Dave Masten stared at his computer monitor over a jumble of screwdrivers, tea packets, and dog-eared physics textbooks that cluttered his desk. “Anybody want to watch this?” he called out to no reply. It was about noon on Thursday, April 11. He scanned his office, a scrubby quadruple-wide trailer at the Mojave Air and Space Port in Southern California’s high desert, but found he was alone.

That isn’t uncommon. The entire team at Masten Space Systems, the rocket company that Masten founded in 2004, numbers 15 people. The seven based in Mojave—mostly young men who wear T-shirts emblazoned with sayings like “I need my space”—spend some time at their desks, working through equations or crafting proposals for clients like NASA. But they are more often found in the converted military garage across the dusty parking lot, tinkering with rocketry.

Masten turned back to his monitor, which was showing the live-stream broadcast of Beresheet, a lunar lander developed by SpaceIL, a privately funded Israeli nonprofit. Beresheet had been launched by a SpaceX Falcon 9 rocket a few months earlier, and had spent the previous week orbiting the moon in preparation for its landing attempt. If it touched down without issue, it would become the first private vehicle ever to land on the moon.

As Beresheet descended, Masten strained to make out the chatter in the background of the SpaceIL broadcast. A few minutes before the targeted landing time, he heard someone say the team had lost contact with the inertial measurement unit, which measures the spacecraft’s acceleration and rotation.

“Shit,” he thought. “They lost the mission.”

Masten’s interest in Beresheet’s flight was personal. His firm is hard at work on its own moon lander.

That lander, the XL-1, is just under three and a half meters (11.5 feet) long and just over three meters wide. With technical input (though not funding) from NASA through its Lunar Catalyst program, the Masten team designed the lander to carry a 100-kilogram (220-pound) scientific payload to the moon’s surface and survive there for 12 days. Three spherical propellant tanks balanced on spindly legs huddle beneath a rectangular solar panel, giving the probe the appearance of a giant ant carrying a matchbox on its back. The tanks hold a proprietary combination of nontoxic liquids that spontaneously ignite when combined, powering four main engines and 16 maneuvering thrusters, all of which hang off the contraption’s sides. The whole thing weighs 675 kg (1,488 lb) without fuel, and 2,675 kg, as much as a Toyota Tacoma pickup, when “wet.” It is simple and cheap, and was promising enough for NASA to select Masten in late 2018 as one of nine companies to take part in the Commercial Lunar Payload Services program (CLPS, pronounced “clips”).

Getting to space has always been expensive; getting to the moon, even more so. Astrobotic, one of the CLPS participants, quotes a price of $1.2 million per kilogram to reach the lunar surface. (Other companies generally refrain from putting a number on it.) As NASA sets out to return humans to the moon by 2024—a surprise deadline recently imposed by the Trump administration—CLPS is an attempt to figure out if private companies can get there quickly and on the cheap. NASA will pay for cargo to be delivered to the moon, but not to design or build the spacecraft that get it there. The aspiration is for CLPS to function like a lunar delivery service.

Masten is the smallest of the nine CLPS companies. Lockheed Martin, with 100,000 employees and a market value of $96 billion, is the biggest. NASA’s latest budget allocates $80 million per year to CLPS, and if the program goes well, this could increase to a total of $2.6 billion over the next decade. Being part of CLPS gives companies the right to compete for contracts through a series of “task orders”—if they aren’t chosen, they aren’t paid. If they are, they get a fixed fee and have to figure out how to use it to get to the moon.

On May 31, the first task order (totaling over $250 million) was awarded to three firms: Orbit Beyond, which will launch in September 2020, and Astrobotic and Intuitive Machines, which plan July 2021 launches. Steven Clarke, NASA’s deputy associate administrator for exploration, says subsequent task orders will create a “good cadence of missions”—initially about two a year, increasing to three or four missions per year by around 2023. None of the CLPS entrants are building a new launch vehicle; they will buy rides to orbit from commercial providers. For instance, Orbit Beyond and Intuitive Machines plan to ride to Earth orbit on a SpaceX Falcon 9.

NASA has not landed a vehicle—let alone a person—on the moon since 1972. Going back just for bragging rights no longer makes much sense. Dave Murrow, a senior manager working on CLPS at Lockheed Martin, says, “Flags and footprints were great in the 1960s—it was very important for us as a nation at that point. But now we need something sustainable.”

It is unclear if there will ever be enough demand for lunar travel to support a healthy industry. The answer will hinge partly on what the landers find on the moon. Marshall Smith, director of human lunar exploration programs at NASA headquarters, believes there is an abundance of water at the moon’s south pole that could be converted to rocket fuel and drinking water for astronauts.

Dean Eppler, a NASA veteran (and economic geologist) who is now chief lunar scientist at the Aerospace Corporation, is less certain. Lunar orbiters, he said at a recent forum his firm organized in Colorado Springs, have gathered about as much information as they can. To figure out if mining the moon for water is viable, “we really have to get down on the ground,” he said. “That’s what the CLPS program is going to be important for. And thank God it’s here, because it would be a hard road without it.”

Ever since the end of the Apollo program, NASA has struggled to reinvent itself as an efficient enterprise. The “Faster, Better, Cheaper” initiative of Daniel Goldin, who ran the agency from 1992 to 2001, is now widely derided; critics accused it of contributing to two failed Mars missions and the 2003 disintegration of the space shuttle Columbia, in which seven astronauts died. “We as an industry got to a spot where, wow, those failures were really painful. We’re not going to do that again,” Murrow says.

Eager to stretch money further without repeating the same mistakes, NASA increasingly relies on private partnerships. Beginning in 2006, it used a concept similar to CLPS to bid on cargo shipments to the International Space Station, recalls Lori Garver, a former deputy administrator of NASA.

The program spurred SpaceX to create the Falcon 9 launch vehicle, which cost about $390 million to develop. NASA estimates that had it developed the vehicle, the cost would have ballooned to between $1.7 billion and $4 billion. But outsourcing is no guarantee of success: a more recent effort to use commercial providers to send human crews into Earth orbit is meeting with the same sorts of delays that NASA’s own programs confront. And Garver questions if the lunar market is big enough to be viable.

CLPS takes a particularly streamlined approach. The CLPS request for proposals was about a dozen pages, compared with the hundreds of documents with endless compliance requirements that normally accompany NASA collaborations. The contracting structure is designed to make protest lawsuits difficult—a wonky but important detail if NASA is to move quickly, since procedural hang-ups are a frequent source of delays. And NASA appears to have gone out of its way to give smaller firms a chance. Though not as small as Masten, the three firms chosen for the first task order are all modest by the standards of aerospace. Apart from Lockheed Martin, only one of the other participants is a large aerospace firm: Draper, a nonprofit corporation that was founded in 1932 as part of MIT.

Chris Culbert, the chief technologist at NASA’s Johnson Space Center (JSC), who manages CLPS, told the Aerospace Corporation forum, “This might be the best chance in many of our careers to actually tell NASA how to do things differently.” Trent Martin of Intuitive Machines, who previously spent a decade each at Lockheed Martin and NASA, is even more effusive: “I’ve been around the agency a really long time, and I’ve never seen anything like it.” If CLPS works as designed, even bigger firms like Lockheed and Draper must prove they can compete on cost and speed with much leaner companies.

For Lockheed, a CLPS contract would be nice. But for the smaller companies, the stakes are higher. Steve Bailey, who runs Deep Space Systems, says he’s “betting the company” on CLPS. Murrow of Lockheed Martin muses, “Sustainable economic activity isn’t going to happen from just one company dominating or monopolizing—it’s going to happen from a diverse set of participants with different strengths and weaknesses, different risk postures, and, frankly, different probabilities of success.”

For NASA, CLPS represents an ideal sort of lean agility. Officials from the administrator, Jim Bridenstine, on down have taken to saying the agency is more interested in taking swift “shots on goal” than in certain but plodding progress.

The first CLPS task order was assigned before NASA had figured out what, exactly, it wants delivered where: “Within the next couple of months we’ll have sorted out which payloads will go on which landers,” Culbert said when the orders were announced. His team at JSC numbers fewer than seven people—a sign, he says, of the trust NASA is putting in commercial providers.

Of the CLPS competitors, Masten Space Systems has an ethos that seems particularly well aligned with this focus on speedy experimentalism. Dave Masten thrives on rapid movement, an inclination that surfaces in his personal life—he marked his recent 50th birthday by attempting a 50-mile trail run—and his professional one. As a rocket engineer, he has long advocated for relentlessly testing and tweaking reusable machines rather than trying to nail designs on the first try. This means that even though Masten is tiny, it can lay claim to forms of experience that larger companies lack. Xombie, Masten’s first operational rocket, has flown 227 times—which the firm claims is a record for any rocket-powered airframe.

For Masten, the CLPS program presents a path to a more stable future, after years of scraping by. But “not to get stinking rich, because this is aerospace, and you don’t actually get rich in aerospace,” Dave Masten says. Masten chose to be chief technology officer of the firm that bears his name so that he can spend more time building rockets, and less worrying about money. Sean Mahoney, a jovial 45-year-old with a rugby player’s build who serves as Masten’s CEO, has a habit of leading new hires to the Roton ATV, a failed rocket displayed in the center of the Mojave spaceport, and telling them: “We’re not guaranteed success. We’re not trying to minimize risk or downside. We’re working for some big stuff.”

As a child growing up outside Cleveland in the 1970s, Dave Masten so loved rockets that his normally education-focused parents let him miss school to watch NASA launches on television.

Eager to experiment with his own designs, Masten would decamp to an open field next to the local elementary school with his younger brother to fly Estes model rockets they assembled from cardboard tubes and balsa wood fins. Inspired by his father, a software engineer who had a penchant for radio-controlled airplanes, Masten wondered: if he fastened wings to one of the rockets and outfitted it with radio control, could he land it like a shuttle?

Eventually he would return to rocketry, but Masten began his career in more earthbound fields. Though he ultimately left college a semester short of graduating, he paid for his mechanical engineering studies by welding for a General Motors supplier before briefly enlisting in the Army, where he learned to drive fuel tankers and despise large bureaucracies.

After moving to Silicon Valley, Masten got involved with various tech firms, including Andiamo Systems, a network hardware company, which Cisco bought in 2002 for approximately $750 million. This didn’t make him Jeff Bezos wealthy, or even Elon Musk rich. But it did give him enough money to dedicate himself to rocketry full time. Along with three others he had met at space conferences and through the Experimental Rocket Propulsion Society (ERPS), an amateur group obsessed with high-power rocketry, he founded Masten Space Systems in 2004.

From its first days in the cramped Santa Clara workshop it shared with ERPS and another company, Masten Space Systems has focused on creating reusable rockets that take off and land vertically. The partners fervently believed this approach would reduce the cost of rocket missions, making space more accessible. “We had half a million bucks,” recalls Jonathan Goff, an amiable propulsion engineer who was one of Masten’s cofounders. “We figured we’d build a flight demonstrator to convince people we knew what we were doing, raise the last of the money to go suborbital—that would take a year, maybe two at the most. And then we’ll either have enough money to go orbital, or raise money to go orbital.”

It did not work out that way. Both Masten and Goff categorize the first vehicle the company built as an abject failure. Flaws in the code meant that when the team eventually tested the craft, tethering it to a crane to keep it from crashing onto the landing pad below, it would take off and spin on the tether like a dizzy puppet.

Meanwhile, Masten and his partners soured on Santa Clara. Their workspace was cramped, and the neighbors often complained about the noise from Masten fiddling with rocket igniters in the back room. Plus, to test their rocket engines away from people, the Masten team had to drive their test trailer—a clunky steel thing they nicknamed the “hot dog stand”—two hours into the Diablo mountain range.

Encouraged by colleagues at another small rocket company, XCOR, Masten Space Systems packed up its gear and drove south. The destination was Mojave, a tiny town in Southern California’s high desert, whose long coexistence with aerospace meant the residents were “more likely to cheer than call the cops” when they heard the roar of rocket boosters, Masten remembers.

By the time Masten Space Systems moved into an old Marine motor pool maintenance building at the Mojave airport (the Marine Corps had taken it over during World War II), the complex was already well known among aerospace buffs. It was in a warehouse in Mojave that Jeana Yeager, Burt Rutan, and Dick Rutan built the Voyager, which in 1986 became the first plane to circumnavigate the world without refueling. In 2004, the Federal Aviation Administration designated the Mojave complex a “commercial spaceport.” A few days later, Burt Rutan’s SpaceShipOne took off from Mojave to become the first private vehicle to enter space with people on board.

Still, Masten’s early relationship with Mojave was ambivalent. He enjoyed the fact that—an hour and a half from Los Angeles, the closest large city—the sky was so inky black he could stand amidst the hangars and gaze up at the Milky Way. The small, dust-beaten town was less appealing. Just a mile or two from where billionaires like Richard Branson and Microsoft cofounder Paul Allen would land in their private jets to check on their rockets, many of Masten’s neighbors suffered from startling poverty and drug addiction.

Not that he saw them much. During his first months as a Mojave resident, Masten spent more nights on a cot set up in the company’s spartan office than in his nearby apartment.

He had plenty to worry about. As the team struggled to fix their unruly rocket, a team from the Discovery Channel visited Mojave to film a program on new space companies. Soon after the crew clicked on their cameras, Masten’s rocket spun out of control during a tethered test flight. The team cut its engine, but by that point it was high enough that when it dropped, the force with which the rocket pulled on the tether tipped the crane forward. The foot of the rocket caught on the launch pad and the machine ripped in two. On tape. “I was pretty sure we were done,” Goff recalls.

By that point Goff had just $50 left to his name; Masten had not taken a salary for years. They were about to close the business and go their separate ways when they received a call from Joel Scotkin, a New York–based investor, who had sold his financial technology consulting firm to Accenture. Scotkin had always been excited by the potential for private companies to transform spaceflight; despite the challenges, he was impressed by Masten’s proprietary engine designs that ran on oxygen and rubbing alcohol. In 2007 he wrote Masten Space Systems a check that, though not huge, allowed the company to push forward. “It was one of those moments where you pull the airplane up but still have grass stains on the fuselage,” Goff recalls.

Things began turning around for Masten in 2009.

In the fall of that year, the company had qualified for a NASA “Centennial Challenge” that judged teams on their ability to simulate accurate moon takeoffs and landings. The first phase of the competition required landing on a well-marked, flat circle 10 meters in diameter. Masten Space Systems entered its Xombie rocket, which won $150,000 for second place behind Armadillo Aerospace, a small Texas rocket startup.

The second part of the challenge a few weeks later carried a million-dollar prize and involved landing on a rocky, cratered surface meant to resemble the moon. Masten decided to use a larger, more powerful aluminum-frame rocket named Xoie. Difficulties with their initial design meant that they were “starting from a pile of parts” a month before the rocket had to be competition ready. They worked 80-hour weeks and tested it about 20 times over the course of several days. The night before their scheduled competition flight, Xoie managed to hover for the required three minutes despite gusts of up to 40 miles (64 km) per hour. Goff recalls thinking: “Holy crap, we’re ready!”

The next morning, though, as an eager audience looked on, Xoie refused to start. Goff slid his finger along the rocket’s engine and felt moisture: it was leaking alcohol, the liquid the vehicle used as fuel.

The company was nearly out of money again. If they lost the challenge, they would likely have to shutter. “We thought, okay, it’s a $300,000 vehicle, but it’s a million--dollar prize, and we’re going to go out of business if we fail anyway. Might as well roll the dice,” Goff recalls.

Finally, they got the engine to light and Xoie roared into the sky. Goff recalls the power of the rocket reverberating in his ribs. He watched, mesmerized, as Xoie hovered high above the Joshua trees and sagebrush of the Mojave Desert, and rejoiced when he saw it touch down in the designated landing area.

But his celebration quickly gave way to panic: no sooner did Xoie land than its oxygen tank burst into flames.

The judges decided to grant the company one more flight attempt the following day—but they would need to be ready at 5 a.m. That gave the team a little less than 12 hours to determine the cause of the fuel leak and fix their charred rocket. “Even when we got the word that we would be allowed to try again if we could repair the vehicle … we were almost ready to give up,” Scotkin wrote at the time in an email recapping the challenge. “Almost all of the MSS personnel looked like walking dead.” 

Goff hurriedly drove to a FedEx facility to pick up a replacement tank. When he returned an hour later, he found that members from other challenge teams had gathered at Masten’s workshop to help. Masten and Scotkin sent him and the rest of the weary team home to rest, and with the support of the volunteers, they got to work.

If they couldn’t get the rocket to stop leaking, mused Keith Stormo of High Expectations Rocketry, a small group from Idaho, perhaps they should build a sump to collect drips and a catheter to divert them from critical parts? Masten’s team found a Rubbermaid trash can lid, glued it on, and used baling wire to hold it tight. Overnight, Masten and a few others patched the tank insulation, repaired the faulty wires, and re-ran the damaged plumbing.

They finished everything mere minutes before their designated launch time. After the team rapidly filled Xoie’s fuel tank with isopropyl alcohol, Masten’s launch director, Ben Brockert, ordered his colleagues to run.

“We were engineers who were fat and out of shape,” Goff recalls, so he and his coworkers were still about 75 feet shy of what would be considered a safe distance from the fully fueled rocket when Brockert began counting down to blastoff.

Xoie completed two flights that lasted the requisite three minutes, and went on to win the challenge with an average landing accuracy of 7.5 inches (19 cm) from the target. The vehicle built by Armadillo Aerospace, which had previously held the lead, came in at 34 inches.

Suddenly, Masten Space Systems had earned $1 million and a reputation for punching above its weight.

In May 2010, Elon Musk sent an email to SpaceX’s propulsion, avionics, and structures teams with a link to a video hosted on a hobbyist website. You could watch Xombie take off vertically, ascend, and then pause in midair as its engine (deliberately) flickered out. As the rocket hurtled downward, the pilot re-lit the engine and the rocket gently descended to the ground. It was the first time a vertical takeoff, vertical landing rocket had ever done so. NASA called the in-air relight “a major step towards flying payloads to suborbital altitude.” In writing to his team, Musk was succinct: “Pretty cool!” Masten says with a chuckle: “I was doing it before it was cool, and now it’s cool and everyone talks about [Musk] doing it and it’s like: uh, okay. He wasn’t the first to do these things.”

Over the past five years, Masten has come close to getting several significant contracts. In 2014, the Defense Advanced Research Projects Agency (DARPA) invited it to compete against industry giants Boeing and Northrop Grumman to build a reusable experimental spaceplane called XS-1. The contract would have paid up to $140 million. “I think they went out of their way to make sure that there was a small company involved to have a chance to prove or not prove itself,” Masten reflects. But he could not hire talent or raise capital rapidly enough. DARPA chose Boeing.

While it hasn’t yet scored a giant contract, though, Masten’s ability to fly and land precisely has proved useful to NASA. Precision is one of the main challenges facing the next generation of landers for both the moon and Mars. Touching down might look easy, Mahoney says, but it’s like “balancing a broom on the tip of your finger that’s spitting fire and getting lighter at the same time.” And that’s on smooth surfaces.

As Lockheed’s Murrow explains, “The interesting places on a planetary body—we know this from going to Mars—are not always the safe, flat places.” Water, he says, is likely to be in the permanently shadowed side of a crater—so you’d want to land near enough to the crater’s edge to explore it, but not so close you fall in.

The 2020 Mars Lander, which is managed by NASA’s Jet Propulsion Laboratory (JPL), uses new precision guidance systems to avoid boulders and find smooth areas for landing. In a series of flights in 2013 and 2014, JPL tested a prototype of the lander’s vision system on Masten’s Xombie rocket. It climbed over 1,000 feet in the air before autonomously landing, guided by comparing images from a digital camera with a known map. In 2017 a newer NASA system, called Cobalt, flew on Xodiac, another Masten rocket. The precise lidar on Cobalt allows a lander to find a flat spot with even greater precision.

Last year Masten tested a sample--return device built by Honeybee Robotics, a firm located not far from JPL in Pasadena, through a NASA-run matchmaking program that pairs experimental payloads with commercial launch vehicles. The “PlanetVac” is essentially a little vacuum that replaces one foot pad of a planetary lander; it’s a simple, lightweight instrument that, if successful, might offer a cheaper and more reliable way of gathering samples.

These collaborations have burnished Masten’s reputation and provided enough cash for it to chug along as, in Masten’s words, “a small company that does a couple of basic services with NASA and the Department of Defense.”

Winning one of the coming CLPS task orders would set it on a different trajectory altogether. But CLPS is no longer the only large competition Masten is in contention for. In late May, NASA announced that six companies had been chosen to develop prototype human moon landers. Of the CLPS nine, there are two firms on the list: Lockheed Martin and Masten.

A few days after Beresheet’s crash, several Masten engineers stood around a whiteboard puzzling over an equation. On a wood laminate table by the trailer’s front door, Mahoney, who commutes to Mojave from Georgia, had deposited a stack of yellowing Atlanta newspapers from July 1969. “Astronauts Test Landing Module; Apollo Moves into Moon Gravity,” announced the Atlanta Constitution on July 19. “Men Walk on the Moon; Eagle Ready to Return,” read the Atlanta Journal.

The day was gusty, even for Mojave, where the wind has been known to topple tractor-trailers. As Dave Masten walked across the parking lot from the office to the aviary where his rockets are stowed, he had to clutch his black wire sunglasses to his head to keep them from flying off.

He poked at the electronic lock on the aviary door and stepped inside the drafty, largely empty warehouse. Walking past walls hung with tools, he affectionately patted Xombie, a 10-foot (three-meter) tangle of steel tubes, aluminum propellant tanks, and carbon-fiber-wrapped pressure tanks, as he might a well-behaved pet. The rocket had been named by several Masten interns after they devoured “every zombie movie Netflix had to offer.” (Nightlife around Mojave, a town of 6,104, is limited.) It was, in the end, well named. After all,  it has flown more times than any other rocket, with plenty of bumps and dents to prove it.

In a week and a half’s time, Masten’s team would have to submit its bid for the first CLPS task order—a detailed plan for how it would get the first bundle of cargo safely to the surface of the moon. Masten’s engineers were beginning to work longer days and gear up for the crunch that inevitably precedes a deadline. “Second-to-last week you start thinking about working 12- to 16-hour days,” Masten says. “Then the last week you get some guys thinking: ‘I’ll sleep when I’m dead.’”

The proposal was not the only thing on Masten’s mind. Another company had recently hired away his only remote pilot, and Masten would have to train a replacement. Money was a constant frustration. Since Masten Space Systems is a small company with relatively few customers, it is particularly painful when those customers—like NASA—don’t pay on time. During the government shutdown earlier this year, Masten and Mahoney had to forgo their salaries to make sure they could pay the rest of their team. Masten sighs. “I mean, that’s happened to me so many times. We’ve got lots of money,” he says, and pauses for effect. “We’ve got no cash.”

Two days after Beresheet crashed into the moon, Morris Kahn, an Israeli entrepreneur who is chairman of SpaceIL’s board, announced that his team was already planning a new mission. If shots on goal really are more important than scoring, the American government will have to respond to failures with comparable enthusiasm. Masten might gamely take risks, but will NASA?

I asked Dave Masten how he felt about the Beresheet crash; after all, it meant he still stood a chance of being at the helm of the first privately funded lunar mission.

“I felt the pain of losing a vehicle,” he replied. “Being first is not much of a motivator for me.” When I asked him to clarify what was, he responded earnestly: “Landing on it, period. You know, the whole purpose of this company was for me to step on the moon.”

Haley Cohen Gilliland is a writer based in Los Angeles.