Why Is The U.S. Space Industry So Obsessed With Vertical Integration?

Apart from passing the Karman line and reaching outer space, there are few boasts in the U.S. space sector prouder than “my company is vertically integrated.”

Vertical integration—the practice of developing and building systems in-house instead of buying from a supplier—is rare in other manufacturing industries, such as consumer electronics or automobiles. However, in the U.S. space business, the logic of do-it-yourself manufacturing is an article of faith, rarely questioned or doubted in public.

• Origins of the strategy go back to post-Cold War space sector consolidation
• Elon Musk’s SpaceX is seen as the foremost practitioner of in-house manufacturing

In the minds of many, the foremost practitioner of vertical integration is Elon Musk’s SpaceX, a pioneering launch vehicle and internet satellite company that, according to lore, made itself out of nothing. Startups looking to make their mark on the space industry often adopt a vertical integration strategy and cite SpaceX as their inspiration.

But vertical integration was not SpaceX’s original plan. In 2001-02, Musk, flush with cash from selling PayPal, tried to buy Russian Dnepr rockets—converted from ICBMs—to reach space on the cheap. Rebuffed by the Russians and unhappy with what U.S. suppliers were charging, Musk hired engineers to develop SpaceX’s first liquid-fueled rocket, the Falcon 1.

That small launch vehicle would feature a booster rocket motor, the Merlin engine, a rocket body, flight computers and many other components designed in-house. SpaceX went on to develop its workhorse partially reusable medium-lift launch vehicle, the Falcon 9, and a penchant for doing most of its development work in-house.

SpaceX now designs everything internally, from its rocket motors and laser communications systems to its satellites and spacesuits. The company also owns its Starbase launch site in Boca Chica, Texas. It operates recovery ships, runs its own flight control centers and is building an astronaut corps.

Other prominent U.S. space companies—such as Blue Origin, Rocket Lab, Relativity Space, Terran Orbital, Planet Labs and Boeing’s Millennium Space Systems—boast a vertical integration strategy as well.

The conditions that produced the U.S. space sector’s obsession with vertical integration started far before the founding of SpaceX and other “New Space” startups, however. In 1993, after the fall of the Soviet Union, then-U.S. Defense Secretary Les Aspin convened major defense and space contractors at the so-called Last Supper and told them that without a major adversary, Washington would drastically reduce its defense spending. There would not be enough funding to go around, so contractors would need to merge or acquire to stay afloat.

In the decades since, waves of mergers and acquisitions have produced corporations with two first names, such as Lockheed Martin, Northrop Grumman and Aerojet Rocketdyne. Below the prime manufacturers, consolidation merged first-, second- and third-tier suppliers specializing in niche components. In the post-Cold War shuffle, a lot was lost.

Rob Meyerson, who was Blue Origin’s president from 2003 to 2018, says that when the upstart launch service provider went to U.S. suppliers looking to buy parts, there was not much left.

“A lot of these waves of acquisition among the supply chain resulted in companies that own product lines, [for instance], valves that they’d never built before,” he says. “They bought a company that used to, 20 years ago, make a 10-in. valve for liquid hydrogen service, for example. And you’d say, ‘OK, that’s in your catalog, I want to buy those.’ And they would say, ‘Oh, we’ve never made that before. We need to charge you X amount of dollars for nonrecurring [engineering] to go requalify our capability to go build it.’”

Forking over cash for a supplier to learn on the job was a bit of a head-scratcher. “You end up paying these companies to get smart and develop capabilities that they used to have,” says Meyerson, who now is CEO of Interlune, a Moon mining startup. “If you have the money, you can make a decision to just do it yourself and vertically integrate.”

The situation was compounded not long after the U.S. space industrial base had reached its zenith of consolidation—perhaps around 2013, when the country’s two remaining large rocket motor manufacturers merged to form Aerojet Rocketdyne. That was when Washington started to worry about its reliance on Russian rockets. NASA used Soyuz launch vehicles to reach the International Space Station, and the Pentagon relied on NPO Energomash RD-180 rocket engines to power the United Launch Alliance (ULA) Atlas V, used for top-secret national security missions.

Concerns on Capitol Hill went into overdrive after Russia invaded Ukraine’s Crimea Peninsula in 2014. The National Defense Authorization Act for Fiscal Year 2015 included provisions to phase out RD-180 engines in national security launches by 2022.

Financed by billionaires, venture capitalists, the U.S. Treasury, low interest rates and eager customers, companies founded in the last decade such as SpaceX, Blue Origin and Rocket Lab went on to execute a do-it-yourself reconstruction of the U.S. space launch sector. Along the way, they gave new meaning to the maxim “necessity is the mother of invention.” Facing reluctance from metal-casting behemoths like Precision Castparts and Arconic to do small-dollar development work for rocket startups, many companies started looking into alternative technologies, including greater use of 3D printing, Meyerson says.

Because rocket and satellite startups of the past decade sought to make big innovative leaps—reusable rockets and low-Earth-orbit satellite constellations, for example—the learning curve was steep.

“I never worked on one rocket my entire time at SpaceX that was the same rocket from the one that launched previously,” says Robert Pakalski, CEO of Datum Source, a procurement software-as-service company. He had worked on new product procurement initiatives for SpaceX, including the thrust structure for the Falcon 9 and Falcon Heavy, and on the Dragon spacecraft. “There’s always different hardware; there’s always improvements going on,” he adds.

Academic research has shown that coordinating product changes with outside vendors can slow down the fail-fix engineering cycle that drives innovation. That is especially true with a product as complex and tightly integrated as a space launch vehicle. There are advantages to controlling more of the process.

“[SpaceX’s] ability to understand every piece and component in their system in detail is built on the fact that they design and build most of what’s in that rocket,” Meyerson says. “They are so hardware-rich. They do a lot of testing. They do really fast analysis, and their flight rate [is rapid]. You’re seeing that right now. [If] you’re flying a lot, you’re incorporating lessons learned from those flights.”

To be sure, SpaceX does not do everything in-house. “SpaceX is entirely vertically integrated as a subassembly integrator,” Pakalski says. “But from the actual individual components that make up the rocket or spacecraft, the majority of that is outsourced versus vertically integrated.”

The company designs most of its components instead of relying on off-the-shelf parts, but those pieces are typically sent to the mom-and-pop job shop ecosystem before returning for SpaceX to put together the subassembly and final product, he explains. SpaceX has a robust internal manufacturing capacity, but its purpose is not necessarily to make every individual part.

“SpaceX has an absolute killer CNC machine shop,” Pakalski says. “It’s probably one of the best in the country, but it’s not an individual business. It’s there to take on the high-[intellectual property] parts and the very critical parts.” The company outsources tens of thousands of parts to suppliers, he adds.

Many small-satellite manufacturers have also chosen vertical integration after discovering that what was available on the market was expensive and tailored for larger spacecraft of legacy manufacturers, Boston Consulting Group Managing Director Matt Martinez says. Some have designed and assembled spacecraft subsystems such as momentum wheels with commercial off-the-shelf electronics, he notes.

“People aren’t going to try to design resistors,” Martinez says. “But there are folks out there saying, ‘Hey, I want to make pretty simple printed circuit boards myself, and I’m not going to outsource those because I can do automated quality control and assembly.’”

The U.S. government also pressures companies to maintain greater cybersecurity control over their supply chains, which can lead to a vertical integration strategy, says Lori Gordon, systems director of The Aerospace Corp.’s Space Enterprise Evolution Directorate.

“Security is very expensive. When you’re horizontally integrating, you have to look at the supplier viability across so many companies, and that’s expensive,” she says. “Companies want to ratchet down, bring everything in-house—bring the skills and capabilities within their virtual walls.”

Still, vertical integration has drawbacks, including high up-front costs. Small-satellite bus manufacturer Terran Orbital has spent heavily on new factories and manufacturing equipment. The company says it makes 85% of components internally, including printed circuit boards, reaction wheels and torque rods, but is also losing tens of millions of dollars per quarter. It is not clear that a return on investment will come anytime soon.

Finding workers with the right know-how to in-source a component can be challenging, and the knowledge of how to clear regulatory barriers on particular systems can be hard to accumulate, Martinez says.

In recent years, several companies have bucked the vertical integration trend and set themselves up as suppliers. Redwire Space has established itself as a Tier 1 supplier, manufacturing everything from solar arrays and cameras to star trackers and radios. In March, SpaceX started selling its laser communications systems to outside customers, the first time it has sold a product directly to another business. Rocket Lab has expanded beyond launch vehicles into building satellites buses and solar panels.

Many of these New Space suppliers were launched by former SpaceX engineers who are using their expertise to establish niche businesses. For example, Ursa Major has focused on manufacturing rocket engines instead of entire launch systems, deeming propulsion systems so complex that the company could gain an advantage by specializing.

“When you look at propulsion, it’s the heart and soul of a rocket,” Ursa Major Chief Operating Officer Nick Doucette says. “It’s most of the cost, most of the complexity, most of the failures.” By pooling into one company engineers with experience building—and blowing up—rocket motors, the startup can develop propulsion systems for clients more rapidly, he says. “We can shortcut a lot of those bad decisions and make those timelines go really quickly,” Doucette adds.

Apex Space, a startup that in June raised $95 million for its Series B round, believes now is a good time to become a space supplier and to support horizontal integration. Apex says its satellite buses are designed to integrate easily with other companies’ propulsion, sensors and subsystems.

“We very much are of the belief that the space industry has finally moved to a period where vertical integration is not the most efficient anymore,” Apex Space CEO Ian Cinnamon says. “That recent shift is, in our opinion, due to the decrease in launch costs and the increase in launch cadence.”

The fall in launch costs—primarily driven by SpaceX’s partially reusable Falcon 9 rocket—means that satellites no longer need to be engineered to optimize their mass, he notes.

“Instead, the No. 1 goal is to get your satellite up faster, because you’re probably monetizing or solving your kind of use case that you need to solve more quickly,” Cinnamon says.

There are a variety of ways to work faster with outside companies, including locating an employee at a supplier’s site, Meyerson notes.

“Vertical integration, it’s not obvious that is the better choice when I look at it, even after having been in it for 25 years,” he says.