Rocket Billionaires

by Tim Fernholz

  A long downtime would also give the political opponents of NASA’s commercial approach more credibility when they argued for a traditional government program for putting people into orbit. At the time of SpaceX’s launch, Orbital had yet to convincingly identify the problem that had blown up its rocket. It was planning to fly its Cygnus spacecraft on the Atlas V in the medium term, an “I told you so” moment for United Launch Alliance. The company was more than happy to step in where a commercial competitor had failed, underscoring its own focus on reliability, and NASA officials applauded Orbital for finding a creative way to keep supplies flowing.

  “I’m sure there are people that are gleeful when others have failures; it’s human nature, and it’s a very competitive environment,” Shotwell said years later. “I do not applaud when other people fail; I don’t get a sense of glee. I think there are some in the industry that do . . . Even [ULA chief executive] Tory Bruno sent me condolences when we had our failure.”

  Less than a month after the accident, Musk spoke to reporters about the progress of the failure investigation board and identified the cause of the failure: a high-pressure helium tank ricocheting around inside the vehicle after a metal strut from an outside supplier snapped under forces well below its certified tolerance. The incident had been classified as a “mishap,” because there was no loss of life or uninvolved public property. Had the consequences been harsher, it would have been an “accident” and investigated by the National Transportation Safety Board (NTSB), with more transparency and independence. As it was, the investigation included eleven SpaceX employees and just one FAA official, who did not sign the final report. Orbital’s failure, also classified as a mishap, had been investigated by a board that included two NASA employees and an independent expert. Some of SpaceX’s critics were skeptical of the process that had led to the company’s final explanation. A separate NASA investigation found several “credible causes” besides a bad strut from a supplier, including “improper installation of the assembly into the rocket” and “individuals standing on flight hardware during the assembly process.”

  If this were true, it was the kind of sloppiness that Mike Horkachuck, the company’s program manager at NASA, feared might seep into the company as it moved from development to operations. The allegations angered SpaceX employees; one told me forcefully that “technicians working on any airframe, rocket, or other hardware are required to walk over it and did not contribute to the strut failure. A board of engineers and the FAA voted and approved the material flaw to be the most probable root cause. Not somebody standing on the strut.” Still, after the process was completed, Gerstenmaier sent a letter to SpaceX—effectively a reprimand—“expressing concerns about the company’s systems engineering and management practices, hardware installation and repair methods, and telemetry systems.”

  Wounded pride or no, the company had to make good with NASA, its most critical patron and customer. SpaceX forfeited a share of its launch fee, worth about $44 million, renegotiated its contract to add additional future flights at discount prices, and invested in added capabilities for the Dragon, like more in-flight power for science projects. The company also undertook a reorganization under chief engineer Hans Koenigsmann, creating reliability teams to build more accountability into design, manufacture, and operations, with more stringent record keeping. SpaceX employees describe this as a time of high stress, with workers being pulled off their main jobs to reexamine every aspect of the Falcon 9 vehicle. Sometimes the people who made the components weren’t with the company anymore. Nonetheless, the teams would “go back to every file, every system, every design, and reevaluate, make sure we are on the right path,” Erin Beck Acain, who worked on the Dragon at the time, told me.

  The NASA inspector general, who examined the commercial cargo program after the two mishaps, suggested that the space agency might learn from the evolution of the EELV program’s approach to mission assurance. Yet it is notable that NASA did not end up adopting the same solution as the Air Force—paying additional money, in the form of cost-plus contracts, for guaranteed reliability. Instead, SpaceX added those new layers of organization at its own cost. That wasn’t the only loss it was taking: The months of delay while it pinpointed the problem with the rocket and implemented the internal changes to fix it meant delaying launches and the revenues that accompany them.

  “The biggest penalty for SpaceX will be delay of launch rate,” Musk told me that summer. “If the flights don’t take off, we lose the revenue associated with that . . . The launch revenue will be meaningful, in the hundreds of millions because of the implied delays.” Leaked financials would later reveal that the company lost $250 million after the accident.

  The company would push back the launch of its superpowered Falcon Heavy launch vehicle, but it did not want to delay a project considered far more vital to the company’s future: reusing its rockets.

  Since Richard Branson launched Virgin Galactic, in 2005, it had struggled to deliver on the commercial promise of the X Prize and SpaceShipOne. Virgin Galactic was partially inspired by Branson’s experience in the airline business, which began in 1984 when his flight to a Caribbean island was canceled. He chartered a plane of his own and sold seats to his fellow disgruntled passengers, and then realized he could do the same thing at scale. Branson had made his fortune as a music impresario and then applied his cheeky touch to a series of ventures that exploited the brand, from retail stores to hotels to cruise ships. In the airline business, he added value by designing a hip experience and marketing campaign, but the planes were the same jets that everyone bought from Boeing and Airbus.

  He sought to follow the same model with Virgin Galactic, but the market for SpaceShipOne had so far been limited to Paul Allen and the X Prize. To obtain both a mother ship and a passenger-carrying space plane, Branson created a joint venture with Burt Rutan’s Scaled Composites, called the Spaceship Company (TSC). It would be dedicated to building the flight hardware that Virgin Galactic—and perhaps, in time, other “spacelines”—would market and operate. In practice, this meant that Rutan’s Scaled Composites team was working on the hardware while Branson’s marketers made promises about it.

  The new, larger vehicle being developed by TSC faced some of the same problems as its predecessor: it was difficult to find a rocket engine small enough and powerful enough to carry it into space. Rutan and his team were still enamored of their hybrid rocket motor, which combined concepts from both solid and liquid rocket engines to fire the space plane once it was dropped from the mother ship, called White Knight Two. The company was using nitrous oxide—the laughing gas you might enjoy during a visit to the dentist—as the oxidizer in its engine.

  One day in 2007, TSC engineers began testing a system that pumped high-pressure nitrous oxide at the Mojave airport. It was a “cold flow” test to check out the plumbing that had been used several times before. With no plan to ignite anything, eleven people watched the test from just a few feet away, rather than from a command center behind a protective earthen berm. Seconds after the test began, however, the tank and the equipment exploded. Two people died on the scene, and a third at the hospital shortly thereafter; three more were injured.

  A shocked Rutan spoke to reporters afterward. “We felt it was completely safe; we had done a lot of these with SpaceShip One,” he said of the test. “We just don’t know.”

  To this day, it isn’t clear what caused the explosion. Nitrous oxide is considered fairly stable, though it can explode if it reacts with certain compounds or reaches high temperatures. That may have been one cause of the accident, which took place on a hot concrete tarmac in the middle of a summer day in the desert. Some experts believe the tank itself was compromised. California’s state safety agency cited Scaled for not training its employees in dealing with the compound or having written safety procedures for how to approach it. Despite the freak nature of the accident, it jibed with suggestions that Rutan’s team flew too much by the seat of their pants, as
they had in the run-up to winning the X Prize. Traditional aerospace tut-tutted about their rough and ready approach, but at the same time needed their creativity—at the time of the accident, Northrop Grumman was in the process of buying out Rutan’s share of Scaled.

  This still left a problem for Branson’s team to resolve, one that bedeviled all of the aerospace start-ups: How do you move from development to operations successfully, maintaining your culture without compromising reliability?

  Change is an inevitable part of the answer. In 2010, at age sixty-seven, Rutan would retire from active work at Scaled. That same year, Branson hired George Whitesides, then the NASA administrator’s chief of staff, as the CEO of Virgin and the Spaceship Company. Whitesides had come far in the decade since he worked at BlastOff. Giving a boost to his mandate was a deal that Branson had cut the year before with the United Arab Emirates. The tiny, oil-rich collection of sheikhdoms sported a sovereign wealth fund dedicated to funneling its petroleum profits toward long-term investments in technology. Branson, in his own telling, flew to Abu Dhabi and sealed a $280 million investment in Virgin in a single day. The company was revitalized, but adaptations—including a redesign of the fuel tank to dispel any concerns about the use of nitrous oxide, hiring an outside company to redesign the propulsion system, and eventually bringing the project back in-house—pushed back Branson’s inaugural flight for years.

  A decade after Scaled Composites had won the X Prize with SpaceShipOne, SpaceShipTwo—the actual vehicle was dubbed VSS Enterprise—was finally entering the critical stage of its test program. It had already been flown as a glider some thirty times to help establish its aerodynamic characteristics, and, starting the year before, it had performed three powered flights where the rocket engine was ignited. The Enterprise had reached a maximum altitude of thirteen miles, and it seemed only a matter of time before it would get to space.

  Early on the morning of October 31, 2014, the Enterprise dropped from its mother ship, fifty thousand feet over the Mojave, for its fourth powered test. Just like SpaceShipOne before it, the vehicle relied on adjustable wing booms that would feather up on reentry to slow it down and maintain the correct angle to avoid breaking apart. During flight, as the space plane accelerated toward the speed of sound, the copilot unlocked this rotating wing. This was standard procedure—but only after the rocket had reached full speed and exited the lower atmosphere. Because the wing was unlocked too early, the aerodynamic force on the vehicle overcame the motors that rotated the boom. The wings deployed suddenly, and the force flipped the Enterprise over backward, tearing it apart.

  The pilot, Pete Siebold, told investigators afterward that he felt enormous g-forces pushing him back into his seat, before hearing the cabin crack apart and feeling the air pulled from his lungs. He blacked out and woke up outside the cabin, plunging toward the ground, more than ten miles below. His face hurt from the cold, and the seal on his oxygen mask had broken, leaving him short of breath. He remembered unbuckling himself from his seat and free-falling. Somewhere between twenty thousand and ten thousand feet above the ground, his parachute deployed automatically, and he was able to land safely in a creosote bush, albeit with an arm broken in multiple places, a fractured collarbone, and bloody scratches. First responders arrived by helicopter after what felt like a “really long time.” Siebold’s copilot, Michael Alsbury, was found dead in the wreckage of the Enterprise. He was the first person to perish testing a commercial spacecraft.

  The accident shook Virgin and the space community, which was still reacting to the Orbital mission that had failed just days before. Siebold and Alsbury were respected test pilots who had dedicated their careers to testing next-generation spacecraft for Scaled Composites; Alsbury was survived by his wife and two small children. The two fliers were spiritual descendants of Chuck Yeager and the other rocket plane pilots who came before them, pushing the envelope in an effort to reach the stars. At Virgin Galactic, heavy hearts concluded that the best tribute to Alsbury would be pressing on with the project. Outside the company, critics laid into Branson’s overoptimistic promises and speculated about long-running problems with the power of the vehicle’s rocket engines.

  The NTSB’s judgment was that Scaled had never considered that a single human mistake could result in the destruction of the entire vehicle. In a situation where busy pilots are required to make precisely timed decisions despite high stress, designers should have anticipated the potential problem; this was an explicit requirement in NASA’s Commercial Crew contracts. But it was also not in keeping with Rutan’s philosophy; the old-school designer once claimed that “if space is going to be cheap, it has to be stick-and-rudder.” The accident sent a different message: space travel was too fast, too complicated, and too dangerous for human hands alone.

  Branson had clearly been stung by media criticism that suggested his ego was behind the accident, but he was using the pique to press on. He decided that the only way forward would be to bring the entire operation in-house, as the other rocket billionaires had. Two of the earliest SpaceX employees, Chris Thompson and Tim Buzza, had joined Virgin in 2012 and 2014, respectively, sharing their approach to spacecraft design. Virgin had already been in the process of buying out Scaled’s share of their joint venture, the Space Company. Now there would be a clear chain of command, with test pilots dedicated to Virgin Galactic’s projects only, not shared among multiple new vehicles. The next SpaceShipTwo, already under construction, would be carefully reexamined for the “human factors” that had been left out of the first version; a former US Air Force test pilot was put in charge of pilot safety.

  By 2017, a new SpaceShipTwo, the VSS Unity, was undergoing glide testing and aiming for a return to powered flight in 2018. The company’s optimism had returned. It raised more money from investors in the Middle East and began to prepare its team for the move to New Mexico to begin normal operations. But, just as the Apollo program had overtaken the work of the rocket plane test pilots in the 1960s, vertical-launching rocket companies like SpaceX and Blue Origin had stolen Virgin Galactic’s thunder. The X Prize win that had catalyzed the company seemed as distant as the idea of a spaceline carrying paying passengers around the world.

  Jeff Bezos’s space company was even older than Virgin Galactic, but his tight-lipped approach protected him from accusations of exaggeration. Yet something—whether it was competitive instincts, the challenge of recruiting the best talent, or sheer pride—compelled him to join the mobs on Twitter in November 2015 to show off the “rarest of beasts” that Blue Origin had birthed: the New Shepard’s reusable booster stage had made it back to earth in one piece after its second test flight.

  Blue had rebounded quickly from the loss of the first booster during the company’s spring flight, producing another flight-ready vehicle. Once again, Bezos arrived at his Texas ranch and joined Blue’s development team in their small control room; the New Shepard runs largely on internal computers, part of the company’s policy of trying to get humans out of the control chain wherever they could to improve reliability.

  On the second trial, the rocket once again took off with a blast of its single engine and ascended to the edge of space; this time, the hydraulics held up throughout the entire descent back to its landing pad. New Shepard’s spindly landing legs unfolded from recessed compartments on the booster’s sides and supported the rocket as it hit the ground. The celebrating team, drenched in champagne, had done what no one had ever done before: launch a rocket to the edge of space and bring it back down again in one piece.

  The same quibbles over suborbital and orbital velocities as the last launch applied here, inspiring a barrage of skeptical tweets from Musk. He noted the record of the X-15 space plane and defended SpaceX’s Grasshopper program and the fact that his orbital rockets had arrived at precise spots above the ocean to prepare for sea landing. But to no avail: Musk had shared videos of his rockets exploding; Bezos had film of his rocket landing gently on the ground, with him standing next to it, a
nd all the bragging rights this entailed. The general public did not honor the finer distinctions about orbital velocities.

  It was the first time that Blue had pipped SpaceX to the post.

  This news came as SpaceX was winding up its investigations into the CRS-7 explosion and applying for a license to return to operational flight. The company planned to launch satellites for Orbcomm before the end of the year. It was the second half of a constellation the company had originally sold to fly on the Falcon 1 several years before. SpaceX would use the opportunity to debut a new, upgraded version of the Falcon 9, called “full thrust,” that promised 30 percent more power thanks to larger fuel tanks and modified engines. This extra fuel could be the difference between the reusable booster crash-landing and making a graceful dismount.

  SpaceX was granted a license to launch its rocket just a few days before its window opened; just as important, it received permission to try and land its first stage at Cape Canaveral. This was a very different task from landing it on an autonomous barge a hundred miles out to sea; though the landing pad at the Cape would be clear, there would be people and expensive infrastructure within miles of where the rocket was intended to touch down. An error would be disastrous; after all, a rocket descending was just a missile without a warhead. Considering that their last flight, less than six months earlier, had been a failure, SpaceX’s executives were making a big ask of their regulators at the FAA and the US Air Force team responsible for the range at Cape Canaveral. They were told to go ahead.