AST SpaceMobile Signs $1.3 Billion Deal With Firefly to Build Its Own Rocket

AST SpaceMobile is paying Firefly Aerospace more than $1.3 billion to build its own rocket. The agreement, announced November 6, gives AST a perpetual license to manufacture Firefly’s Miranda and Vira engines, plus the engines themselves for the first three flights of a new vehicle called Blacklift. It is a satellite company deciding that owning its launcher is worth more than renting one.

Blacklift is described as a 20-ton, reusable rocket built to carry up to five of AST’s BlueBird satellites per launch. First flights are targeted for late 2028. Until then, AST keeps flying its satellites the way it always has: on Falcon 9 and New Glenn, bought from someone else.

Why a Satellite Operator Wants to Build Rockets

AST SpaceMobile’s entire business is built around BlueBird, a direct-to-cell satellite with a phased-array antenna around 64 square meters across, large enough to talk to unmodified phones on the ground without any extra hardware. That size is also the constraint. BlueBirds are too big to fly in bulk on most rockets, which caps how fast AST can scale its constellation no matter how much capital it raises.

The risk of depending on someone else’s rocket showed up directly in April 2026, when AST lost BlueBird 7 after a New Glenn upper-stage anomaly left the satellite in an unusable orbit. AST expects to recover the hardware cost through insurance, but insurance does not rebuild a constellation on schedule. More than 100 additional BlueBirds are still contracted to launch on external providers through 2028, so Blacklift does not replace that dependency soon. It is a hedge against still needing someone else’s rocket a decade from now.

The stakes are also just bigger than they were when AST first signed its Falcon 9 and New Glenn contracts. BlueBird satellites are in production through BlueBird 51 at AST’s Midland, Texas facility, and the company’s direct-to-cell business model depends on getting a constellation large enough into orbit fast enough that unmodified phones get continuous coverage, not occasional passes. Every satellite that sits on the ground waiting for a launch slot is a satellite not earning subscriber revenue.

What $1.3 Billion Actually Buys

The deal breaks into three pieces, and the split matters more than the headline number. About $800 million, split between$400 million in cash and 10 to 14 million shares of AST stock, buys a perpetual license: once the engines are delivered, AST can manufacture Miranda and Vira itself, in a new engine facility it plans to build in the United States. A separate $250 million covers Firefly actually building and delivering engines for Blacklift's first three rockets. The last$250 million, spread over three years, pays for Firefly’s engineers to help AST actually integrate the thing.

That structure tells you what AST is really buying. Not just engines. The right to stop buying them.

The Engine Doing the Heavy Lifting

Miranda is not a clean-sheet design built for this deal. It is a scaled-up version of Firefly’s flight-proven Reaver and Lightning engines, using the same injector design and the same patented tap-off cycle, where hot gas is bled from the main combustion chamber to drive the turbopump instead of adding a separate gas generator. Each Miranda produces 230,000 lbf of thrust. Firefly has already run it through more than 175 hot fire tests totaling over 4,200 seconds of burn time, including at least one flight-like test at 110 percent of rated performance.

The same engine already powers two other Firefly vehicles, Antares 330 and Firefly’s own Medium Launch Vehicle, where seven Miranda engines combine for 1.6 million lbf on the first stage. AST is not funding an unproven engine from scratch. It is buying into hardware with a test record and two other customers already committed to flying it, which is exactly what a satellite company with no rocket-building history needs if it wants an engine Blacklift can trust on its first flights. Thrust alone does not determine how much mass an engine can send to orbit; that comes down to specific impulse and the rocket equation, the real reason a kerosene-fueled engine like Miranda trades some efficiency for the thrust density a reusable first stage needs.

What Blacklift Still Has to Prove

Vira, the vacuum-optimized variant of Miranda, will power Blacklift’s upper stage. Beyond the 20-ton, reusable, five-BlueBird-per-launch description, AST has not released Blacklift’s actual payload mass to orbit, which makes it hard to size against Falcon 9’s 22,800 kg to low Earth orbit or New Glenn’s 45,000 kg. Firefly CEO Jason Kim called the engine selection the hardest, highest-stakes decision a company makes when it sets out to build its own rocket. AST president Scott Wisniewski framed the payoff as compressing what is normally a five-to-ten-year rocket development cycle, by starting from an engine that already has a flight history instead of starting from nothing.

AST will detail more of the vehicle at a public unveiling event in New York on November 12.

A Launch Market Where Owning the Rocket Is the Point

Blacklift fits a pattern that goes beyond AST. The constellation operators moving fastest, SpaceX chief among them, are the ones that do not have to negotiate for a launch slot shaped around someone else’s manifest. A satellite built to a rocket’s constraints is cheaper to launch but harder to scale. A rocket built to a satellite’s constraints is the opposite trade, and it only pays off once a constellation is large enough to fill the vehicle on every flight.

AST is betting that BlueBird’s unusual size, the very thing that makes it hard to launch today, is exactly why owning a rocket sized to match will matter more than the $1.3 billion it costs to get there. Blacklift will not fly for two more years. Whether that bet was worth making depends entirely on whether Miranda’s ground test record holds up on a vehicle that, as of November 2026, does not yet exist.