Is AST SpaceMobile Running the SpaceX Playbook in Reverse?

Key Highlights

• We read Blacklift as the SpaceX model run in reverse: payload-first vertical integration, aimed beyond 2027 at the post-2028 launch crunch.
• AST SpaceMobile will build Blacklift, a reusable rocket rated at roughly 20 tons to low Earth orbit, designed to carry up to five BlueBird satellites per launch starting in late 2028.
• AST's agreement with Firefly Aerospace is valued at more than $1.3 billion and covers 27 Miranda and nine Vira engines for the first three rockets, three years of engineering services, and a perpetual license to build both engines at a new US facility.
• Firefly reports more than 175 Miranda hot-fire tests totaling over 4,200 seconds, equivalent to roughly 20 full first-stage flights, though neither Miranda nor Vira has yet flown.
• The announcement follows a year in which Vulcan was grounded for nine months and New Glenn lost a BlueBird and then its pad, while conflicting industry reports have SpaceX reducing/halting Falcon 9 bookings beyond 2028.

The News

AST SpaceMobile announced Blacklift on 6 November and unveiled the vehicle at an event in New York on 12 November. The rocket is a purpose-built, reusable launcher powered by Firefly Aerospace's Miranda and Vira engines, intended to give AST regular, unconstrained access to orbit and better ownership economics after a year of third-party launch disruptions. Blacklift is designed to complement, not replace, AST's existing launch agreements, which cover more than 100 BlueBirds to be deployed across 2027 and 2028, and it would complete the company's end-to-end vertical integration. AST SpaceMobile announcement (Form 8-K, Exhibit 99.1)

Analyst Take

The launch market spent 2026 reminding satellite operators who actually sets their schedules. ULA's Vulcan has been grounded since a February solid booster nozzle anomaly, though its core stage compensated and the mission reached orbit, and ULA targets a return to flight on 13 November. Blue Origin's New Glenn, which delivered NASA's ESCAPADE probes and landed its booster in November 2025, left AST's BlueBird 7 in an orbit too low to sustain in April. Blue Origin then lost its pad in a May hotfire accident and has moved its return to January. SpaceX has reportedly stopped taking Falcon 9 reservations beyond 2028, according to TechSpot, and Elon Musk has said on X that the shift away from Falcon depends on Starship flying reliably several times a week. The booking cutoff is clear. The date Falcon 9 actually retires is not, because it depends on Starship. Whether Starship succeeds or stalls, it looks to us more closely attuned to SpaceX's own ambitions across communications, orbital compute and human spaceflight than to serving AST. Note: Elon Musk controls a post-IPO SpaceX and Starlink, and Jeff Bezos owns Blue Origin and founded Amazon, whose Leo constellation competes for the same customers.

Into that gap, AST SpaceMobile is building its own rocket. Our first read is that this is the SpaceX model run in reverse. SpaceX and Rocket Lab built launchers, then built (or bought) spacecraft businesses to fill the manifest and build cadence for scale. AST built the satellite factory first and is now building the launcher that factory needs.

The skeptical case deserves a full hearing. AST's third-party agreements already cover more than 100 BlueBirds across 2027 and 2028, and neither Firefly Miranda nor Vira has flown. We remember the struggles Amazon Leo (nee Kuiper) had while waiting on United Launch Alliance Vulcan, which relied upon Blue Origin untried engines. History also may make the AST plan ambitious. According to Ars Technica, Falcon 9 took 4.7 years from announcement to orbit, the fastest of any US medium-lift vehicle, and Vulcan took 8.7. Blacklift's late 2028 target implies roughly 2. Add FAA launch licensing, range scheduling, and the cost of staffing an engine factory with propulsion talent that every new launcher is competing for. On that math, Blacklift arrives after the problem it was announced to solve, if it arrives on time at all.

We think that reading misses the calendar. Blacklift appears aimed at the post-2028 window, when the Falcon 9 commercial manifest is expected to thin and AST's government and carrier demand is expected to compound. In our experience on the buy side, a credible make-versus-buy option is one of the few levers that moves a single-source supplier, though whether it moves launch pricing before first flight remains to be seen.

What Was Announced

AST's release describes Blacklift as a roughly 20-ton first-stage-reusable rocket carrying up to five stacked BlueBirds per flight, with first launch targeted for late 2028. At the 12 November event, AST added the architecture: three stages, all-composite construction, nine Miranda engines on the first stage, two Vira engines on the second, and one Vira on the third. AST's event video cited a design cadence of up to two launches a month. Abel Avellan, AST's founder, chairman, and CEO, separately framed the goal as at least 72 satellites a year, which implies closer to 15 launches annually. He described first-stage reuse as a minimum viable product requirement.

The propulsion deal with Firefly Aerospace is valued at more than $1.3 billion. AST pays roughly $800 million up front for a perpetual license to manufacture Miranda and Vira ($400 million in cash plus 10 to 14 million Class A shares), $250 million for the engines on the first three rockets, and $250 million for three years of engineering services. The delivery covers 27 Miranda and nine Vira engines, which by our math is nine and three per vehicle, consistent with the stage layout. It should be acknowledged that Firefly reports more than 175 Miranda hot fires totaling over 4,200 seconds, which the company equates to roughly 20 full first-stage flights. That testing includes a flight-like run at 110% of the engine's 230,000 pounds of rated thrust. AST's release says the engine architecture has proven flight heritage, which overreaches the case. That heritage belongs to Firefly's smaller Reaver and Lightning engines, and Miranda and Vira have yet to fly.

The structure has precedent. According to Reuters, Firefly licensed Reaver to Astra in 2021 with competitive restrictions attached. This time Firefly calls the arrangement non-competitive, keeps ownership of the engine intellectual property, and continues selling launches to other customers. AST is buying the right to build the engine, not the design itself.

The design choices matter more than the specs. A single payload and a single destination strip out the requirements creep that tends to bloat general-purpose rockets. At the event, Avellan framed the third stage as a deliberate exception to his own best-part-is-no-part rule: staging lower and slower should make first-stage recovery far simpler than a Falcon-style return from higher energy. AST says the composite tanks reuse the material system, tooling, and Midland floor space already building BlueBird structures, and that the first tank set should be in testing by January. The stack-compression rig AST built to qualify five-satellite stacks is, in effect, already a piece of Blacklift ground equipment. Taken together, we see Blacklift as the launch equivalent of an ASIC, a chip designed for maximum efficiency in a single use case. AST's next-generation chip in its satellites passes up generality for what the company says is roughly ten times the processing bandwidth of its Block 1 satellites. Blacklift makes the same trade on the ride to orbit.

Market Analysis

Launch supply for large constellation payloads is tighter than the headline launch counts suggest. Multiple analysts predict a shortfall of up to 150 launches through 2030, with the crunch possibly lasting into the late 2030s. Vulcan's return this week is a commercial mission, and the Space Force has said it will study early flights before restoring national security assignments. New Glenn's January return will carry its reliability questions at least into mid-2027, though the program has already demonstrated booster landing and reuse. Rocket Lab pushed Neutron's debut to no earlier than 2027 while reportedly growing its manifest to 16 missions and lifting the base price to $70 million on several of them, citing tightening industry capacity. That is a big signal to us; when the rocket that hasn't flown yet is raising prices, buyers have their answer about leverage.

The 2028 Falcon 9 horizon is itself conditional. Media reports indicate that Musk has tied the wind-down to Starship flying dependably several times a week. That means a Starship slip could extend Falcon's commercial runway and soften the near-term case for Blacklift. We would treat that as a timing risk to the thesis, not a reversal of it, since a Falcon extension on SpaceX's terms still leaves AST negotiating for capacity rather than owning it.

Rocket Lab's income statement shows where launcher-first integration has led. Its Space Systems segment, which includes recent acquisitions, generated $189.5 million of $234.1 million in revenue. AST's version starts from captive payload demand, which Avellan put at roughly five tons per BlueBird, and builds the vehicle around it. This doesn’t read to us as a potential distraction. Much more is known about rocket design in the Falcon 9 and AI era, making vertical integration far less of a risk than when SpaceX did it.

For carrier partners including AT&T, Verizon, and the operators in Japan, Europe, Canada, Brazil, and Saudi Arabia, Blacklift changes little before 2029. What matters near term is AST's guidance for intermittent nationwide US service by year-end 2026 and roughly 45 BlueBirds in orbit by mid-2027. AST ended the quarter with about $3.2 billion in cash and restricted cash, which funds the Firefly deal without forcing a choice between rockets and spectrum. The downside is concrete. If Blacklift fails or slips badly, AST may have spent $400 million in cash and issued 10 to 14 million new Class A shares for a license it cannot use, before counting engine factory capital and the payroll of a propulsion team.

For the Pentagon, which has watched its two-provider assured-access model narrow to one provider for much of 2026, another domestic launcher with dual-use payloads fits a stated policy preference. The Space Force's Space-Based Airborne Moving Target Indicator (SB-AMTI) program already carries about $100 million in AST revenue, and Rocket Lab booked a dedicated Neutron launch under the same program in Q3. That suggests the government is buying both payloads and rides from vertically integrated suppliers in parallel. Firefly gets capital to scale Miranda production ahead of its own Eclipse debut.

Looking Ahead

The key trend we'll be monitoring is whether Blacklift's Firefly engine proves itself on someone else's rocket first. The real test is in late-2027, when Reuters reports that Eclipse, the Firefly and Northrop Grumman rocket also powered by Miranda, is planned to make its first launch. According to Avellan, that launch is from Wallops Island in Virginia, the same site AST plans to use. That flight is likely the first hard data on whether AST's late 2028 target is a plan or an aspiration.

The next few months set the context. Vulcan and Starship are scheduled to fly on 13 November, AST is due to reach six ASIC satellites a month in December, and January brings both New Glenn's return, with BlueBirds 17 to 19 expected early in the first quarter, along with AST's first composite tank tests.

Starship's flight rate will set Falcon 9's real exit date in the background. If Eclipse slips or Midland's tanks fail qualification, the third-party manifest becomes AST's whole launch plan through 2029 and the company spends $800 million on a rocket that does not yet exist. Recall what happened to the Amazon Leo launch cadence that relies on a bevy of third parties working on different clocks. It is obvious to us that AST does not want to be the next chapter in that case study.

Disclosure, a member of the author’s family works in a line manufacturing role at Blue Origin.