by Tim Fernholz
It had taken exactly two days since SpaceX filed its suit for it to hit home with its competitors, who claimed in court that the injunction even endangered near-term launches, because it would prevent them from paying Russian advisers for their support. The injunction was overturned about a week later, when the government’s attorneys were able to produce letters from sanctions enforcers making a distinction between Rogozin and the company he controlled, NPO Energomash. But ULA’s weaknesses were now on display.
Musk and Gass kept up the drumbeat of rhetoric against each other’s firms. At a press conference announcing the lawsuit, Musk invoked the old-fashioned American virtue of competition and hammered on ULA’s connection with the sanctioned oligarchs, noting that “it would be hard to imagine some way that Dmitry Rogozin is not benefiting personally from the dollars that are being sent there.” ULA’s team took the position that SpaceX’s challenge was a threat to national security itself: “SpaceX is trying to cut corners and just wants the USAF to rubber stamp it,” Gass told the Washington Post that summer. “SpaceX’s view is just ‘trust us.’ We obviously think that’s a dangerous approach and, thankfully, so do most people.”
Musk would take the war of words further. The Air Force official who had okayed the block buy at the center of the legal dispute, Roger “Scott” Correll, retired from the Pentagon shortly after making the decision. He would emerge as the vice president of government relations at Aerojet Rocketdyne, the engine maker that worked closely with ULA. It was undoubtedly a case of the revolving door at work in the military-industrial complex, but Musk suggested something worse in a series of tweets.
“V[ery] likely AF official Correll was told by ULA/Rocketdyne that a rich VP job was his if he gave them a sole source contract,” Musk wrote. “Reason I believe this is likely is that Correll first tried to work at SpaceX, but we turned him down. Our competitor, it seems, did not.”
This was, of course, an explosive accusation. In 2003, an Air Force official named Darleen Druyun was found to have negotiated a post-retirement job at Boeing while also negotiating the government purchase of in-air refueling tankers from the aerospace giant. She and her manager, CFO Michael Sears, were fired and wound up serving short jail sentences for their roles in the scandal. Now, with Correll, Aerojet immediately said Musk’s allegations were without merit and that Correll’s hiring had been entirely aboveboard. SpaceX spokespeople were taken by surprise when Musk shared his allegation and were bombarded by calls after the impromptu comments. Correll himself did not comment on the allegations, and no legal action was undertaken.
Despite Musk’s loose talk, SpaceX did not want the details of this case spilling out in public. As soon as the company had filed its protest, it also asked that court records be sealed to protect proprietary information; indeed, Judge Braden would order the companies to stop talking to the press about the issue. But a careful examination of the redacted record and discussions with government insiders give a fairly clear picture of what happened.
The substance of SpaceX’s complaint rested on the precise timeline of the contract awards and the distinctions among the Defense Department’s political leadership, the Air Force purchasers, and the lawmakers on the Hill who controlled the purse strings. The fact that this political battle spilled over into federal court shows how tangled the web of influence surrounding the US rocket monopoly had become. Judge Braden declined ULA’s request to dismiss the case, tartly noting that “the court does not request or need the views of the [ULA] . . . [which] has no basis to challenge [SpaceX’s] standing in this case.” Instead, she ordered the Air Force to disclose the details of ULA’s contract to SpaceX’s attorneys and ordered them to prepare a settlement proposal for mediation; the eventual mediator would be former US attorney general John Ashcroft.
The government and United Launch Alliance tried to argue that the contract issued by the Air Force did not violate the directive issued by Kendall in 2012 to buy more launches through open competition, and that SpaceX had missed its chance to protest. SpaceX said the opposite. Lawyers dug through complex contracts and cost estimates throughout the fall. This was vital, because it prevented the Air Force from hiding behind the excuse that SpaceX would not be able to meet its mission requirements; while the company was not yet prepared to fly the largest spy satellites, many of its smaller missions—particularly the newest generation of GPS satellites—were well within the company’s bailiwick.
The writing was clearly on the wall for ULA to see. In Congress, Senator McCain was working to pass a law banning the import of Russian rocket engines. George Sowers, who had been the primary designer of the Atlas V and warned his managers about SpaceX early on, was assigned to lead a team that would develop a plan to compete with SpaceX. “That’s a big deal,” Sowers told me later. “Imagine setting up a company to be a government-regulated monopoly, subsisting mostly on [Federal Acquisition Regulation] contracts, and transform it to be commercially relevant.”
The transformation did not come quickly enough for ULA’s parent companies. In August 2014, Gass, then fifty-eight, announced his retirement from the launch monopolist, in view of “the changing industry landscape.” An executive from Lockheed Martin’s ballistic missile business, Tory Bruno, was brought in as CEO with a mandate to make ULA competitive. In the next year, Bruno would lay off a dozen executives and begin preparations for a broader transformation that would mean reducing the entire workforce by 30 percent in the years ahead.
But change had to be more than just cutting the fat. ULA would have to come up with a whole new vehicle to replace its two expensive rockets if it wanted to begin competing with SpaceX. The company called its new rocket design Vulcan, still unable to escape the tug of Greek mythology. It would upgrade in steps, beginning with a powerful new booster stage that would, in effect, be able to replace both Atlas and Delta boosters in flying ULA’s Centaur second-stage vehicle. Then the company would replace the Centaur with something called the Advanced Cryogenic Evolved Stage, or ACES, with even more power to shoot large satellites directly into ultra-high orbits. This was all fine and good, but the biggest problem would be coming up with a whole new engine, and fast, since importing the RD-180 just wasn’t politically tenable.
With its monopoly in danger, ULA was already looking at serious cost cutting. Finding a billion or so dollars to invest in a powerful orbital rocket engine would mean cutting back on the steady flow of cash heading to its parent companies, and having to do so in a joint venture that already boasted a complicated cost-sharing agreement. Public markets were likely to look askance at putting up the funds to help a stumbling legacy company compete with new entrants unburdened by historical costs.
Yet, amazingly, there was someone out there who had the money to pour into such a venture, someone who was already competing with SpaceX. His name was Jeff Bezos.
Blue Origin had an existing relationship with ULA; when Bezos’s company was dabbling in NASA’s Commercial Crew partnership, it intended to use the Atlas V to carry its capsule into space. One of Blue’s business development executives was Brett Alexander, who had worked at NASA on the space taxi program and had consulted with Sowers at ULA. He stayed in touch with his old employer, giving him hints about the “supersecret” engine development work going on at Bezos’s closely guarded space firm. When ULA began looking for a new engine, Blue pitched it some promising ideas and, just as important, adapted its plans to ULA’s preferences for Vulcan, increasing the thrust by 25 percent for the larger vehicle. The engineering appeared to make sense, but, as with NASA, the promise of self-financing sealed the deal.
“The business deal they were willing to offer us was a business deal you only dream about,” Sowers told me. “You have a supplier to take all of the development risk on themselves, and most of the cost.”
Other companies were considered for the job, and Aerojet Rocketdyne was brought on as a competitor, thanks in part to the longtime contractor’s backing in Congress. But it was clear that Blu
e Origin held the pole position, a fact made clearer after remarks by ULA’s vice president of engineering, Brett Tobey, to a classroom full of students were leaked online.
Tobey was a victim of his own generosity. In March 2016, he candidly shared with students at the University of Colorado the reality of the aerospace business, right down to the importance of political connections. “McCain basically doesn’t like us,” he said on the recording. “He’s like this with Elon Musk, and so Elon Musk said, ‘why don’t you guys go after ULA and see if you can get that engine to be outlawed?’” On the other hand, Tobey noted that “we have this friend—I told you about that big factory down in Alabama, Decatur—this is Senator Richard Shelby from Alabama.”
But from a business standpoint, his most telling remarks concerned the engine competition between Blue and Aerojet.
“We’re sitting here as a groom with two possible brides,” Tobey said. “We’ve got Blue Origin over here, the super-rich girl, then we’ve got this poor girl over here in Aerojet Rocketdyne . . . The chances of Aerojet Rocketdyne coming in and beating the billionaire is pretty low. We’re putting a whole lot more energy into BE-4, Blue Origin.”
Tobey would resign after his views were published by the media, but the reality he described was one lamented by those companies without a wealthy patron like Musk or Bezos. “You have these well-capitalized businesses with very, very long-term plans, without any near-term or medium-term expectation of return on investment,” James Maser, the SpaceX president who later ran Aerojet, told me. “That has really upset the traditional business model.”
The new partnership between the country’s biggest aerospace contractors and Bezos’s private R&D shop signaled the pressure SpaceX was putting on the traditional rocket launch business. “There is no way that ULA would have considered buying engines from Blue Origin except for the pressure that SpaceX put on them,” Mueller, SpaceX’s engine guru, said later in leaked remarks of his own. The deal was something of a masterstroke for ULA; the threatened incumbents had found an ally in exactly the kind of disruptive company eating away at their bottom line. SpaceX had denied Blue Origin a launchpad in 2013. Now Blue was stepping in to bolster SpaceX’s most powerful rival just when Musk’s company had the government contractor on the ropes. If battle lines hadn’t been drawn before, they were now.
To be sure, the intervention couldn’t come quick enough to prevent Musk from wedging his company into the competition for government launches. ULA made one last legal gambit in early 2015, arguing that a recently passed funding bill—which technically banned the use of Russian rocket engines while making an exemption for ULA’s contracts—had ratified the block buy. Judge Braden once again slapped down this argument, saying that SpaceX still had a case and that she was prepared to rule on it if mediation failed to result in a settlement. In the end, a sealed settlement was indeed agreed to by SpaceX and the US Air Force.
Despite the legal wrangling, the settlement did not kill ULA’s block buy; defense officials ultimately would not risk the supply chain for the heavy rockets that only ULA could provide. For this reason, the deal could be seen as a defeat for SpaceX. At the same time, the publicity of the suit and the scrutiny of the deal made clear that this paradigm could not continue, and for that reason SpaceX claimed at least a moral victory. And perhaps more: the block buy’s estimated cost fell to $11 billion as officials shifted more optional launches into competitive bidding ahead of schedule.
SpaceX was offered a path to certification ahead of schedule, and, later in 2015, the company would win the right to launch a GPS satellite for the Air Force, its first successful bid on a national security contract. This would mark the end of ULA’s monopoly on American national security launches, a major moment for SpaceX, which had been derided by the incumbent virtually since its founding.
Musk’s company was now flying satellites for all comers, and pushing toward human spaceflight through NASA’s Commercial Crew program, though budget cuts had pushed back the first planned flight another year. Yet this was hardly enough for Musk, or for the team at SpaceX. “If all we do is be yet another satellite launcher . . . [or] only as good as Soyuz in cost per person to orbit, that would be okay, but really not a success in my book,” Musk had said in 2007. By 2015, his company was on its way to matching the full spectrum of what rocket engineers had accomplished before. To truly change the game—to make Martian colonization practical—SpaceX would have to do something no one had ever done before. It would have to make its rockets reusable, to drive the cost down not through mere efficiency but by way of a total paradigm shift. This vital fact was clear to Musk and his team. But it was also clear to Bezos and his team. Which is why, the same year Musk sued the Air Force, he also took Bezos’s company to court. Why?
Blue Origin had patented his idea to land rockets on boats.
13
Reduce, Reuse, Recycle
It would be a mistake to consider that reuse is the alpha and omega of breaking innovation in the field of launchers.
—Stéphane Israël, CEO of Arianespace
Since the very beginning of SpaceX, the company had dedicated itself to a fairly obvious intuition: What if we don’t throw away the rocket when we are done using it?
Nearly every rocket ever made has been designed to be thrown away. The space shuttle orbiter was unique as a reusable space vehicle, but it relied on jettisoning its solid fuel rocket boosters and an enormous fuel tank mid-flight; the boosters were recovered from the ocean, refurbished, and reused. Satellite-launching rockets were thrown away entirely. The reasoning for this reliance on expendability was simple: There’s already a tiny margin of error involved in reaching orbit—remember, most space vehicles are 85 percent propellant by mass. Adding the gear to make a rocket reusable increases mass and further reduces that margin. And bringing that rocket back down to earth in one piece involves exposing it to the harsh conditions of reentry, which could harm it enough to make reusability pointless.
Rocket builders also didn’t think it was worth investing that money in solving these problems, because there just weren’t enough launches happening to justify the expense. It was cheaper to make throwaway rockets than to invest in reusable rockets—unless, that is, you thought you would be able to fly them a lot more often than before. In 2015, George Sowers, then the vice president of human launch services at United Launch Alliance, was intrigued by SpaceX’s ideas and attempted to figure out if they were onto something. He came away unconvinced.
“Can you actually bring it back and refurbish it for lower cost than to build a new one?” Sowers asked me rhetorically. “I’ve done a lot of analyses; I’ve convinced myself that, at least with today’s technology, the answer is no.” Executives at other big rocket makers echoed that view.
Musk and his team disagreed from the get-go. If he was going to spend so much of his own money, particularly on the expensive engines, he wasn’t about to throw them away. More important, reusability was the only way to lower the cost of launch dramatically enough to match his ambitions. Musk had a simple analogy: The cost of building his rockets was comparable to that of a 737 airliner, but because they were thrown away after a single use, their cost per flight was wildly higher. While the cost of building the rocket was around $54 million, the propellant used in each Falcon 9 flight cost only about $200,000. If the company could reuse even just the first stage, it estimated that it could cut the price by about a third. SpaceX’s expendable rocket was already dramatically cheaper than those of its competitors, but effective reusability would essentially end the competition entirely.
For the Falcon 1, SpaceX engineers originally expected that the first stage would be able to return to earth via a parachute, where it would then be fished from the sea. The challenges of testing that rocket and its eventual cancellation meant that this plan never came to fruition. As the Falcon 9 was developed, it became clear that parachutes would be insufficient for returning the twenty-ton first stage back to earth. Return from
space brings us back once again to the physics of getting there. Remember, your vehicle’s velocity must exceed 17,500 miles per hour to stay in orbit. To return to earth, you have no choice but to plunge back into the atmosphere at extremely high speed. As you do, the vehicle smacks into the air in front of you, cramming the gas tightly together and making it enormously hot. While it’s comparatively easy to make metal structures that can withstand the physical force of reentry, dealing with the high temperatures is a more difficult engineering challenge.
Space vehicles have typically relied on special shapes to move that heat away from vital areas, and special materials are used that can absorb this energy. The Apollo and Soyuz space capsules let their blunt bellies take the heat of reentry before deploying parachutes. The much larger, reusable space shuttle counted on its heat shielding to absorb the force, then used its glider body to slow down in a series of wide turns before landing on the runway. Yet it, too, was a cautionary tale—not just because the Columbia disaster revealed how vulnerable the shielding could be, but because the expense of refurbishing it turned out to be far higher than expected.
SpaceX’s competitors expected that the upstarts would learn the same hard lesson. “There was a chief engineer of another launch provider—I will not say the name—who told me, categorically, to my face, ‘You will never land a first-stage booster,’” Martin Halliwell, the chief technology officer at European satellite giant SES, said in 2017. “‘It is impossible, and even if you do it, it will be completely wrecked.’”
There was a third way to go, beyond parachutes and heat shielding—at least in theory. It was called retropropulsion, which means flying the rocket backward toward earth on a cushion of hot gas generated by its rocket engines. This was the iconic image of the golden age of science fiction: a rocket landing engine first on an alien planet. The technique had been used in the Apollo lander because the moon has so little atmosphere, but trying it on a planet with a thicker atmosphere would be far more dangerous. Yet if SpaceX could rely on powered flight to slow down the rocket, it would need less expensive heat shielding. Retropropulsion was also a vitally important technology for SpaceX’s larger mission of building a city on Mars.