Rocketry is making leaps forward now, and companies are experimenting with a variety of new launch techniques. Yet Relativity Space may well outdo them all, ushering in a rocket manufacturing revolution that may end up being about far more than simply access to space.
Relativity Space’s revolution? 3D printing entire rockets. Nearly everything in their rockets, from the tip of the nose to the engine nozzles, comes out of a 3D printer. If it works, it could promise to be revolutionary.
The challenges and opportunity of metal 3D printing
3D printing metal (also known as “additive” manufacture) can be both difficult and dangerous. It can disperse toxic powders and fumes into the environment. The process can be slow compared to regular metal manufacture, involving detailed follow-up machining to ensure parts are finished and within tolerances.
It is also extremely hot compared to plastic printing, achieving temperatures over 2500 °F degrees (1371 °C). The repetitive heating and cooling of the printing process can create stresses, cracks, and voids within the metal body. Manufacturers need to perform internal inspections before the metals can be used for sensitive tasks … and almost every task is sensitive when it comes to rocket launches.
Still, interest in additive manufacturing endures, as creators chase the dream of creating an infinite variety of on-demand, custom-made parts. The space sector has had some successes as well: both Jeff Bezos’ Blue Origin and Elon Musk’s SpaceX use small-scale metal 3D printing to create some specialized parts.
Two college friends found Relativity Space
For Relativity Space founders Tim Ellis and Jordan Noone, though, that just wasn’t enough.
Ellis and Noone met in college, and both ended up in the space business: Ellis went to Blue Origin, and Noone went to SpaceX. Both were convinced that a strong investment in 3D metal printing would demonstrate its viability, agility, and short lead-time. Higher-ups supported and encouraged their efforts, and they had some successes. But, as Jordan Noone said in an interview with SpaceQ, it became clear that the two larger companies were too committed to traditional supply chains, processes and approaches.
That’s when they started Relativity Space. Noone and Ellis realized that in order to succeed, they needed to commit entirely to 3D printing rockets, and only a startup could be agile and enough to risk it.
Noone pointed out that it was tremendously different from traditional rocket manufacture. Truly taking advantage of 3D printing in rockets required both a revolutionary approach and an integrative approach, one that combined entirely new approaches to both 3D printing and rocketry. They had no choice: as Noonan said, while “other companies develop printers and rockets, doing both changes how you build the part, how you inspect the part, and how it’s ready to fly.”
It’s new territory.
Noone said that this made staffing up very tricky. They needed aerospace expertise, but they also needed metal 3D printing expertise. They needed to acquire and onboard talent quickly, but almost nobody has knowledge of both fields. Noone said that it was a unique challenge that meant that Relativity Space had to hire and train a large cohort of both aerospace and non-aerospace talent. They’ve succeeded in bringing on top-quality people, like SpaceX’s Zachary Dunn, but since they’re exploring new technologies in two fields at once, everybody is new. Everybody is learning.
Yet while staffing has been an issue, funding hasn’t. While early investors were skeptical, Relativity Space’s promise (as well as its founders’ experience) landed the company a coveted spot in the Y Combinator program. As they graduated Y Combinator, developed their company, and hit their milestones, they’ve attracted over US$185 million in investment — including a stunning US$140 million Series “C” funding round led by Bond and Tribe Capital. They’ve also attracted the support of NASA, which is letting them use the Stennis Space Centre in Mississippi for testing, and Launch Complex 16 at Cape Canaveral in Florida. There’s tremendous interest in their project.
Relativity Space’s innovations: Stargate, Terran and Aeon
Out of this whole process came Relativity’s three key projects: the “Stargate” 3D printer, the “Terran” rocket, and the “Aeon” engine. (All three come from the computer game “Starcraft”, which is a favourite of Ellis’.)
The “Stargate” is the 3D printer that manufactures their rockets. The world’s largest 3D printer, it was entirely designed and built by Relativity Space, and fills up a whole room in Relativity’s factory in Los Angeles. It doesn’t look like an advanced manufacturing facility. If anything, it just looks like a set of robot arms over a rotating base. One of the arms liquifies Relativity’s proprietary aluminum-based metal alloy, and deposits it one layer at a time on the spinning base. The other arms support the process: polishing the metal, removing imperfections in the surfaces, and inspecting the material to ensure that it doesn’t have any of the cracks, stresses and voids that have been an issue in 3D metal previous printing. It all happens without human intervention.
Stargate is critical in the production of their “Terran 1” rocket. The Terran 1 is Relativity Space’s signature product, a rocket with a nominal payload of 900 kilograms to low earth orbit. Relativity boasts that their rocket is “among the most cost-effective launchers in the world”, and they aren’t wrong: a dedicated mission is only US$12 million, which is lower than the cost of launching with a company like Blue Origin or SpaceX.
Their promise of “raw material to flight in less than 60 days” is even more dramatic, but their “Aeon 1” engine proves it’s possible. Since Relativity Space doesn’t need to acquire and combine mass-produced parts from faraway manufacturing facilities, and since retooling isn’t a factor, they’ve been able to quickly iterate on the process of reducing the number of rocket and rocket engine component.
This process has had striking results. A traditional rocket — and especially its engine — consists of thousands upon thousands of carefully manufactured and assembled parts. Relativity Space’s promotional material says that the Terran 1 rocket consists of only 730 parts, and the Aeon 1 engine has only 100 parts. Yet in our interview, Noone said that further refinements have meant that “we’re already lower than that,” giving the example of the single-piece 3D printed injectors they used to demonstrate the project’s viability.
“Manufacturing as software”
Noone said that none of this would be possible if it weren’t for a key distinction: the Stargate learns.
Those robot arms that produce and inspect and repair and finish the Terran 1 — as well as the smaller 3D printers that produce many of the Aeon parts— use machine learning extensively at every stage of the process. Noone said everything that Relativity Space does is fed back into their machine learning algorithms.
Noone and his team are learning a lot about additive manufacturing, but it’s the printers themselves that are learning how to print, inspect, and verify parts more effectively and more quickly every day. So even the problems and errors that plague 3D metal printing are actually helping Relativity Space. Every issue is just more data for the AI to learn from.
Huge potential for launches and spinoffs
This could have dramatic effects, then, on manufacturing as well as space.
Yes, Relativity’s inexpensive launch capabilities are attracting clients. While Noone was not at liberty to discuss all of their potential clients, he did mention that Canada’s own Telesat is betting on Relativity Space to launch a portion of their LEO broadband satellite constellation. Relativity had made that announcement public in April of 2019 with Telesat’s approval.
As Noone pointed out, though, it’s much bigger than that. Nobody has ever had the resources or opportunity to learn this much about additive manufacturing. “Manufacturing as software” means that creating a single part or product is as economical as a million-part run; even creating new parts only waits on Relativity’s engineers and Stargate AI’s to sort out the optimal way to produce them. Iteration becomes as fast for hardware as it is for software. While Noone didn’t get into details, this could change manufacturing in ways that go far beyond rockets.
It’s also incredibly scalable. While only one Stargate currently exists, any new Stargate will be as capable as the first. A Stargate in their Mississippi facility has the same capabilities as the one in Los Angeles; and any future Stargates in any future factories would as well. A Stargate (with accompanying small printers) in Jakarta would be as capable as the one in Los Angeles. Terran rockets, with Aeon engines, could be built and launched almost anywhere.
And, as Noone made clear, one destination was always in mind.
Noone and Ellis always had their eyes on Mars, even during their time at Blue Origin and SpaceX.
When asked, Noone said that he didn’t believe that Relativity had all the answers, or that they could do it all themselves. But they do have their role to play. They want to send the printers to Mars as their contribution to the effort, along with enough feedstock to allow the printers to operate until local materials can be found. Someday soon, Stargates could be a standard part of humanity’s presence on Mars, fabricating parts, tools, vehicles, and perhaps even housing.
Yet, far sooner, they could end up being an important part of life right here on Earth. After all, Amazon started as a bookseller. If Noone and Ellis really do overcome the challenges and make metal 3D printing practical, there may be no limit to what their company could accomplish.