Firefly is building fast and breaking things on path to a reusable rocket

BRIGGS, Texas—The new medium-lift rocket under development by Firefly Aerospace and Northrop Grumman will eventually incorporate a recoverable booster that will return to its launch site in Virginia for reuse.

Firefly has previously suggested rocket reuse is on the roadmap for the new rocket—known, for now, only as the Medium Launch Vehicle (MLV)—but officials revealed new details of the plan during a recent visit by Ars to Firefly's rocket factory in rural Central Texas.

“Northrop and Firefly have a similar perspective and that is, for that class of rocket, reusability is a requirement for a bunch of reasons," said Bill Weber, Firefly's CEO. "Economically, it becomes an advantage because we don't have to go build additional floor space... Similarly, the pricing structure for customers starts to get super competitive, which we absolutely love, and we’ll be right in the middle of.”

Firefly is one of several companies racing to field new medium-class rockets, and all will be at least partially reusable. Rocket Lab, perhaps the most dominant company in Firefly's class, is developing the Neutron rocket as it continues flying the smaller Electron launcher, which now has amassed 50 missions. Relativity Space, a well-funded private company based in California, is developing the partially reusable Terran R rocket after abandoning its smaller Terran 1 vehicle following just a single test flight. Stoke Space is working on a novel rocket design with a reusable booster and upper stage.

All of these launchers are sized to compete with SpaceX's Falcon 9 rocket, the current market leader. They will eventually join the US military's roster of launch providers for national security missions, which currently only includes SpaceX, United Launch Alliance, and, most recently, Blue Origin.

Necessary to compete

Until now, Firefly has provided few details of its rocket reuse roadmap. But details revealed to Ars show the MLV will employ a familiar method of recovery.

"For our base design, we're designing around return to launch site propulsive landing," said Merritt D'Elia, manager of propulsion for the MLV program. "We'll iterate through all of these things, but we're fundamentally architecting for reusability. Not just architecting, we're planning to do it."

Firefly may decide to include an option for downrange landings on barges at sea, as SpaceX does with Falcon 9 and Blue Origin plans to do with the New Glenn rocket. Rocket Lab and Relativity also plan to perform downrange rocket landings. However, D'Elia said that approach is expensive, requiring upkeep of ships, and delays the return of boosters to the launch site for refurbishment.

MLV's booster recovery technology testing will begin with the rocket's first flight, according to D'Elia, when Firefly will fly control thrusters to demonstrate the first stage's flip maneuver to return to the launch site after separation from MLV's upper stage.

Weber, Firefly's chief executive, said the company's goal is to recover a full MLV booster and reuse it by about the sixth flight of the rocket. "As it stands right now, it'll be somewhere around Flight 6," he said. "That is likely when it works well to introduce that capability into flight."

On the ground, Firefly is designing Miranda engines to be able to perform multiple burns on a single flight, a capability it will need for propulsive landings. Engineers are testing MLV's composite structures to ensure they can withstand multiple launches and landings, including the heat of reentry back into the atmosphere.

"To meet launch cadence, to go as fast as we need to go, to do it at costs that make sense, and to do it in a way that we’re not doing more damage to the planet along the way, I don’t know how you can achieve that without reusability," Weber said.

In some ways, MLV is an upscaled version of Alpha. The new rocket will be powered by seven Miranda engines on the first stage, each generating 230,000 pounds of thrust at full power, using the same tap-off engine cycle Firefly uses for the smaller Reaver and Lightning engines on the Alpha rocket. And MLV's structures will be made of carbon composite materials, just like Alpha.

But MLV will be more than twice the diameter of Alpha, requiring new tooling on Firefly's production floor. Firefly more than doubled the size of its engine factory to accommodate production of the larger Miranda engines. MLV will be able to deploy payloads of up to 16 metric tons (35,000 pounds) into low-Earth orbit when its booster flies in expendable mode.

The two-stage Medium Launch Vehicle will stand nearly 184 feet (55.7 meters) tall, with seven Miranda engines generating more than 1.6 million pounds of thrust at peak power.

Credit: Firefly Aerospace

Nearly two years ago, Firefly Aerospace and Northrop Grumman entered into a partnership to jointly develop the Medium Launch Vehicle. For Firefly, this collaboration opened a path to growth to pursue more lucrative parts of the launch market for heavier satellites. In Northrop Grumman's case, this was a partnership of necessity as the supply of Russian rocket engines for its Antares rocket dried up after Russia's invasion of Ukraine. Northrop flew its last Antares rocket with Russian engines last August to launch a Cygnus supply ship to the International Space Station.

Without the Antares, Northrop Grumman has turned to SpaceX to launch Cygnus missions on Falcon 9 rockets until at least next year. Firefly's initial focus is on developing the Miranda engine and the MLV's first stage, which will be combined with the Antares rocket's solid-fueled upper stage to enable a faster resumption of Antares launches. This rocket, called the Antares 330, will fly three times to complete Northrop Grumman's cargo contract with NASA.

Firefly is moving fast to deliver the first booster stage for the Antares 330 to Northrop Grumman early next year, according to Weber. Once the booster is transported to the Antares launch site at Wallops Island, Virginia, technicians will complete assembly and install the rocket's upper stage. Antares 330 is scheduled to make its first launch in late 2025 or early 2026, Weber said.

While this is happening, Firefly plans to develop a brand-new upper stage with a single engine named Vira, based on the Miranda design. Combining the new booster stage with the upper stage will create the full Medium Launch Vehicle. Northrop Grumman is partially funding the MLV development program, and the aerospace giant will marry its flight-proven avionics and software with Firefly's propulsion and carbon composite propellant tanks.

All in one field

A half-dozen test stands dot the landscape on a more than 200-acre prairie about an hour's drive north of Austin. Here, Firefly Aerospace builds rockets and puts hardware through stress testing before heading to orbit.

Firefly has operated here for nearly a decade, staying alive through bankruptcy, a lawsuit over trade secrets, and the divestment of the company's majority shareholder, Ukrainian businessman Max Polyakov, at the behest of the US government.

Despite the upheaval, Firefly has risen to a position among the top few of dozens of startup rocket companies established in the wake of SpaceX's dominance of the US launch industry. Now, Firefly is branching out into commercial lunar landers and orbital transfer vehicles, while its first main product, the Alpha small satellite launcher, finds its footing after four flights.

Alpha is designed to haul up to a metric ton of payload into low-Earth orbit, about three times more than Rocket Lab's Electron. The fifth launch of Firefly's Alpha rocket could take off as soon as Monday night from Vandenberg Space Force Base in California with a batch of eight small satellites sponsored by NASA.

Firefly's first full-length Miranda engine, serial number MD-2, goes on the test stand in Briggs, Texas.

Credit: Firefly Aerospace

Firefly continues to iron out issues with the Alpha rocket. After Monday night's launch, Firefly will conduct demonstrations with the upper stage in orbit to verify that changes made since the last launch in December fixed a problem that prevented the restart of the upper stage engine. This capability isn't required to deliver NASA's payloads into orbit Monday night but is necessary for future Alpha flights.

There are dozens more Alpha launches in Firefly's order book, a backlog recently propped up by Lockheed Martin's purchase last month of up to 25 Alpha flights to loft small satellites. All of Firefly's launches to date have departed from California, but the company announced last week that it will also base Alpha launches from Wallops Island, Virginia, using the same pad as the Antares 330 and MLV. For now, at least, Firefly is deferring construction of a launch site at Cape Canaveral, Florida.

Firefly engineers continue tweaking Alpha's design to improve its reliability and efficiency, but most of the floor space inside the company's rocket factory is now devoted to the larger MLV.

Culture shift

The first handful of Miranda engines are now in production, and Firefly has completed nearly 20 test-firings of Miranda engines at the company's "Rocket Ranch" here in Briggs. In May, one of Firefly's first Miranda engines blew itself apart during a hot-fire test, an unsurprising outcome in early-stage full-scale testing of a new rocket engine. It had fired successfully more than 10 times before then.

"This was a destructive anomaly," D'Elia said, showing off mangled engine parts on a shop floor. "This chamber got thrown out in a field and landed in the mud. That’s part of the process. Through all of our programs, the times when the vehicle components, engines, or structures take a step change in maturity is always around an anomaly because you learn so much about your thing by having it fail, by figuring out why it failed, and then making sure that never happens again. And the item—the tank or the engine—takes a step change in maturity.”

"They’re often unplanned, but they’re often educational," said Jeff Duncan, Firefly's director of mechanical engineering. "This was a spectacular anomaly.”

In the six weeks since the engine blew up, Firefly engineers found the cause and implemented fixes in the next Miranda test engine and put it on the stand to kick off another series of hot-fire tests. It takes less than five minutes to walk from Firefly's engine factory to the test stand, so engineers inspect one engine after a test-firing and then quickly change how the next one is being built.

"These are the times when you learn how the engine really works because you have to figure out how the engine failed," D'Elia said. "Nobody likes to destroy hardware. We're not ever in the business of recklessly putting hardware in harm's way" ("We do sometimes!" Duncan interjects).

A composite production test article for the first stage of the Medium Launch Vehicle and Antares 330 rockets.

Credit: Firefly Aerospace

The Miranda engine that failed on the test stand was a "franken-engine" of sorts. Leading up to the explosion, engineers swapped out the engine's initial combustion chamber for another one with improvements to better withstand the super-hot gas generated when kerosene and liquid oxygen propellants mix and ignite. This is design iteration in practice.

"On the first chamber, you could see erosion, and you could see wear on it. Maybe it's not a dramatic anomaly, but you say, that's not right, you tune it, you evolve it, and this one doesn't have any—it looks great," Duncan said, inspecting the Miranda engine's remains.

"Except it blew up," D'Elia joked. "But that's the idea. These are development engines, and if you're not learning something from your development engine, you're wasting your time."

Sometimes, Firefly's comfort with test failures has clashed with the conservative culture of Northrop Grumman, one of the largest US defense contractors. But a visit to Firefly's engine shop allayed those concerns.

"If you never destroy anything, your engine is too heavy," D'Elia said. "We are trying to build a really efficient, high-performance thing here, and you’ve got to figure out where the line is to build that. If you just build a battleship that never fails in development, then you will build an inherently inefficient device.

"We don’t ever want it to fail once we’ve said, 'This is the design," D'Elia said. "But we’ve got to figure out where is that line, so we can go right up to it, but not cross it."