by Tim Fernholz
“While it was wonderful that Boeing bid and won and was investing, it was pretty clear to me they were never going to be the economic competitors with SpaceX that we ultimately wanted,” Garver, the deputy NASA administrator at the time, told me later. She recalled an invitation from Bezos to visit the company’s design and test facilities. “He toured me around, and it became one hundred percent clear that they are in it for the long haul. They had engine bells lined up, thirty of them.” On a visit to Blue’s Texas facility, Garver saw a test stand as big as the one NASA used to test the Saturn V rocket engines, which was capable of withstanding eleven million pounds of thrust. She asked the program manager, a young engineer, how much it had cost to build, and he estimated about $30 million. She told him that it was costing the space agency $300 million to refurbish NASA’s stand to test the SLS. “Yeah, I know,” the engineer told Garver. “I used to work at NASA there; that’s why I left.”
He wasn’t the only one. As the space shuttle program shut down after its final flight in 2011, its home base at the Kennedy Space Center had to transition. After some fits and starts, the center’s leadership decided on a master plan that would cut costs by selling or leasing surplus hardware and facilities. The biggest surplus on offer was the right to lease the Space Launch Center 39-A, which included a launchpad and tower, fueling gear, and warehouse facilities for working on rockets and for strapping satellites and spacecraft on top. SLC-39A was iconic as the launch site of both the Apollo moon rockets and the space shuttle missions. Now it would go up for sale to the highest bidder who could make use of it.
The two primary bidders were SpaceX and Blue Origin. SpaceX won the lease in 2013, with the expectation of using the site to launch two rockets: the Falcon 9 and a new, more powerful rocket it was developing called Falcon Heavy, which would effectively combine three Falcon 9 boosters into one twenty-seven-engine vehicle fit for deep-space launches. Acquiring SLC-39A would mean that the company was leasing three different launch sites from the government: two at Cape Canaveral and one at Vandenberg. Musk wasn’t done: in 2014, he would lease a patch of land in the very southern tip of Texas to build his own launch facility, which remains under construction as of press time.
Blue was not happy with NASA’s decision and challenged the choice on the grounds that it would share SLC-39A with other users—arguably delivering more public bang for the buck—while SpaceX wanted exclusive access. The arbiters sided with SpaceX, saying that NASA had never expressed a preference about exclusivity. Musk was blunter about the decision in a 2013 interview with Reuters. “I think it’s kind of moot whether or not SpaceX gets exclusive or non-exclusive rights for the next five years,” he said. “I don’t see anyone else using that pad for the next five years . . . It’s a bit silly because Blue Origin hasn’t even done a suborbital flight to space, let alone an orbital one. If one were to extrapolate their progress, they might reach orbit in five years, but that seems unlikely.”
Later, he clarified his views on those odds, emailing a reporter to say that, “frankly, I think we are more likely to discover unicorns dancing in the flame duct.”
12
Space Race 2.0
I am not surprised that Germany has awakened to the importance of [rocketry] . . . I would not be surprised if it were only a matter of time before the research would become something in the nature of a race.
—Robert Goddard, 1923
SpaceX had been founded at a tough time for rocket makers. Sea Launch, the European champion Arianespace, and Lockheed Martin’s joint venture with Russia’s Khrunichev State Research and Production Space Center had been the key players in the market for launching private satellite systems at the turn of the century. But after the tech crash pulled the rug out from under ambitious satellite entrepreneurs, the rocket makers were forced to cut prices dramatically as the demand for their vehicles dried up. They could barely give their wares away, selling launches at well below cost.
In 2007, however, the cycle began to reverse. Gwynne Shotwell had a front-row seat as she traveled the world, hawking SpaceX’s rockets to potential customers. The creation of United Launch Alliance put a formal close to any expectation that the new EELV-class rockets would aim at the commercial market. As that tie-up progressed, Lockheed pulled out of its joint venture with Russia to market the Proton rocket—there were too many potential political conflicts of interest. The Proton was already seen as something of a low-quality rocket. That was, in fact, part of the pitch behind Lockheed’s partnership—selling discounted packages of cheaper Proton and “Cadillac” Atlas rockets to serve different needs. But with the dissolution of the international partnership, accountability waned and failures began to pile up, including the dramatic loss of a Japanese communications satellite in September 2007.
“Despite the incredible capability and robust designs, Russian space technology production always suffered from quality problems,” Mark Albrecht, who ran the partnership for Lockheed, told me. “While we could provide no direct technical assistance to Russian space technology companies, our conversations about quality control, our approach to independent testing and validation did rub off on them. Once American aerospace partnerships ended, quality began to lapse even further.”
Sea Launch, the joint venture between Boeing and a Ukrainian rocket maker, was a bit cheaper. But the system entailed sending the rocket to sea for weeks at a time on a floating platform, which limited the number of launches it could perform each year. In January 2007, a Sea Launch rocket turned into a ball of fire after a “foreign object” tore through the engine shortly after launch, causing delays to their future manifest and raising launch insurance rates.
That left private satellite companies with limited options. Arianespace was reliable, but very expensive. Japan had a heavy rocket, but it was also expensive and focused on its own domestic market. As big satellite companies looked ahead, they were planning to revitalize satellite constellations launched in the nineties that had ten- or fifteen-year lifespans. If they wanted to avoid competing for a limited supply of expensive choices, they’d need to take a page from NASA and invest in creating a new capability. The Falcon 9’s development cycle coincided with rising prices and reduced options in the rest of the rocket world, which made buyers willing to put down cash on an untried system in order to save money in the future.
“People wanted to be part of that big change. I don’t know if anyone looked at it as a revolution at the time, but people wanted to have additional access to space,” Shotwell said of the rising prices and the Sea Launch and Proton accidents. “Failures are bad for industry. It’s hard to grow the market size when people are worried about ‘How in the world are we going to get this thing to space?’” SpaceX’s offer of a reliable, lower-cost rocket was welcome news in the industry.
That timing—or luck—helped SpaceX survive as it struggled to complete the qualification process of the Falcon 9 and the Dragon for NASA. In 2012, five months after the first rendezvous in orbit, another Falcon 9 was teed up on the Cape Canaveral launchpad for the first official contracted cargo mission to the space station. This time, the vehicle was carrying more than just expendable food and water; it brought up replacement parts for the station’s life support systems, mechanical apparatus for performing experiments, and scientific samples in freezers. SpaceX, always eager to maximize its resources, also tucked an experimental satellite for the company Orbcomm behind the Dragon; the payload had originally been scheduled for Falcon 1 but was bumped up to the big rocket.
The Falcon 9 left the pad on a tower of flame, but just over a minute into the flight, observers on the ground saw a flash and a spray of debris from the base of the rocket. The thrust chamber in one of the nine engines had burst during flight, likely due to a manufacturing flaw. The debris were aerodynamic panels bursting free to relieve pressure in the vehicle. Flight computers shut down the destroyed engine and adapted to a new trajectory. The rocket was slowed, but it was not stopped.
The n
ewly created Falcon 9 was able to carry the second stage to separation. After the two stages split, the Dragon was delivered to its rendezvous point with the ISS. The engine failure, however, made it impossible for SpaceX to fly the Orbcomm satellite to its proper altitude, which registered it as a partial failure. Still, Orbcomm’s team said they were able to gather some data during the brief time their satellite operated in its too-low orbit before burning up.
Despite this hiccup, what really mattered was the success of the primary mission despite the engine loss. This proved Mueller and Musk’s promise that their rocket could survive this exact scenario, something that couldn’t be said of any other rocket since the Saturn V. With four orbital flights under its belt, SpaceX had proven that it had not just an effective rocket, but a rugged one.
The next year, 2013, would see the company flying another Dragon mission to the space station. More important, the company launched its first two commercial satellite contracts. Since SpaceX had only launched its own Dragon, these missions gave it an opportunity to show that its rocket could play well with spacecraft designed by others. It could also demonstrate the protective carbon-fiber nose cone that fit over satellites, called a fairing and made in-house by SpaceX at a cost of $8 million. First, the company launched a small satellite for the Canadian Space Agency from Vandenberg. At that launch, the company also debuted a new version of the rocket, dubbed the Falcon 9 v1.1, just like an iterated software program. This vehicle had significantly more powerful engines, in an easier-to-assemble arrangement, and larger fuel tanks.
This upgrade significantly increased the power and efficiency of the rocket. That mattered enormously to the second satellite launch of 2013, a communications satellite for a company called SES. This was a big deal, because SES is a giant in the satellite industry, a Luxembourg-based company operating dozens of satellites in orbit and acting as a blue-chip purchaser of rockets. Like most of the major satellite firms, it invested in the most expensive satellites, enormous machines placed carefully at ultra-high-altitude orbits around the earth.
This is special real estate; it’s called a “geostationary” orbit because a spacecraft at that altitude must fly at the exact same speed as the earth’s rotation. It allows the satellite to track one specific spot on the planet below, effectively “hanging” over a region. Broadcasters love this altitude, because it provides more consistent, reliable coverage than satellites launched at lower orbits that might go around the planet fifteen times a day. And because broadcasting is the most lucrative business in space, launching satellites to geostationary orbit is the most lucrative business in rocketry.
Naturally, getting a satellite up this high requires a powerful rocket, hence the Falcon 9 1.1. It also requires the second stage of that rocket to fly a careful maneuver to put the satellite on the right path. SES had partnered with SpaceX before almost any other major satellite operator, so that it would not have to rely only on super-expensive, government-produced rockets to find these characteristics. As a result, the Luxembourgers paid less than $60 million for a launch that might have cost more than $160 million on the open market. It wasn’t just nervous SpaceX employees watching in December 2013 as the countdown commenced at Cape Canaveral. SES employees who had bet on the risky upstart knew that a failed launch would mean more than the loss of their very expensive three-ton satellite—it would mean relying on older, more expensive rockets for years to come.
Five previous attempts to launch this mission rocket had been scrubbed, but on the sixth attempt, as the clock hit zero, the rocket’s engines rumbled to life. It launched at sunset, the sky behind it painted with baroque purples, and soared into the night like a dagger of flame. The mission was another success and, as was now a tradition, the company’s Los Angeles employees went wild in the cafeteria and balcony that overlooked the glass control room. “NASA helped develop a capability where the US can finally regain dominance in launch,” Shotwell said that year. SpaceX had a spacecraft that could do work for NASA and a rocket that could do work for the private sector. Now they were going to make money—real money.
When the Falcon 1 test program was scattering rocket parts across a Pacific atoll, SpaceX’s employees had received sympathy emails from their former colleagues at the other prime companies. But once the Falcon 9 was flying, those good-faith gestures dried up, and SpaceX’s government relations team noticed an uptick in criticism from lawmakers and the media. Their competitors had stopped ignoring SpaceX, and now rival lobbyists and public relations teams were pushing back hard, and with good reason: SpaceX had its sights set on the most lucrative prize in the launch world, a $19 billion contract to fly five years’ worth of space missions for the US government.
In addition to private satellite operators and NASA contracts, Elon Musk wanted to break into the final segment of the launch business: national security. The US military and intelligence communities, after all, are the biggest launch market in the world, operating one of the largest and, arguably, most important satellite constellations. The military’s geeks had backed SpaceX in its early days, when SpaceX was focused on the Falcon 1 and launching small satellites quickly.
The rocket that the company ultimately put up for sale, the Falcon 9, didn’t fit the bill. Instead, it was a direct competitor to the rockets being used by the United Launch Alliance monopoly that the US Air Force had endorsed in 2006: the Atlas V and the Delta IV. That crucial decision to allow the EELV program’s two competitors, Boeing and Lockheed, to combine in a joint venture relied on the assumption that new rockets would not be available soon. So when SpaceX showed up six years later with a far cheaper rocket than what ULA had on offer, suddenly the meetings in contractor boardrooms and Capitol Hill offices became much more awkward.
This was especially so because in 2009, the new Obama appointees at the Defense Department had woken up to the fact that they did not understand why the prices of ULA’s rockets continued to rise. In 2007, the Pentagon had declared the program’s acquisition phase completed, which meant less government oversight. Now the DoD launched a number of investigations into exactly what was happening in ULA’s production system, with the ultimate goal of figuring out how to cut costs. Particularly important was a survey of ULA’s suppliers, ordered by the Air Force and performed by the monopoly itself, which suggested that if the government ordered forty launches over five years, the rocket builder would then be able to deliver some discounts.
This influential survey would later be found to be deficient. Outside evaluators discovered that it had been accompanied by a letter urging respondents to “justify” a purchase strategy to “enhance our collective business.” One executive told auditors that ULA “wanted certain answers” from the subcontractors it surveyed. At the same time, ULA was providing contradictory answers to the government. In some presentations, it said its suppliers were running below capacity and in danger of going out of business, which might lead to higher government subsidies. In others, it claimed they were busy and financially healthy. Pentagon officials didn’t even bother to examine the data underlying the survey.
The inquiries into the EELV program, conducted on several fronts, came to a head in 2011 as the Air Force was contemplating the block buy of five years’ worth of launches for the Air Force and the National Reconnaissance Office. The lawmakers in charge of approving the funding for this purchase asked government auditors if the Air Force was competent enough to buy rockets without getting fleeced. The answer was definitely not a yes. Their report noted that 20 to 60 percent of ULA’s reported costs were either “unsupported or questioned”—a vital problem if you are paying a contractor for their costs plus a guaranteed profit. Both defense officials and ULA acknowledged “that launch prices may increase substantially in the coming years.” The auditors recommended that the Pentagon slow down and do more due diligence before committing to such a massive expense.
“It was damning for this procurement,” one lobbyist who followed the issue closely told me. After the audi
t was released, lawmakers ordered the Pentagon to recertify the purchases from ULA as a major acquisition program. This would give the government the ability to pull back the curtain and examine the situation in more detail. The immediate result of this decision is called a “critical breach of Nunn-McCurdy,” the kind of fate bewailed by the overworked aides scurrying around the capital trying to keep the edifice of government standing.
Named after the lawmakers who wrote it, this was a law put in place in the 1980s to automatically terminate overspending defense programs if immediate action wasn’t taken. Critics of the EELV program say Bush administration officials reduced reporting requirements precisely to avoid the consequences of such a breach. But the sheer amount of cash being plowed into the launch program made hiding the overruns impossible. This was especially true at a time when the front-page news in Washington was about brutal political clashes between Tea Party Republicans and the Obama administration over public spending. These debates cost the US government its AAA borrower rating, resulting in harsh restrictions on defense and discretionary spending alike.
The Pentagon’s 2012 forecast of a 58 percent increase in cost for the EELV program was a cry for help. The launches cost, on average, over $400 million each—more than four times what SpaceX would have bid at the time. Buying rockets from ULA would become the fourth-biggest procurement expense in national security, lagging behind only advanced jet fighters, submarines, and destroyers for the Navy. One DOD evaluator, attempting to make sense of the overages, found that the contract structure “implies that money has been spent on effectively idle personnel.” That report observed that, while some causes of high costs—the vagaries of the US space program and the international launch market—were unavoidable, “the final cause is poor program execution due to an environment in which little incentive for cost control, or threat of termination, exists for the vast proportion of EELV’s content.”