AST SpaceMobile’s decision to build its own rocket is a capital-allocation warning disguised as vertical integration.

The company has signed Firefly Aerospace to supply 36 engines for a new heavy-lift vehicle called Blacklift, including 27 Miranda engines and nine vacuum-optimized Vira engines, in a deal valued at more than $1.3 billion. AST also receives a license to manufacture those engines itself, while Firefly keeps ownership of the underlying intellectual property. The target is an initial Blacklift launch in late 2028.

The conventional take is that AST is securing reliable access to space and reducing its dependence on SpaceX. That is only half right. AST is not buying launch independence. It is buying another full-stack development program, with all the schedule risk, manufacturing complexity, qualification work, ground infrastructure, range coordination, and launch-cadence problems that come with it.

The cash and stock structure matters. Firefly gets a large contract and equity exposure, while AST converts part of its own market valuation into propulsion capacity. That is not the same as paying a launch provider for delivered missions. AST is effectively funding a new rocket while still having to build and operate the satellite network that generates the revenue case for the rocket in the first place. The deal may eventually lower per-satellite launch costs, but it raises the amount of capital AST must spend before the constellation can produce meaningful service revenue.

There is a legitimate strategic rationale. AST’s BlueBird satellites are unusually large, and Falcon 9 availability, rideshare constraints, and deployment timing can become bottlenecks as the constellation expands. Owning a launch vehicle could give AST control over payload integration and mission timing. But those benefits arrive only after Blacklift works, and the 2028 schedule is aggressive even before accounting for the fact that AST has never developed an orbital-class rocket.

That is the part the press release cannot make disappear: AST is attempting to become a launch company because its original business depends on launching a large number of satellites. Firefly gains a funded anchor customer. SpaceX loses a future customer only if Blacklift reaches flight and then flies often. AST shareholders inherit the technical and financial downside.

This is not primarily a rocket story. It is the beginning of a contest over who controls the bottleneck between satellite manufacturing and deployed service.

AST’s business model has always carried an uncomfortable dependency. The company can design a powerful direct-to-device constellation, raise capital, sign telecom partnerships, and win government work, but none of that matters if satellites remain on the ground. SpaceX has been the obvious answer because Falcon 9 is available, proven, and increasingly optimized for high-cadence constellation deployment. That convenience is also strategic dependence.

AST is trying to internalize the bottleneck. The problem is that launch is not a single bottleneck. It is a chain. Engines are one link, and Firefly’s Miranda and Vira programs address that link. They do not solve turbomachinery production rates, propulsion test capacity, avionics qualification, structures, flight software, stage integration, pad operations, range access, recovery decisions, or launch insurance. A rocket is not made dependable by attaching dependable engines to it.

The license provision is especially revealing. AST can eventually manufacture the engines, but Firefly retains the intellectual property. That arrangement gives AST a path toward control without requiring it to recreate every propulsion capability from scratch. It also lets Firefly monetize Miranda beyond its own launch vehicles. In practical terms, Firefly is turning propulsion into a platform business, while AST is taking on the role of systems integrator and launch operator.

That can work, but the economics are brutal. The first three Blacklift rockets require 36 engines, yet a constellation needs a sustained launch campaign, not a demonstration. If AST wants meaningful control over deployment, it must eventually purchase or build enough vehicles to create a cadence. Three rockets establish neither production learning nor operational resilience. They establish an expensive first batch.

The winners are clearer in the near term. Firefly gets revenue, validation, and a strategic customer willing to finance engine scale. Its engine business becomes less dependent on Firefly’s own launch cadence. Investors should value that more highly than another optimistic projection for Firefly’s launch backlog.

The loser is the assumption that satellite operators can cheaply become launch operators. AST may be able to do it, but the transition will consume management attention and capital at precisely the moment the company needs to focus on spectrum, terminals, constellation deployment, and customer acquisition. The market will applaud vertical integration until the first delay forces a choice between funding satellites and funding the rocket.

1. **The contract’s payment schedule and stock component.** A headline value above $1.3 billion is not equivalent to $1.3 billion in cash. Track how much AST pays up front, how much is tied to milestones, how much arrives as restricted stock, and whether the arrangement changes as Blacklift slips. The financing mechanics will show whether this is a committed development budget or a headline commitment spread across an uncertain program.

2. **Firefly’s Miranda production rate and test evidence.** Watch for hot-fire results, qualification milestones, repeatability across engines, and evidence that Firefly can produce engines at a rate compatible with constellation deployment. One successful engine test proves almost nothing about factory throughput. The real signal is whether production moves from bespoke hardware to repeatable industrial output.

3. **AST’s next launch commitments.** If Blacklift is genuinely about strategic independence, AST will continue using third-party launchers while Blacklift is developed. The important question is whether the company reserves enough Falcon 9 or other commercial capacity to protect the satellite deployment schedule. A sudden reduction in external launch bookings would suggest AST is substituting a promise for capacity.

4. **The first Blacklift vehicle’s architecture and pad location.** AST will need to disclose far more than engine count. Payload capacity, staging, launch site, integration model, recovery strategy, and range arrangements will determine whether the 2028 target is ambitious or unserious. A launch date without infrastructure is marketing, not a schedule.

The broader message is uncomfortable for every satellite company that believes launch access is merely a procurement line item. The market is moving toward vertical integration, but vertical integration does not eliminate risk. It relocates risk into the balance sheet and makes the customer responsible for failures that used to belong to the launch provider.

AST has bought control over one critical input. It has not bought a working rocket, a launch cadence, or a profitable constellation. Firefly may emerge as the real winner because it is selling the most valuable part of the program while keeping its intellectual property. SpaceX remains exposed only if Blacklift becomes operational, repeatable, and cheaper enough to matter.

Until then, Blacklift is not a credible alternative to Falcon 9. It is a very expensive option on one.