small band of government officials along with them, who wondered whether the staff would burn out before the rocket fired.

“That was a big question in my mind—how could they keep that pace?” said Dave Weeks, a NASA official on loan to DARPA to help oversee the launch.

There was a cordial but uneasy relationship with Musk and the government team, especially at first. Controlling and detail-oriented, Musk wasn’t particularly interested in their insights or suggestions. It was his rocket—not theirs—and he was suspicious of outsiders.

Weeks could understand why. Weeks had met Musk in 2002 or early 2003, when he and Gwynne Shotwell, who would later become the president of SpaceX, visited NASA’s Marshall Space Flight Center. They weren’t long into their meeting when a fire alarm sounded and they had to evacuate the building. Outside, a security guard approached as they waited to go back in, and said that he’d been informed that Musk, a green card holder born in South Africa, hadn’t gone through the required background check the center required of foreign nationals.

The spaceflight center sat on the US Army’s Redstone Arsenal, and, escorted by security, they went to the officers’ club there.

“I thought this was not a great start,” Weeks recalled. Musk was polite and gracious, but “miffed a bit,” said Weeks.

Essentially, Musk had been kicked off the Marshall Space Flight Center. Then, he was kicked off Vandenberg Air Force Base. And here was the government, working with them side by side on Kwaj. Musk remained polite and calm. But he kept a distance. To the greatest extent possible, SpaceX was going to do this on its own.

He “wasn’t high on government help,” Weeks said.

ON MARCH 24, 2006, just four years after Musk had founded SpaceX, the seven-story-tall Falcon 1 rocket stood on the pad on Omelek Island, a couple miles north of Kwaj, ready at long last to launch. The flight had been delayed repeatedly over a period of months, and frustration was mounting, as were doubts about whether the rocket would ever get off the ground. But Musk was determined. He had fought so hard to get here, invested so much of his money. He had battled with the government, the entrenched contractors, and had pulled off stunts to get noticed by NASA, hoping they’d see that his company was for real.

Nothing would prove that, and quiet his critics, like a successful launch. Now there was something else on the line. In early 2006, NASA announced that it was starting a program to help commercial companies like SpaceX build rockets and spacecraft that would eventually be able to fly cargo and supplies to the International Space Station.

With the space shuttle set to retire in the coming years, NASA, under the administration of George W. Bush, was making a bold bet—that the private sector would be able to provide a taxi-like delivery service to the station, an orbiting laboratory, some 250 miles high. If the commercial sector could take over the relatively routine operations in what is known as low Earth orbit, that would allow NASA to do the hard stuff, to explore in deep space.

Michael Griffin, the NASA administrator at the time, believed that by investing in a few companies to help them develop their rockets and spacecraft, NASA could touch off an industry of viable commercial rocket companies. NASA, then, would become a customer, buying a service to launch supplies to the space station.

“Commercial enterprises historically, if they can succeed at all, they’re cheaper than government enterprises, more dynamic,” he said.

Griffin wanted just to provide “seed money,” and was concerned that if “you have a lot of government money into the enterprise, then it becomes a government enterprise. And that’s what I wanted to avoid.”

He carved out a $500 million proposal in his budget, called the Commercial Orbital Transportation Services program, or COTS, that would be divided by two or three companies. How did he come up with that amount? “Truthfully, I just made it up,” he said. It was enough to be significant for a startup. But in the space business, a couple hundred million dollars can burn up quickly, so the companies would have to come up with funding of their own.

For SpaceX, winning a spot in the COTS program wouldn’t just be a much needed stamp of endorsement from NASA, but the investment would provide a measure of stability. SpaceX had to be considered a long shot. First, it had to convince NASA officials that the business was going to be around in a few years. NASA didn’t want to invest in a company that would evaporate overnight.

NASA wanted to know “who would our potential investors be, how much money did Elon have to invest?” Shotwell recalled.

Then there was another question about the capabilities of a company that had never launched anything, yet was talking about one day going to Mars. “Can this very young, brash company do what they say they were going to do?” she said.

While it was preparing to launch the Falcon 1 from Kwaj, SpaceX crews in California were working on the next rocket, the much more powerful Falcon 9, which could be used in the COTS program.

Virtually everyone who wasn’t at Kwaj was assigned to the Falcon 9 program. It was all hands on deck. Starting as soon as it received the request for proposals, the company was “largely shutting down and everybody working on this particular effort,” Shotwell recalled. “At the time, it was that critical to the company and to our future.”

The competition put even more pressure on the crews in Kwaj, who had delayed the launch again and again through 2005 for various technical issues. But now they were finally ready to go.

To launch a rocket successfully on the very first try would be an enormous coup. The history of rocket development is filled with blooper reels of rockets blowing up every way possible, on the pad, just above the pad, or going way off course, spinning wildly like a balloon losing air until they come crashing down in a fireball.

“A million things have to go right in order to have a successful launch, literally, and only one thing has to go wrong to have a really particularly bad day,” Shotwell would later say.

Leading up to the launch, Musk acknowledged the possibility of failure, telling a reporter that the company could withstand a failed launch or two. “If we have three consecutive failures… it’s not clear to me that we know what we’re doing and maybe we should go out of business.”

Sitting in the control room, Walker, the DARPA program manager, had another goal for the launch: “I was hoping no one was going to get hurt.”

THE MOMENT HAD been four years in the making, and it all came crashing down in less than one minute. The count from T-minus 10 seconds seemed to proceed smoothly. The engine ignited, and the rocket lifted up from its pad on Omelek. But 34 seconds into the flight, the engines stopped firing. The team was watching a live feed from a camera, and suddenly the island below the rocket wasn’t getting smaller anymore. Someone in the control center said, “That’s not good.”

Fifty-nine seconds after it lifted off, the rocket crashed into the water just offshore.

Musk had been boasting about this rocket for years, and had invested so much into it. He knew the chances of success were slim, but still, this hurt. All of those hopes vanished in an instant, as pieces of the rocket, once so sturdy and erect, now lay strewn across the reef, polluting the pristine waters like so much litter.

“I remember how sad Elon was that day,” said Steve Davis, a young prodigy Musk had hired out of Stanford grad school. “Everyone was just super sad.” The rocket had been paid for almost exclusively by Musk, and at some point the money would run out “at which point we’re all done,” Davis said.

Hans Koenigsmann, one of SpaceX’s first employees, was also devastated. “My wife told me that after the first Falcon 1 failed, I didn’t talk to her for a month. And I didn’t even know. I didn’t talk at home at all,” he said. “It was a little bit more heartbreaking because we actually collected the wreckage. We picked it up. It’s better when it goes out of sight.”

SpaceX was trying to make spaceflight affordable, to open it up to the masses. To take humans further than even NASA had. That’s what drove them, even then. And if they couldn’t get the Falcon 1 to work, it wouldn’t just be the failure of a company, it would be as if “we have adversely affected where humanity could go. At least that’s how I felt,” Davis said. “We all felt that we were kind of the shot to do it.… Who else was doing it? There wasn’t anyone.”

It also hurt because Musk and his team at SpaceX knew that their critics were undoubtedly preparing their I-told-you-so, tsk-tsking postmortems. Legitimate, on-the-level criticism Musk could take; it was the condescension he couldn’t stand.

Still, any brooding would have to come later. In the moment, Musk tried to remain cool and professional—the whole team did—as they assessed the situation. The DARPA and air force team was impressed that the people from SpaceX, many of them young and relatively inexperienced, had remained so calm as they watched their rocket hit the water and explode, no matter how badly it hurt inside.

Afterward, Musk tried to stay positive, saying in a statement that “we had a successful liftoff and Falcon made it well clear of the launchpad, but unfortunately the vehicle was lost later in the first stage burn.”

They waited for low tide to start to recover the pieces of the rocket. As the SpaceX team pulled pieces of the wreckage out of the water, Weeks, the NASA official on loan to DARPA, tried to impose some order. The debris needed to be cataloged and marked, he said, collected as evidence for the investigation into the cause of the mishap. This needed to be an orderly process, not a free-for-all, yard-sale cleanup.

But Musk, smarting from the sight of his rocket strewn in pieces throughout the reef, was annoyed and pulled Weeks aside. “He wanted to know why I was directing his people,” Weeks said. This was his rocket, his explosion, his mess.

Except that it wasn’t, not entirely. Weeks reminded him that while it was his rocket, it had just exploded on a government facility on a mission that the government had invested in. There would be a government investigation, and the government would decide when and whether he’d be able to launch again.

“Elon likes to be in control,” Weeks said. “And you couldn’t blame him. This was his money, and he had put $100 million into it.”

DARPA’s Mishap Investigation Board, comprising Walker and Weeks and a couple of others working alongside SpaceX, completed its work by July 2006—in less than four months. The problem, it concluded, had been a fuel leak caused by the failure of a single nut of a fuel pump. The corrosive atmosphere of the islands, the salt air that Musk had worried about, had caused it to corrode and then fail.

With the cause found, DARPA cleared SpaceX to get back to flight. Despite the failure, DARPA was ready to stand by SpaceX and fund another try.

The education of SpaceX was only just beginning. And it had learned a powerful lesson about the perils of spaceflight: an entire rocket could go up in flames because of the failure of a single piece of hardware no larger than a pebble.

IN AUGUST 2004, Musk gathered the entire company, all eighty employees, in the cafeteria. He looked somber, or tried to. But he couldn’t keep the poker face for long. He was too excited.

“We fucking won!” he said.

SpaceX had won one of the COTS awards, a prize that could be worth as much as $278 million.

The staff cheered wildly. “There was this eruption of joy, people jumping up and down,” Shotwell recalled. “This was a huge deal for us, obviously. That it was no longer just Elon pouring his own personal money into it. I think he felt very—I don’t think vindicated is the right word, but I think he probably felt a great sense of relief that what he had built up until then was worth something, and NASA had recognized it.”

After the failure of the Falcon 1 launch, this was an enormous boost, one that gave them hope and motivation to carry on. The other winner was Rocketplane Kistler, an offshoot of the company that had won the sole source contract that Musk had sued over in 2004 and had recently emerged from bankruptcy.

Musk was ebullient, proclaiming it a significant step toward achieving its goal of dramatically lowering the cost of spaceflight. “I think it could be some of the best money NASA’s ever spent,” he said.

Others weren’t so sure they’d be able to drop the cost of space—something companies had been trying to do for decades.

“I’m tired of hearing that; it never comes true,” groused John Pike, a space analyst at globalsecurity.org, a think tank. “There have been no improvements of per-pound launch since John Kennedy was president.… I think it’s just going to be a good way to burn up a bunch of money.”

SpaceX and Rocketplane did seem a risky pair for NASA to bet on. Neither company had launched a rocket, and Kistler’s finances seemed shaky. It was unclear whether either of them would be able to come close to launching anything safely into orbit—let alone a spacecraft that would have to chase after the International Space Station, orbiting Earth at 17,500 mph. Skeptics inside NASA thought it was nuts, a complete waste of time and money.

To Musk, the supporters “were like the weird rebels within NASA,” he said. “They were given a project that everyone expected to fail, and they were just damn determined to make sure it didn’t. And they did so because they really believed that things needed to change. They did it for really good reasons. They cared about the advancement of spaceflight, and they really worked their asses off to help us succeed.”

But even they admitted they were bushwhacking into unknown territory.

“It was new to everybody,” said Marc Timm, a program executive at NASA. “The first meeting, we had no clue how we were going to do it. We had representation from legal, safety, Space Station, NASA Headquarters, and just sat down with a clean whiteboard and said, ‘How can we do this?’”

As daunting as it was, it was thrilling to be part of something completely new in an agency that had grown bureaucratic and stodgy, that was often far more comfortable saying no because it was unsafe.

Unlike the cost-plus contracts that sustained the traditional contractors, paying them even when they went over budget or over schedule, these awards were known as Space Act Agreements. The companies would only get paid once they reached certain milestones. By design, the money from NASA was not enough to keep the companies going on their own. They had to find additional outside investment, or customers.

If they didn’t, free market forces would take over, and they’d fail.

“Commercial companies had not done this before,” said Scott Horo­witz, then NASA’s associate administrator for the Exploration Systems Mission Directorate. “There was a high risk in them being able to develop the capability, which was supposed to come online when the shuttle retired in 2010. As much risk as there was on the technical side, there was even more risk on the financial side. To be quite honest, the business cases were not that robust.”

For years, commercial companies, SpaceX chief among them, had said that they could build rockets more safely and efficiently than NASA. That it was time for the private sector to take over. This was going to be their chance to prove it.

“There’s a lot of people that have been running around for years in conferences, yelling and screaming, ‘Get out of our way! Give us a chance! We can do it for one tenth of the price, and much more reliably. And a whole lot faster and better than NASA,’” said Griffin, the former NASA administrator.

“They’ve been screaming that for years. This was like, ‘show me.’”

No one was more vociferous than Musk. And he reveled in the opportunity to prove himself, especially with the strict, “show-me” way the payments were structured. The bolder, the better, he figured. The arrangement, he knew, favored a scrappy startup like his, and, for once, put the less nimble, more risk-averse traditional contractors at a strategic disadvantage. They could play it safe. He had nothing to lose.

“The funding is milestone-based, so if we don’t achieve milestones, there’s no money spent,” he said at the time. “It’s not like a standard government cost-plus effort where the worse they do, the more money they get. This is the case where if we don’t do what we say we’re going to do, we don’t get paid. So it’s a no-lose proposition for the taxpayer.”

“I think the fear on the part of some people out there is not that we will fail, but that we will succeed,” he added.

For SpaceX, the $278 million was a windfall. The deal was like a “Good Housekeeping seal of approval” that gave the company credibility in the marketplace, recalled Tim Hughes, SpaceX’s senior vice president and general counsel. If NASA trusted them, then commercial satellite manufacturers could as well.

But for “a lot of people in the big aerospace world this was considered a token amount of money to placate these commercial guys so they could stop complaining to Congress,” Musk said, years later. “Just give them enough to hang themselves.… Let’s just give these annoying commercial people enough money so that they can fail, and then we can say that was dumb. We don’t have to do that again.”

SpaceX was still not regarded as a threat. Not by the Lockheeds and Boeings, which had large, multiyear contracts in hand that could sustain them as long as there was support in Congress. And while the program was a bold step for NASA, one that SpaceX and other startups jumped at, it was largely dismissed by the rest of the industry. Lockheed Martin, Boeing, Northrop Grumman all passed, not thinking that this new era would last.

They were focused on another NASA program that was a central part of the Bush administration’s official “Vision for Space Exploration.” Called Constellation, it was the White House’s grand plan to bring humans back to the moon by 2020 and eventually to Mars. Instead of just a few hundred million dollars at stake, the construction of a pair of new rockets, the Ares I and V, a lunar lander, and a spacecraft, the Orion, was a prize worth tens of billions over many years. In comparison, the COTS program was mere crumbs. COTS was a side program, carved out by Griffin; Constellation was the big program of record that the big contractors remained focused on.

“They screwed themselves because they were just arrogant and complacent,” Musk said later. “Look, Boeing doesn’t get out of bed for less than a billion.”

(In response, Boeing noted that “Apollo and the programs that inspire all space enthusiasts would not have been possible without Boeing. At the turn of the 21st Century, before Musk entered the space business, Boeing was building the International Space Station with NASA, where we’ve kept astronauts safe and continuously on orbit for over seventeen years.” And the company said that “while others talk about aspirations and hopes, we actually do things in space and will deliver on our commitment to America’s journey to Mars. That’s what we get out of bed for.”)

Looking back on it, Griffin said he was surprised that the legacy contractors “did not see the handwriting on the wall that there would be in the future some commercial space enterprises.” They all had the “muscle mass,” he said. They were the “trained athletes on the field,” who could have “undercut even the best entrepreneurial firm” had they chosen to bid on COTS. Had they wanted it, the prize was theirs for the taking.

Griffin saw a lesson that could be taught in business school. NASA, the customer, was indicating it was moving in a different direction. But its traditional contractors ignored it, or didn’t take it seriously, or weren’t nimble enough to move with it. That created an opening. A small one. A mere crack a couple million dollars wide that was easy to ignore when they were marching through a gilded, billion-dollar doorway.

“The entrepreneurial firms can exist only if the larger established contractors create a niche for them,” Griffin said. “If I’d been running Boeing or Lockheed or Northrop, I would not have done that. I would not have allowed a fledgling competitor to invade my space.”

TWO YEARS LATER, however, the invaders were stumbling. The insurgency that COTS was supposed to foment was flaming out in SpaceX’s failed launch attempt, and subsequent delays, which fueled doubt and a lack of funding that put Rocketplane Kistler out of business. The millions that NASA had awarded the company as part of the COTS program were not enough to sustain it. And now the space agency was forced to go out and find another company to participate in the program, as critics questioned whether these young, inexperienced companies were worth NASA’s time and investment.

After its 2006 failure, it took SpaceX a year to fly again. This time, on March 20, 2007, it performed much better, flying to 180 miles, well past the edge of space. Its second stage separated, and the onboard camera showed the booster falling back toward the ocean below. In the control room, the SpaceX employees could see the curvature of Earth and the blackness of space.

“I’m going to watch that video for a long fucking time!” Musk said. “Congratulations, guys.”

But then, before the second stage could get to orbit, the second stage started wobbling uncontrollably before it fell back to Earth.

“This was a pretty nerve-racking day,” Musk said afterward. “The rocket business is definitely not a low-stress business, that’s for sure.”

Still, he found solace in the fact that the rocket had indeed made it to space. “I don’t think I’m disappointed,” he added. “In fact, I’m pretty happy.”

As he later pointed out, it was a test flight. The whole point was to see how the system worked, and root out any problems.

On August 3, 2008, the third launch attempt also came just short of getting to orbit. At stage separation, the first and second stages collided, causing yet another failure.

Publicly, Musk remained resolute. “SpaceX will not skip a beat in execution going forward,” he said. “We are in a very good financial basis here. We have the resolve. We have the financial base. And we have the expertise.”

He added: “For my part, I will never give up, and I mean never.”

Despite those rosy predictions, the truth was, SpaceX was hurting. Musk was burning through the $100 million he had invested of his own money, while it had also earned a $20 million investment from Founders Fund, the Silicon Valley venture capital firm started by Peter Thiel, who knew Musk from their early days at PayPal..

The challenge “was really keeping the company going financially while we were struggling. That was my contribution primarily,” SpaceX president Gwynne Shotwell recalled. “I didn’t get to do as much engineering as I would have liked to, but continually convincing customers to invest in SpaceX, and to take the risk associated with buying launches from us. I was focused on keeping the company alive, keeping people paid while we were struggling and getting through it. Because I knew we’d get through it. Technically, I knew we would get through it. It was a matter of could we get through it financially and stay stable.”

The three failures were “really incredibly painful,” Musk said later.

If this fourth launch didn’t work, “we would have ceased to exist. I was out of money.”

LESS THAN TWO months later, on September 28, 2008, SpaceX tried again. To prevent the first and second stages from colliding, the engineers adjusted the timing ever so slightly.

“Between the third and the fourth flight, we changed one number, nothing else,” said Hans Koenigsmann, SpaceX’s vice president of mission assurance. “That was the time we needed to separate the two stages.”

That was the technical solution. Tim Hughes, SpaceX’s general counsel, had a more superstitious one: adding a pair of four-leaf clovers to the mission patch. Throughout NASA’s history, the agency had created patches for each launch, artistic talismans that captured the spirit of the adventure while, hopefully, bringing luck. It was a ritual that dated back to Mercury, Gemini, and Apollo, and on through the shuttle. Astronauts were like ballplayers, believing that the power of symbols and a certain set of rules, however absurd—never step on the foul line—bequeathed success, a line drive up the middle, or not dying in a fireball just off the launchpad.

Knowing the importance that symbolism and superstition had for its astronauts, NASA allowed its crews to design their own patches. For Apollo 11, which would become one of the most recognizable, the astronauts decided to omit their names in a self-effacing gesture, though listing the crew members had been the practice in the past.

“We wanted to keep our three names off it because we wanted the design to be representative of everyone who had worked toward a lunar landing,” said Michael Collins, a member of the Apollo 11 crew. “And there were thousands who could take a proprietary interest in it, yet who would never see their names woven into the fabric of a patch. Further, we wanted the design to be symbolic rather than explicit.”

Neil Armstrong didn’t want the word eleven spelled out, as it was in the original design, because non-English speakers wouldn’t be able to read it. So, instead it would go by the numeral, 11.

SpaceX, now desperate for what could be its fourth and final flight, was willing to believe wholeheartedly in the power of the patch. It needed all the luck it could get. The four-leaf clovers were added to the patch.

THIS TIME THE Falcon 1 rocket got to orbit in a flawless flight that was perfect from the countdown to stage separation, which ended with a crescendo of cheers from SpaceX’s more than five hundred employees.

The company had broadcast the launch on the web, in a flawed, rudimentary real-time show anchored by amateurs whose enthusiasm and passion were genuine—they were SpaceX employees, after all—but who were ill trained for the rigors and length of almost an hour on air.

The moment Falcon 1 had achieved orbit, the commentator declared that “Falcon 1 made history as the first privately developed launch vehicle to reach Earth orbit from the ground.”

He then took a moment to explain the long and improbable odyssey that had led to this feat:

“SpaceX has designed and developed this vehicle from the ground up, from a blank sheet of paper they’ve done all the design, all the testing in house. We don’t outsource, and we have achieved this with a company that is only now five hundred people. And it has all occurred in under six years.”

For this launch, Musk was in California with the bulk of his team, not in Kwajalein. He emerged before the throng in a nondescript polo shirt, trying to rise to the occasion for his cheering employees standing around him three and four deep. Some had brought their children to the factory to witness the launch, and they hoisted them on their shoulders for a better view of Musk. But he was utterly overwhelmed and at a loss for words.

“That was freakin’ awesome,” he managed in a halting cadence, then stated the obvious: “We made orbit.”

After thanking his team, he said, “There were a lot of people who thought we couldn’t do it—a lot actually.” He let out a little maniacal, take-that laugh of redemption, before continuing.

“But as the saying goes the fourth time’s the charm, right? This really means a lot to SpaceX. Getting to orbit, that’s just a huge milestone. There are only a few countries that have done it. It’s normally a country thing, not a company thing. It’s an amazing achievement.”

As the words left his mouth, it was as if just then, for the first time, he realized the enormity of the accomplishment—it was what countries had done. And as that notion settled in, he once again became speechless.

“My mind is kind of frazzled,” he stuttered, exhausted. “Man, it’s definitely one of the greatest days of my life, and probably for most people here. We’ve shown people we can do it. This is just the first step of many.”

The Falcon 1 would lead to the Falcon 9, an even more powerful rocket—with nine engines, compared to one—that was under development, he said. And then there was the Dragon spacecraft, a capsule being designed to take cargo to the International Space Station. That was the next prize, the natural follow-up to the COTS program, which, after all, was supposed to help develop the capability to at least fly supplies—food, equipment, science experiments, toilet paper—to the station.

It was as if the opening of this one door would lead to another and then another, achievement begetting monumental achievement, one small step, leading to one giant leap springing outward to a point on the horizon that only he could see.

What else was possible?

“A lot of things,” he said, allowing his mind to wander into the future.

“Ultimately,” he said, “I think even getting to Mars.”

That remained to be seen. But from then on, every single SpaceX mission patch would have a common feature, one often hidden in the recesses: a four-leaf clover.

AT THE END of 2008, SpaceX got a gift two days before Christmas: a $1.6 billion contract from NASA to fly cargo in its Dragon spacecraft on as many as twelve flights to the International Space Station. For SpaceX, it meant that the company had finally arrived, and had been given NASA’s full, unmitigated endorsement. These would be real missions—delivering thousands of pounds of supplies to the space station—feats that would require the company to successfully launch its Dragon capsule to orbit, which would then fly and dock with the space station, traveling at 17,500 mph. It was an achievement that had been accomplished only by nations, and only a select few—the United States, Russia, Japan, and the European Union.

When Musk got the call from NASA, he was overcome. The last few years had been a struggle, filled with failed launches, even a divorce. But now six years after founding his space company, he had been vindicated. And when the NASA officials finished telling him that SpaceX had won the contract, he unceremoniously blurted out, “I love you guys!”

He also changed his login password to “ilovenasa.”

WHILE MUSK WAS charging ahead, Bezos was still moving slowly, taking small, deliberate baby steps. More than a year after Blue Origin had flown Charon, the massive dronelike vehicle powered by the four jet engines acquired from the South African Air Force, the company was ready to fly again, this time with something that looked like a big, white gumdrop, a flying, flat-bottomed Humpty Dumpty.

Before dawn on November 3, 2006, a cold morning in West Texas, a flatbed truck rolled the vehicle out of its hangar to the launchpad. And as the sun came up, turning the distant mountains purple, employees and their families gathered in the bleachers, some with blue cheerleader pom-poms, ready to watch the launch on a jumbotron.

The vehicle they had come to see was named Goddard, after Robert Goddard, the father of modern rocketry, who in 1926 became the first to ever launch a liquid-fueled rocket. He was a builder and a dreamer, who in 1919 wrote a paper, “A Method of Reaching Extreme Altitudes,” published by the Smithsonian Institution, that touched on the possibility of developing a rocket to reach the moon.

At the time, the notion of reaching the moon seemed as far-fetched as it was ridiculous. Goddard was derided as “moony” and “crackpot,” and even the New York Times wrote a scathing editorial in 1920 under the headline “A Severe Strain on Credulity,” which scoffed at the idea, saying a rocket could not work in the vacuum of space.

“That Professor Goddard, with his ‘chair’ in Clark College and the countenancing of the Smithsonian Institution does not know the relation of action to reaction, and of the need to have something better than a vacuum against which to react—to say that would be absurd,” declared the Times. “Of course he only seems to lack the knowledge ladled out daily in high schools.”

Goddard responded by saying that “every vision is a joke until the first man accomplishes it; once realized it becomes commonplace.”

But the ridicule led Goddard, a shy man who preferred working alone, to an even greater measure of reclusiveness and dedication to a long-term vision of spaceflight that he knew would take many decades to fulfill.

“How many more years I shall be able to work on the problem I do not know,” he wrote in 1932. “I hope as long as I live. There can be no thought of finishing, for aiming at the stars, both literally and figuratively, is the work of generations, so that no matter how much progress one makes, there is always the thrill of just beginning.”

He died in 1945, without having lived to see humans go to space. But just before the Apollo 11 moon landing in 1969, he received a belated, postmortem vindication. By then, it was abundantly clear that rockets could indeed work in space, and the Times issued a correction to its editorial, a half-century after it was published.

“Further investigation and experimentation have confirmed the findings of Isaac Newton in the 17th century, and it is now definitely established that a rocket can function in a vacuum as well as in an atmosphere,” it read. “The Times regrets the error.”

FROM A PURELY chronological standpoint, it made sense that Bezos named his first rocket after the father of rocketry. But they were also kindred spirits. Like Goddard, Bezos was dedicated to taking the long view, with Blue seeing it as an enterprise that would take generations to complete. Like Goddard, Bezos believed that the impossible could be made routine. And like Goddard, Bezos’s company shunned the press, keeping its work secret, carefully protected from scrutiny and the criticism that would surely follow.

In fact, Bezos was such a fan that he chose “Goddard” as the middle name for one of his sons.

After years of work, Bezos and Blue Origin had arrived at the debut of its first launch, modest though it would be. Still, there was indeed a feeling of the “thrill of just beginning,” and the company made it festive. A cowboy cooked biscuits over an open fire. For the kids, there was a bouncy castle.

The countdown played over a loudspeaker. Goddard lifted off, flying to 285 feet and then touching back down in a flight that lasted all of thirty seconds. Bezos celebrated with a giant bottle of champagne, and joked that his only job during the launch was to pop the cork. The launch was more successful than the uncorking, though. Bezos broke off the top of the cork, leaving the rest still in the bottle.

Goddard’s short hop was another small and modest step forward for the company, and it stood in stark contrast to the giant leap Musk had taken. SpaceX’s first successful launch wasn’t just a couple hundred feet. It wasn’t even a suborbital spaceflight. SpaceX instead went straight to orbit, achieving an extraordinarily difficult task that requires the rocket to go so fast that it essentially is continually falling around Earth. It was typical SpaceX: Head down. Plow through the line.

SpaceX was now looking ahead to its more powerful Falcon 9 rocket, which would launch from Pad 40 at Cape Canaveral Air Force Station. But Blue Origin remained steadfastly committed to its approach, even if it seemed the company was barely leaving the gate.

In following this deliberative path, Goddard would be followed by New Shepard, a rocket named for Alan Shepard, the first American to reach space. The company’s progress, then, would mimic the step-by-step evolution of American spaceflight. Goddard’s rocket design helped lead to Shepard’s suborbital, up-and-down flight in 1961, which lasted just 15 minutes and 28 seconds. NASA didn’t put a man into orbit until John Glenn circled Earth the following year.

Blue Origin didn’t say anything publicly about its launch or its next steps, until two months later. In a blog post, Bezos wrote:

“Accomplishing this mission will take a long time, and we’re working on it methodically. We believe in incremental improvement and in keeping investments at a pace that’s sustainable. Slow and steady is the way to achieve results, and we do not kid ourselves into thinking this will get easier as we go along. Smaller, more frequent steps drive a faster rate of learning, help us maintain focus, and give each of us an opportunity to see our latest work fly sooner.”

It didn’t matter how far ahead the hare was; the tortoise was content to keep the methodical pace Bezos had outlined for the company in his 2004 letter: “Be the tortoise and not the hare.” It would stay true to its motto, “Gradatim Ferociter,” or “Step by Step, Ferociously,” while repeating over and over:

Slow is smooth and smooth is fast. Slow is smooth and smooth is fast. Slow is smooth and smooth is fast.