by Tim Fernholz
“Speaking very broadly on behalf of established aerospace—we really believed our own press releases,” says Albrecht, who made a career in aerospace after his time in government as CEO of Lockheed Martin’s International Launch Services. “We believed that space was really hard and really expensive and really difficult because that’s the way we were doing things for the government.”
Now that assertion would be put to the test. Lindenmoyer began figuring out how to put that $500 million to work. Michael Wholley, the general counsel at the space agency and a former fighter pilot, helped him settle on a legal mechanism called a Space Act Agreement to avoid the usual red tape. It had a unique historical origin: the man who had been tasked to write the original National Aeronautics and Space Act—the 1958 legislation called for by President Eisenhower to create NASA in the first place—was a nervous young attorney who feared some snafu in the legalese that might imperil the race to top Sputnik. To give himself some cover in case he had forgotten something important, the lawyer added a final clause that gave NASA the authority to make any agreements it needed to in order to fulfill its mission. “He basically said, ‘If I’ve forgotten something, use this,’” Wholley said.
Besides eliminating red tape, this “Other Transaction Authority” also allowed NASA to create incentives for participating companies to make significant investments of their own capital into the effort, leveraging the government’s money. This was vital to ensuring that sufficient funding could be found for NASA’s Constellation program. The first rocket in the program, Ares I, was billed as the primary replacement for the space shuttle, capable of safely carrying crew to low earth orbit. It was to be accompanied by a much larger cargo-carrying rocket called the Ares V. However, both were designed largely with planetary exploration in mind, not the simpler task of space station service.
Griffin had been thinking about how to build rockets for a long time; with James French, another veteran aerospace engineer, he literally wrote a textbook on the topic. When he consulted for Musk and accompanied the entrepreneur on his rocket-finding expeditions to Russia, Griffin had made a proposal: let him hire a team of engineers to build a rocket with components from existing supply chains. His knowledge of the industry would allow him to deliver a rocket that would suit the mission perfectly. By this point Musk was already skeptical of the aerospace status quo, and getting into what John Garvey, Tom Mueller, and others were up to out in the desert. By the time he offered Griffin a job as chief engineer at SpaceX, the two men failed to see eye to eye.
Now Griffin would make an offer that Musk—and every other rocket entrepreneur—couldn’t refuse. It amounted to a huge influx of seed money to develop the technology that Musk desired to get to Mars. The space agency put out a call for participants in a technology development partnership. Applicants would need to demonstrate a service wherein a spacecraft would carry cargo, pressurized and unpressurized, to the space station on a reliable rocket. If they could accomplish that, NASA hoped they would then progress to carrying astronauts.
NASA’s lawyers worked over the Christmas holiday to figure out the details of avoiding procurement regulations. In early 2006, NASA rolled out the official announcement for proposals—they didn’t call it a “request for proposal” because that term of art was reserved for traditional contracts. This would simply be a development effort to stimulate private-sector space capability. Beyond the basic capabilities desired, companies would be free to develop their own offer from start to finish, describing what they’d do with a share of the $500 million.
Instead of requirements, the companies would propose their own “milestones” for evaluation, which would be tied to financial payments. For years, space start-ups had been telling NASA that they could do a better job building rockets. Now they’d have a chance to prove it.
8
A Method of Reaching Extreme Altitudes
It would cost a fortune to make a rocket to hit the moon. But wouldn’t it be worth a fortune? The great pity is that I cannot commercialize my idea.
—Robert Goddard, 1920
I think I’ve come to realize what makes orbital rocket development so tough,” Elon Musk wrote in an update sent to fans of his company and posted on SpaceX’s website on New Year’s Day 2005, shortly before Griffin was nominated to take over at NASA. In the early years of the company, Musk would write blog posts, heavily salted with the rocket jargon he had picked up in his studies, describing the work of his team or offering candid commentary on the status quo: “We drew some of our ideas from an old Thor rocket and its mobile launcher that are sitting in a museum at Vandenberg. It is not clear to me why those ideas were abandoned.”
Yet the previous year had been a wake-up call for Musk, a sign of how hard it would be to realize his vision at SpaceX as delays in launching the Falcon 1 mounted. This resulted in an introspective turn. “It is not that any particular element is all that difficult, but rather that you are forced to develop a very complex product that can’t be fully tested in its real environment until launch and, when you do launch, there can be zero significant errors,” Musk wrote, implicitly comparing his current work to his past life in the software industry. “Unlike other products, there is no chance of issuing a bug fix or recall after liftoff. You are also forced to use very narrow structural safety margins, compared to an aircraft or suborbital rocket, to have any chance of reaching orbit at all and must hit a bull’s eye when you do.
“Having seen us go through the wringer to make this work (and it’s not over yet), I have a lot of respect for anyone that has tried to develop a serious launch vehicle.”
Musk was clearly thinking about his work in terms of Silicon Valley aspirations. At the end of one of his 2004 blog posts, he noted that his team would go over their rocket “with a microscope” to make sure everything was ready ahead of a flight, because, “as Andy Grove said, ‘Only the Paranoid Survive.’” He linked to a copy of the Intel CEO’s book of the same title at Barnes & Noble’s website—not Amazon—apparently never being one to give a competitor even a tiny advantage.
In 2004, the company erected its Falcon 1 rocket at Vandenberg Air Force Base in California. The military installation is most famous for testing nuclear ICBMs and hosting US antimissile interceptors. Like Cape Canaveral, it is specially positioned to provide access to space. Unlike the Cape, which makes it easy to launch satellites eastward, in the direction of the earth’s rotation, Vandenberg is best for launching satellites that revolve around the earth from south to north, a path known as a polar orbit. This is ideal for, among other purposes, spy satellites that aim to cover as much as of the earth as they can with their prying eyes. SpaceX was set up to use a small launchpad as a site to test their rocket operations close to its Los Angeles base.
Working out the ground systems was an important preflight exercise, but SpaceX was still waiting for Mueller’s team to fire the Merlin engine at full duration. This was accomplished by the end of the year; while Musk was meditating on the challenges of building a rocket, the company had progressed to testing the fully constructed vehicle.
By this time, SpaceX faced competition for its airspace: the National Reconnaissance Office was scheduled to fly a top-secret spy satellite from Vandenberg, which required a long lead time for attachment to the rocket. The satellite in question is known as Keyhole-11 (KH-11) and is similar to the Hubble Space Telescope—a telescope tube the size of a school bus—but pointed down at the earth instead of up. The SpaceX launch would involve an experimental rocket flying over this multi-billion investment in space hardware, something the powers that be weren’t going to risk. The spy satellite would be a passenger on the last flight of the Lockheed Martin–built Titan IV, from the late eighties. Delays in launching the aging rocket ate into SpaceX’s plans to debut a new one.
The company got tired of waiting. To escape regulatory restrictions, SpaceX had already made plans to launch Falcon 1 missions from a tiny outpost of US power deep in the Pacific—the Kwajal
ein Atoll, in the Marshall Islands. With the days of nuclear testing behind it, the outpost functioned as a kind of target range for Vandenberg: whenever the need to demonstrate the capability of the American nuclear counterstrike arose, a target would be deployed from “Kwaj” for a Vandenberg-based interceptor to attack. The atoll was also home to powerful radar stations and even a huge laser target that had been designed to be struck from space during the Strategic Defense Initiative. The place fit into Musk’s pseudo-supervillain lifestyle, but his main concern was that SpaceX’s rocket was lagging behind his aspirations. It was time to get flying. C-17 Globemasters, the enormous big-bellied aircraft used to transport military vehicles, began hauling sections of the Falcon 1 across the ocean to the company’s new primary test site. It was a small island in the atoll called Omelek.
It was tough going for the Falcon 1 team on the island. They lived on the eponymous main island, Kwajalein, and their commute was a forty-minute boat ride across the atoll to Omelek. To pass the time, engineers gave one another impromptu lectures or earned their diving certifications. But insects, sunburn, and crippling boredom belied Musk’s cheery description of the island as “a tropical paradise,” as he wrote in one blog update. The remote facility, far from the manufacturing floor, workers’ homes, or regular supplies of electricity, was hardly ideal for rocket testing. The salty air and humidity would corrode electronics, and vital supplies were hard to come by.
In one test cycle, the team ran out of liquid oxygen—half of the critical propellant needed to light the engine. A combination of a broken storage valve, unexpected high temperatures, and poor planning left them with nothing to do except charter more planes to fly to Hawaii for tanks of LOX. Another anecdote, relayed in a blog post written by Musk’s brother, Kimbal, during a visit to the “rocket island,” underlines the SpaceX team’s Herculean efforts. The electronic circuits that powered the Falcon 1’s computer systems were acting up, and the engineers decided to replace them. The rocket was pulled down and the circuit boards were removed and given to an avionics engineer, Bulent Altan, who flew back to California overnight. That same day—a Sunday—a SpaceX intern was dispatched by plane from California to Minnesota to pick up new components from a supplier. Altan and the intern met at SpaceX headquarters on Monday morning, assembled the circuits, tested them, and packed them up. Altan then flew back to Kwaj, landing at 6:00 a.m. to begin installing the new components and reconstructing the rocket. The total turnaround time for this mission, according to the younger Musk, was eighty hours.
Once they finally got the whole system ready for launch, the winds rose to the point that launching the rocket would be too dangerous. The team began to pump fuel out of the vehicle so they could lower it safely to the ground. In the process, a bad electrical connection kept a valve from closing, creating a vacuum in one of the tanks that caused it to buckle and become useless. More delays. SpaceX had entered 2006—and now approached its four-year anniversary—without flying a rocket at all.
In March 2006, the company winched a repaired Falcon 1 back up to vertical on the launchpad on Omelek. It was time to try for a first launch again. At the moment of ignition, everything looked like it was going well—“nominally,” in rocketspeak. The engine fired and the rocket rose above the atoll, carrying a satellite designed by students at the US Air Force Academy as a practice payload. But after thirty seconds of flight, the engine kicked out—flames appeared to burn topsy-turvy around the base of the rocket, without the controlled oomph of the engine’s guiding hand. The rocket, now just a heavy metal tube surprised to find itself several thousand feet in the air, plunged onto a nearby coral reef. The students’ satellite was thrown clear by the impact and burst through the roof of the improvised machine shop that the company had constructed on Omelek. The first flight of the Falcon 1 was a failure.
It would be months before the cause of the crash was identified by an investigation conducted jointly by SpaceX and DARPA, the ostensible customer for the flight, and led by Worden.
“I went out there to watch their practice launch, and they were going to launch a few days later,” Worden told me, recalling how he wound up leading the investigation. “It looked to me like a bunch of kids trying to write software rather than rocket engineers trying to do hardware. So I wrote a rather scathing report, sent it to Elon and the DARPA director.” This prompted a bit of a scrap with Musk, who mocked Worden as an “astronomer,” which he is.
“Okay, look, I’m not criticizing your engineering, not your rocket technology,” Worden remembers saying to an irked Musk. “I was a US Air Force operations guy. I said there were certain characteristics of groups that succeeded, and he didn’t have many of those. He had a lot of characteristics of those that failed. It goes back to what [Admiral Hyman] Rickover used to say about nuclear submarines: the devil is in the details, and so is salvation.”
A fuel leak had allowed kerosene to drip down onto and into the engine; after liftoff, the engine itself caught fire. The extra consumption of fuel by the unexpected fire caused the pressure in the engine to drop, effectively shutting the whole thing off. Investigators found that corrosion had occurred around an aluminum nut securing a fuel pump, sometime during the eighteen hours of launch prep or during the three months the Falcon 1 spent in a warehouse without temperature or humidity controls.
Musk told reporters that the company would replace the aluminum fasteners with stainless steel ones to avoid the problem in the future. “The irony is we are replacing them with a cheaper component to increase reliability,” he lamented. Many at the company blamed the failure on the circumstances that had forced them to launch from the island—to wit, the ever delayed Titan launch at Vandenberg operated by their rival Lockheed Martin.
“The first launch failure was heartbreaking, because we were fifty, sixty people, maybe more,” Hans Koenigsmann said later, recalling dispirited engineers collecting broken pieces of the rocket off the beach. “I spent probably three or four months on the launch site in the middle of the Pacific for that. At the end, it didn’t fly very far. We learned a lot of things we did wrong, and learning sometimes hurts.”
While Koenigsmann and the Falcon 1 team attempted to suss out what had gone wrong with their first flight attempt, Musk and the rest of the team had to shift focus. They were in the process of bidding to participate in NASA’s space taxi program, and they had to convince the agency that everything they had learned made them worthy of a shot at servicing the space station.
The response to NASA’s call for a new private-sector orbital transit system was robust. Twenty-one plans arrived—from small companies like SpaceX and SpaceDev, which helped build the engines for the X Prize–winning SpaceShipOne, as well as from “prime contractors” like Boeing and Lockheed Martin. While the program was officially agnostic about which companies would be chosen, it became clear that traditional aerospace firms weren’t prepared to do a new kind of business.
“The larger companies requested more funding; some of them requested the whole thing, which was not a good fit,” Lindenmoyer said. “I thought we were pretty clear about that, but that didn’t work . . . We also were looking at projected pricing: if the system we were helping develop was so expensive that only the government can afford it, that’s not something we wanted, either. Some of the larger companies didn’t rate very well on that element.”
The six finalists were all new space companies, and a committee of NASA officials evaluated each company in three areas: the feasibility of its technology, whether it had the potential to become a sustainable business in the future, and its prospects for obtaining financing outside of the government. The NASA team sought help answering the latter two questions by recruiting Alan Marty, who had worked at several tech companies and then led a team of venture capitalists at J. P. Morgan. Marty’s job was to help NASA’s executives get into an entrepreneurial mind-set. He brought dozens of copies of Clayton Christensen’s book The Innovator’s Dilemma—an iconic Silicon Valley tome abo
ut how stagnant companies are disrupted by start-ups armed with outside-the-box thinking—to hand out at every NASA meeting he participated in. With Marty guiding the financial evaluation and Lindenmoyer on the technical side, NASA began considering its options.
SpaceX’s pitch stood out initially, for a number of reasons. Thanks to Musk’s personal wealth, it had already gotten a good start in developing a new rocket engine, the Merlin, and a launch vehicle, Falcon 1, which had undergone one test flight already, albeit a failed one. None of the other companies were close to that level of full-scale testing. SpaceX had a plan to attack markets beyond just NASA by flying satellites for corporations, the Air Force, and the academy. And Musk had the company thinking about human spaceflight: it already had schematics for a spacecraft, called the Dragon. Musk said the vehicle was named after the titular beast in the song “Puff the Magic Dragon,” an ironic riposte to claims that SpaceX was a pipe dream.
But SpaceX hadn’t found a way to pay for the Dragon. There weren’t many people clamoring to spend $60 million to fly a few people into space; the most a space tourist had paid so far was $20 million to visit the space station on a Russian rocket. Now NASA was not just offering SpaceX the seed funding to build a human-carrying spacecraft, but promising a commercial market for space transit that could fund the company’s ultimate aim of exploring Mars. Especially after the failure of the first Falcon 1 launch, a new stream of revenue was vital to the company’s hopes. It was so important that, aside from the team working toward the second test of the Falcon 1 out in the Pacific, the rest of the company threw themselves into preparing their pitch for the space agency.
