The major customer of the rocket company has started to build its own rockets.

A satellite firm that has mapped out production plans for its 51st satellite is still struggling to find a way to send its spacecraft to orbit. It no longer wants to wait for rockets.

On November 9, U.S.-based satellite-to-cell connectivity firm AST SpaceMobile released its Q3 financial report.

The two satellite factories owned by the company in Texas are ramping up production, with an expected output of 6 new-generation BlueBird satellites per month by December.

However, as of the release of this financial report, AST only has 12 BlueBird satellites operating in orbit. Another 3 satellites were delivered to the launch site as early as September, and are still waiting for SpaceX to confirm the launch date.

Satellite production is accelerating, but the constellation deployment progress has not kept pace completely. Back in March this year, AST planned to have 45 to 60 satellites in orbit by the end of 2026; its latest target has been adjusted to around 45 satellites by mid-2027.

Just three days before the financial report was released, AST announced a bolder decision: to build its own rockets.

This reusable launch vehicle, named Blacklift, is designed with a low Earth orbit payload capacity of 20 tons. It is scheduled to make its first launch at the end of 2028, and can carry up to 5 BlueBird communication satellites per mission.

To take this step, AST has signed an agreement worth over 1.3 billion U.S. dollars with U.S. commercial space firm Firefly Aerospace, covering engine purchases, manufacturing license access, and engineering support for the next three years.

AST is not short of rocket suppliers. It has signed launch service agreements with enterprises including SpaceX and Blue Origin, and plans to deploy more than 100 BlueBird satellites in total by the end of 2028 through existing and additional launch arrangements.

Given that launch services are available on the market, why spend huge sums to build rockets on its own?

01 The constellation held back by launch progress

AST SpaceMobile's goal is to allow ordinary mobile phones to connect to mobile communication networks via low-orbit satellites without replacing terminals or installing special additional equipment.

Different from traditional satellite communication services that mainly target dedicated terminals, AST hopes to connect satellite communication capabilities to existing cellular networks through mobile operators such as AT&T, Verizon, Vodafone and Rakuten, to provide services to ordinary mobile phone users.

The prerequisite for this business is to send a sufficient number of satellites to orbit first.

The large-sized BlueBird satellites adopted by AST are equipped with deployable phased array antennas. The phased array of the second-generation satellite has a maximum area of about 223 square meters, which can deliver stronger signal coverage and support direct communication between ordinary mobile phones and satellites.

However, the large antennas also make the volume, mass and deployment requirements of the satellites more complex, putting forward corresponding requirements for the rocket's payload capacity, fairing space and launch mission scheduling.

In September 2024, AST launched 5 first-generation BlueBird satellites at one time via SpaceX's Falcon 9 rocket.

Two months later, the company announced new launch service arrangements with suppliers including Blue Origin, planning to deploy about 60 second-generation BlueBird satellites between 2025 and 2026.

Under the original plan, AST was responsible for manufacturing satellites, while professional rocket companies took charge of launch missions. This arrangement would allow the constellation construction to move forward without AST bearing the huge cost of independently developing rockets.

However, the actual deployment progress failed to meet expectations.

On December 23, 2025, AST launched its first second-generation BlueBird satellite via an Indian launch vehicle. In March 2026, the company still planned to have 45 to 60 satellites in orbit by the end of that year, with one launch mission executed every one to two months on average.

By November 9, this target had been adjusted to around 45 satellites in orbit by mid-2027. The deployment scale originally scheduled for the end of 2026 has been delayed for at least half a year.

This change cannot be entirely attributed to rocket supply. Satellite manufacturing, integration testing, regulatory approval and mission preparation may all affect the deployment progress. AST is also pushing for upgrades to satellite processing chips and adjustments to manufacturing processes.

Nevertheless, the uncertainty of external launch services has indeed caused actual losses to the company.

On April 19 this year, AST's BlueBird 7 was launched on Blue Origin's New Glenn rocket. Although the satellite entered space, it was sent to an orbit lower than planned, could not maintain scheduled operations, and was eventually deorbited.

According to the latest 10-Q document submitted by AST to the U.S. Securities and Exchange Commission, as of September 30, the company has received 32.4 million U.S. dollars in insurance compensation. After deducting the insurance payout, the accident still caused a loss of about 126 million U.S. dollars.

In accordance with the launch service contract, AST expects to get one additional replacement launch.

But neither the insurance compensation nor the replacement launch can recover the lost satellite, nor the time consumed by rescheduling related missions.

Worse still, even launches that do not end in accidents may disrupt the pace of constellation deployment.

AST disclosed in its November 9 financial report that BlueBird satellites No.14 to No.16 were delivered to Cape Canaveral, Florida as early as September, and are still waiting for SpaceX to confirm their launch dates.

BlueBird satellites No.17 to No.19 are expected to be delivered for launch before the end of November, and are scheduled to lift off on Blue Origin rockets in the first quarter of 2027.

BlueBird satellites No.20 to No.25 are scheduled to be delivered between December 2026 and January 2027, to participate in dual-use government and commercial missions related to the U.S. Defense Production Act priority rating.

The real problem is that AST cannot independently decide when these satellites will be sent to orbit.

As the scale of satellite production expands, the launch link may become a key factor restricting the speed of constellation deployment.

AST plans to reach a production speed of 6 satellites per month this December, equivalent to a theoretical annual production capacity of 72 satellites. But even if the satellites are manufactured successfully, they still need rockets to send them to orbit in a timely manner.

To guarantee the constellation deployment progress, AST has pre-purchased a large number of launch services. According to its latest quarterly report, the company currently has contracts for 10 future launch missions, plus additional launch opportunities expected to be obtained under the U.S. Defense Production Act priority rating, as well as non-binding agreements for at least 5 more launches.

The company estimates that these arrangements will support the deployment of more than 100 BlueBird satellites in total by the end of 2028.

However, AST also acknowledged in its risk disclosure that the existing long-term launch contracts do not cover all satellites planned for deployment. There is still a gap between signing a launch agreement and getting a confirmed launch window.

For AST, the speed of constellation deployment directly affects the launch of its commercial services. The company estimates that about 25 BlueBird satellites can provide discontinuous coverage in key markets; to achieve continuous services in major markets such as the United States, Europe and Japan, about 45 to 60 satellites are needed.

If satellites cannot enter orbit on time, network coverage and commercial operations will be difficult to advance as planned.

AST has already invested huge sums in building satellite factories and established cooperation with multiple global mobile operators. As satellite production speeds up, access to stable and predictable launch services is becoming a key factor affecting its business plans.

Independently developing the Blacklift rocket is exactly an additional option AST has added for its future constellation deployment.

02 Spending 1.3 billion U.S. dollars on engines: how will the rocket be built?

Building a rocket is far more complex and expensive than building a satellite.

A launch vehicle capable of sending a 20-ton payload to low Earth orbit needs to solve a series of engineering problems, including high-thrust engines, rocket body structure, propellant supply, flight control, stage separation, and ground testing.

If reusability is also required, further solutions must be developed for first-stage recovery, landing control, post-return inspection and maintenance, and re-launch.

For AST, which has been mainly engaged in satellite R&D and operation before, this means stepping into a brand new technical field.

AST did not choose to develop all core systems from scratch, but started with engines first.

Under the agreement signed with Firefly on November 6, the entire transaction is divided into three parts:

The first part is the engine manufacturing license. AST will pay 400 million U.S. dollars in cash and issue 10 million to 14 million Class A common shares to Firefly, to obtain a permanent manufacturing license for the Miranda and vacuum-version Vira engines.

Based on the valuation at the time of the agreement announcement, the consideration for the license part is about 800 million U.S. dollars.

The second part is engine procurement, with an amount of 250 million U.S. dollars. Firefly will deliver 27 Miranda engines and 9 Vira engines to AST, for the first three Blacklift rockets.

The third part is engineering support, also worth 250 million U.S. dollars, with a service period of three years.

The three parts add up to more than 1.3 billion U.S. dollars.

More importantly, the 1.3 billion U.S. dollars is not the full development budget for Blacklift.

AST still needs to complete the overall rocket design, structure manufacturing, system integration, flight test, launch facility and recovery system construction. The company has planned to build new engine manufacturing facilities in the United States, but has not fully disclosed relevant investments.

As of the latest quarterly report, the two parties expect the agreement to close in the fourth quarter of 2026.

Is it worth spending so much money to acquire engine technology?

For AST, the direct value of this deal is to reduce the workload of developing the power system from zero.

Firefly's Miranda engine has completed more than 175 hot fire tests, with a total test time of over 4200 seconds, including flight condition simulation tests under 110% performance conditions.

Miranda and Vira inherit part of the technology accumulated from Firefly's previous Reaver and Lightning engine projects.

For Firefly, this is another profitable business.

For a long time, commercial rocket companies mainly earn revenue through full rocket development, launch services and related aerospace system delivery. Engines are often regarded as core technologies that support the competitiveness of the entire rocket.

But this time, Firefly chose to sell the engine manufacturing license to another enterprise that is preparing to enter the rocket manufacturing field.

It can not only get revenue from early engine supply, but also directly monetize the technical assets formed by years of R&D investment through licensing and engineering services.

Jason Kim, CEO of Firefly, said in the announcement that this deal will help the company speed up launch vehicle production and expand its technology R&D capabilities.

The company plans to use the newly added funds to expand recruitment, pre-purchase long-lead materials, and add engine testing facilities at the Rocket Ranch in Briggs, Texas.

This has created an interesting situation: AST spends money to acquire the ability to build its own rockets, while Firefly uses this money to strengthen its own rocket business.

Moreover, Firefly did not sell the entire engine intellectual property rights to AST. According to the announcement, Firefly still retains relevant intellectual property rights and will continue to serve other government and commercial customers.

Of course, this model also has potential problems. If AST further develops its own rocket manufacturing capabilities in the future, whether the two parties will form competition in some businesses still depends on the specific scope of the technology license and future market strategies.

This transaction is valuable for both parties. AST gets engine technology and engineering support, while Firefly gets capital and orders. It is still unclear whether building its own rockets will be more cost-effective for AST than purchasing launch services.

Blacklift is designed with a low Earth orbit payload capacity of 20 tons, and can carry up to 5 BlueBird satellites per mission. But AST has not yet disclosed the manufacturing cost of a single rocket, number of reuses, recovery and maintenance costs, and expected annual launch volume.

Compared with purchasing launch services on a per-mission basis, independently developing rockets requires bearing huge R&D and facility investment, and then amortizing the cost through continuous launches. The more launches there are, the easier it is to dilute fixed costs. If the time satellites spend waiting for launch can be shortened to allow communication services to start operation earlier, additional revenue may also be generated.

AST has a certain capital base. According to its November 9 financial report, as of September 30, the company holds about 3.2 billion U.S. dollars in cash, cash equivalents and restricted cash, with a Q3 operating revenue of 42.3 million U.S. dollars.

However, satellite manufacturing, external launch procurement and ground system construction all require continuous investment. There is no clear budget for how much money can be allocated to Blacklift.

AST's satellite communication business is also expanding. On November 9, the company announced that it is advancing the Rakuten AST joint venture project with Japan's Rakuten, planning to build a satellite-to-cell network for the Japanese market. The project is expected to get government support of up to 148 billion yen.

This cooperation may increase future satellite deployment demand, but the relevant funds are not R&D funds for Blacklift.

Blacklift also faces a time problem.

According to the plan, this rocket will not start launching until the end of 2028, while AST's existing constellation deployment arrangements already cover the period up to the end of 2028. In the next two years, the company will still rely on suppliers such as SpaceX and Blue Origin to complete most of its launch missions.

Whether Blacklift can recover its upfront investment mainly depends on how many satellites still need to be launched after 2028, and whether the rocket can maintain a sufficient launch frequency and reduce costs through reusability.

All these need to be verified by actual operations. AST is investing huge sums in independently developing rockets right now, which can neither replace external launch services in the short term, nor bring clear cost advantages. It is betting on the continuously growing satellite launch demand in the future and the long-term benefits brought by it.

03 Rocket firms build satellites, satellite firms build rockets

AST's decision to independently develop rockets cannot be separated from SpaceX.

In the past, commercial rocket companies were mainly responsible for launches, while satellite enterprises were responsible for manufacturing and operation. SpaceX broke this division of labor through Starlink.

SpaceX not only owns launch vehicles such as Falcon 9, but also operates its own low-orbit satellite internet network. Relying on its own rockets, it can schedule launch missions according to Starlink's constellation deployment needs, and coordinate the progress of satellite production and deployment.

High-frequency launches and reusability also help reduce launch costs.

For satellite operators that rely on external launch services, this advantage is very difficult to replicate.

SpaceX has another identity: a competitor in the satellite communication market.

AST has long purchased launch services from SpaceX. In September 2024, 5 first-generation BlueBird satellites were sent to orbit on Falcon 9; in June and August 2026, AST launched 3 second-generation satellites via Falcon 9 respectively.

The two parties are client and supplier in the launch business, but have a competitive relationship in the satellite-to-cell market.

AST acknowledged in its latest 10-Q document that the number of third-party suppliers that can meet its satellite launch needs is limited, and some of these enterprises also operate businesses that compete with AST.

In other words, AST needs to rely on its competitors to provide a key service.

Amazon, which is also facing launch supply issues, has chosen a different path.

In April 2022, Amazon signed up to 83 launch arrangements for its then Project Kuiper satellite network.

Among them, ULA will undertake 38 Vulcan Centaur launches, Arianespace will undertake 18 Ariane 6 launches, and Blue Origin will provide 12 New Glenn launches plus up to 15 optional missions.

In December 2023, Amazon purchased 3 more Falcon 9 launch services from SpaceX.

Even though Starlink is a competitor of Amazon Leo, Amazon is still willing to purchase launch capacity from SpaceX. It did not choose to independently develop large rockets, but chose to guarantee constellation deployment through multiple suppliers and long-term contracts.

However, large-scale procurement cannot completely eliminate launch supply risks.