AST SpaceMobile ($ASTS): Can Deployment Become Recurring Revenue?

01 Scenarios and the conditions behind the bull and bear cases

What Would Falsify This Reading

Sustained, separately disclosed SpaceMobile revenue and usage would disprove the view that consumer monetization remains unproven. Repeated launch or deployment failures, or materially weaker manufacturing progress, would challenge the assumption that the planned fleet can be built on the stated timetable. A funded national project with definitive contracts and clearly attributed payments would strengthen the financing case; an expected project value without those documents would not. These tests follow the August 10, 2026 update and the June quarterly report; they are conditions, not price targets. Sources.

02 How the space-to-phone network works, and what its speed claims establish

The attraction of AST’s architecture is the possibility of reaching a phone already in someone’s pocket. In the June 30, 2026 business description, SpaceMobile uses satellites, ground gateways and the partner operator’s cellular network to serve an unmodified mobile device outside terrestrial coverage Quarterly report. This is a wholesale distribution model: the operator relationship is part of the product, not simply an advertising arrangement. A useful connection still requires spectrum access, integration with the operator’s core network, usable capacity and permission in the relevant country. Quarterly report.

Large phased arrays address a difficult radio link between a moving spacecraft and a small handset. Their size can help the link budget and beam control, but it does not abolish shared capacity. A peak rate obtained under favorable test conditions says something about technical capability; it does not state the sustained rate available to all users in a crowded service area. The June 30, 2026 report and the May 11 update cite a demonstrated peak of 98.9 Mbps with an in-orbit Block 1 satellite and an unmodified phone over international waters Accounts · First-quarter update. That is a dated demonstration, not an average commercial user experience. Accounts · First-quarter update.

The August 10, 2026 release says Block 2 satellites are expected to deliver peak rates approaching 200 Mbps Second-quarter update. The same update describes the proprietary ASIC’s processing capacity and expected improvement in throughput. Those are design and forward-looking claims. The measured result so far is the Block 1 peak; the higher Block 2 rate has not yet been demonstrated on a commercial network Second-quarter update. The question is whether the architecture produces reliable service under the traffic, geography and spectrum conditions that actual operators need. Second-quarter update.

The June 30, 2026 report describes Block 2 arrays of up to approximately 2,400 square feet, more than three times the Block 1 array size, and a design intended to improve bandwidth Quarterly report. It also discusses AST5000 production and the possibility of continuing FPGA-based production in parallel Quarterly report. This qualification matters: a new chip entering steady production is not evidence that every satellite already in orbit uses that chip. Manufacturing choices, software integration and qualification can change the mix of hardware that supports the network. The 10-Q says the AST5000 chip is expected to support up to 40 MHz per beam and up to 10,000 MHz of processing bandwidth per Block 2 satellite, that ASIC production has reached a steady stage, and that Block 2 satellites based on FPGA chips will continue to be assembled and launched in parallel; these are company expectations, not measured network results. Quarterly report.

Ground infrastructure deserves the same attention as spacecraft. Gateways move satellite traffic into the cellular network; operator integration handles authentication, service delivery and local requirements. The August 10, 2026 update describes nearly 50 gateways at various stages of completion, installation and planning Business update. A planned gateway, an installed gateway and one carrying commercial traffic are different states. Counting all of them as operational would remove the very execution steps that the company still has to complete. The same release reports an initial 3,000 digital cells activated across the continental United States in preparation for a beta service in 2026, described as scaled non-commercial usage with strategic operator partners. Business update.

The analytical test is therefore broader than whether a phone can connect once. It includes repeatability, handoffs as spacecraft move, available coverage, capacity under load and the cost of maintaining service. AST’s demonstrated link is meaningful evidence that the concept can work. Public documents do not establish a mature network-wide average speed, paid subscriber cohort or unit economics for the consumer service. Until those measures are published, a demonstration peak cannot stand in for them.

03 Fleet inventory, launch failures and the manufacturing timetable

Counting the constellation by operational state gives a clearer picture than a single headline total. The June 30, 2026 quarterly report describes the five Block 1 BlueBirds launched in September 2024, BlueBird 6 launched in December 2025 and deployed in February 2026, and BlueBirds 8–10 launched on June 17, 2026 with array deployment completed in July Quarterly report · August update. The August 10 update adds the launch of BlueBirds 11–13 on August 5, 2026 Quarterly report · August update. Launch is a milestone; completed deployment, testing and commercial readiness require their own evidence. Quarterly report · August update.

The August 10, 2026 issuer release reports 13 spacecraft in orbit Accounts and fleet discussion · Update. For commercial BlueBird inventory at the August 10, 2026 update, the relevant count is 12, calculated as five Block 1 satellites plus BlueBird 6 and BlueBirds 8–13; the broader spacecraft count includes the BlueWalker 3 test spacecraft Accounts and fleet discussion · Update. BlueBird 7 is excluded because it was deorbited Accounts and fleet discussion · Update. Neither count means that every spacecraft has completed all commercial activation steps. This distinction prevents an attractive headline total from being mistaken for a fully usable consumer fleet. Accounts and fleet discussion · Update.

BlueBird 7 shows the gap between a launch and a working network Initial filing · Quarterly report. The April 20, 2026 filing said that during the New Glenn 3 mission the upper stage placed the satellite in a lower-than-planned orbit; it separated and powered on, but the altitude was too low to sustain operations with its on-board thrusters, so it would be de-orbited, and its cost was expected to be recovered under insurance Initial filing · Quarterly report. The later June 30, 2026 quarterly report records deorbiting and the actual accounting outcome: a US$125.911 million loss net of insurance, with US$32.5 million insurance recovery, of which approximately US$21.6 million had been received and US$10.9 million remained receivable at quarter end Initial filing · Quarterly report. Against the April expectation, the recoveries booked by June 30 covered only part of the loss; the 10-Q adds that AST expects a replacement launch under the terms of its contract with the launch provider. Initial filing · Quarterly report.

Insurance supports the financial response to a lost asset; it does not restore the lost slot in the constellation or the time needed to replace it. Manufacturing another satellite, obtaining a launch opportunity and deploying it take resources even when part of the asset cost is recoverable. A credible fleet assessment therefore needs both the technical status and the cash and accounting consequences. Treating the insurance figure as proof that the incident had no economic effect would ignore schedule risk.

The June 30, 2026 MD&A states a target of approximately 45 BlueBird satellites in orbit by early 2027, replacing the target of approximately 45 satellites in orbit by the end of 2026 stated in the April 20, 2026 filing and the May 11 release; the annual report had planned approximately 45 to 60 Block 2 satellite launches by the end of 2026, at one launch every one to two months on average, and the July 15, 2026 filing first moved the target to early 2027. The earlier plan counted Block 2 launches, the later targets count all BlueBirds in orbit April filing · May release · July filing · Quarterly report · Production update. The June 30, 2026 report describes agreements covering launches of more than 60 Block 2 satellites Quarterly report · Production update. Those are future transport arrangements, not existing assets in orbit. The August 10, 2026 release places BlueBirds 17–46 at various production stages and said BlueBirds 14–16 would be ready to ship shortly; as of October 5, 2026 no AST filing or release had announced their launch date Quarterly report · Production update. A partially assembled unit is not a completed satellite waiting only for a rocket. Quarterly report · Production update.

The June 30, 2026 report says AST has completed the investments needed to assemble, integrate and test up to six Block 2 satellites per month and is still accelerating toward that production run rate; at the filing date it had completed fully assembled microns for up to 36 BlueBirds, including those already shipped or launched, with BlueBirds 14 to 46 in various stages of production and integration, some in final pre-shipment quality review Quarterly report. Capacity indicates the potential rate supported by facilities; completed output shows the rate achieved. Comparing those measures without their definitions can make a ramp look finished. The useful next disclosures would connect components, completed spacecraft, shipment, launch and deployment, showing where the bottleneck actually sits.

Coverage also follows a threshold rather than a smooth count. In the June 30, 2026 MD&A, management says 25 BlueBirds (five Block 1 and 20 Block 2) can support selected commercially attractive markets with non-continuous coverage, approximately 45–60 can enable continuous service in key markets, and approximately 90 can address additional target geographies Coverage planning. These are company planning estimates with geographic and operational assumptions. They are not a promise of universal continuous service wherever a phone can be carried. The same filing states management’s belief that the company is fully funded for the costs to manufacture and launch a constellation of approximately 90 BlueBirds; that is a company assessment, not an audited measure, and it rests on the per-satellite cost estimate discussed in the cash chapter. Coverage planning.

04 FCC permission and spectrum rights are conditions for service

The FCC order adopted April 21, 2026 authorizes a constellation of up to 248 satellites, including 223 additional spacecraft beyond prior authority, subject to the attached conditions FCC order. It is an enabling regulatory decision, not evidence that the authorized fleet has been manufactured or that customers can already purchase nationwide continuous service. The order was granted by the FCC’s Space Bureau and other bureaus under delegated authority. Set against the authorized ceiling of 248 satellites, the August 10 count was 12 commercial BlueBirds in orbit, a Merlintrader comparison. FCC order.

The authorization carries interference, coordination and operational obligations. The April 21, 2026 order includes requirements to protect other users, coordinate with federal systems and public-safety operations, and stop or modify operations in relevant interference circumstances FCC conditions. The conditions for certain bands and airborne use cannot be generalized into unrestricted permission for every service configuration. A network can have a valuable authorization while still needing engineering, agreements and compliance work to operate within it. Among the specific conditions, AST may not provide 800 MHz service to airborne devices, must comply with the public-safety licensee protection requirements for 800 MHz and 700 MHz operations, and must immediately cease operations that cause harmful interference to an affected U.S. federal system; for the 223 additional satellites, use of the 430–440 MHz band is limited to emergency operations of no more than 24 hours when no other band is available. FCC conditions.

The same FCC order establishes deployment milestones of half the authorized constellation by August 2, 2030 and the full authorized constellation by August 2, 2033, with the license term extending to September 19, 2039 April 21, 2026 order. These regulatory milestones are different from management’s nearer-term launch targets. Missing a commercial target does not automatically mean the license has expired; holding the license does not mean the commercial target has been achieved. Failure to meet the regulatory milestones, however, may reduce the authorization to the number of satellites in use at the milestone date and would forfeit AST’s surety bond; the company must certify compliance with the first milestone, or report that it was missed, by August 17, 2030. April 21, 2026 order.

Regulation is also geographic. U.S. authorization does not grant permission in Japan, Europe or every other market represented by an operator relationship. The June 30, 2026 quarterly report describes international regulatory approvals and integration as prerequisites for service Quarterly report. National regulators can require different spectrum arrangements, traffic handling, infrastructure and coordination. A mobile operator’s customer base therefore represents a route to potential distribution only where the network can legally and technically serve it. Quarterly report.

AST’s spectrum portfolio combines shared mobile-operator spectrum with mobile satellite service rights and agreements. The August 10, 2026 release describes a strategy targeting approximately 100 MHz in the United States and more than 60 MHz globally on a market-by-market basis Release · Spectrum discussion. The word “targeting” and the geographic qualification are part of the claim. The S-band ITU priority rights, held through an entity acquired on September 25, 2025 and covering frequencies in the 1980–2010 MHz and 2170–2200 MHz ranges for use in low Earth orbit, and the Ligado agreements are rights and contracts, not unconditional ownership of operating licenses in every country. Release · Spectrum discussion.

Spectrum can improve the business case by supporting more capacity or government applications, but it can also create recurring payments and complex dependencies before service receipts arrive. Interference restrictions may narrow where or how a band is used. Priority rights may need national implementation. Access agreements can require cash, shares or revenue sharing. For this reason, spectrum is best evaluated as a set of rights, conditions and costs rather than a single impressive bandwidth total.

The April 21, 2026 FCC order also imposes reporting and coordination related to orbital debris and other affected operations FCC order and conditions. Those obligations matter because a large constellation interacts with other spacecraft, astronomy and federal users. A conditional approval reflects the regulator’s framework for managing those issues; it does not establish that all collision, optical or interference risks have disappeared. The order records a NASA letter stating that initial analyses project light pollution from the AST constellation above thresholds recommended by the astronomical community, and notes AST’s preliminary coordination agreement with the National Science Foundation and regular meetings with NASA under a Space Act Agreement. FCC order and conditions.

05 Operators, Europe and Japan: distribution before proven adoption

The August 10, 2026 issuer update reports signed partnerships with more than 60 mobile network operators whose combined customer bases exceed three billion subscribers Business update. These are the operators’ customers, not paying AST SpaceMobile users. The distinction is central to the valuation question: an operator relationship can shorten the route to market, but it cannot establish the fraction of customers who will use the satellite service, how often they will use it or what revenue AST will retain. Business update.

The June 30, 2026 report describes a space-based wireless connectivity agreement with AT&T and a definitive commercial agreement with Verizon, announced on October 8, 2025, both covering the continental United States (excluding Alaska) and Hawaii, and a ten-year commercial agreement with Saudi Telecom Company signed on October 29, 2025 Commercial agreements. Different agreements have different territory, exclusivity, prepaid amounts and revenue-sharing terms. A memorandum of understanding is not interchangeable with a definitive commercial contract. Even a definitive agreement remains subject to operating readiness and applicable regulatory requirements. Commercial agreements.

Operator distribution has a practical advantage: the service can fit into an existing billing and customer-support relationship. It also leaves AST dependent on another party’s rollout decisions, pricing and marketing. The public material does not disclose a uniform take rate, retained revenue per user or churn measure across the partner network. Without those inputs, any per-user revenue model for the partner base would rest on assumptions rather than disclosed data. The August 10 release adds company statements that network integration and testing are under way with Vodafone, Orange, Telefónica, Vodafone Ukraine and Deutsche Telekom in Europe, and in Canada, Japan and Saudi Arabia, subject to final regulatory approvals, and that a joint venture planned by the top three U.S. operators is expected to enable space-based cellular broadband nationwide; no terms for that venture are given. Business update.

European development includes SatCo, the joint venture with Vodafone. The June 30, 2026 report describes equal ownership established in July 2025 and an equity-method investment rather than a consolidated subsidiary SatCo note. The transaction described in the June 30, 2026 accounts assigned US$23.5 million of non-cash consideration to rights, divided between an equity investment and an interest-bearing receivable SatCo note. The split was a US$5.9 million equity investment and a US$17.6 million receivable from SatCo carrying 6.6% annual interest. AST recognized equity-method losses of US$1.9 million in the second quarter and US$5.3 million in the half year, and related-party revenue from gateway equipment sold to SatCo of US$1.9 million and US$9.8 million respectively, eliminating the intra-entity profit through the equity-method loss. This structure matters because related-party gateway revenue and the venture’s finances do not become independent consumer demand merely by appearing in a recognized-revenue line. SatCo note.

At June 30, 2026, the SatCo equity investment’s carrying balance was zero, with further accounting effects applied to the related receivable Quarterly report. That is an accounting outcome under the equity method. It does not mean the European venture had been dissolved or that all future opportunity was worth zero. Conversely, a continuing venture does not erase the losses or make its distribution plan a completed commercial rollout. Both statements can be true without contradiction. Quarterly report.

Japan requires an especially careful separation of service plans and national infrastructure funding. The July 15, 2026 filing describes preliminary selection of a proposed Rakuten–AST venture for a J-LEO initiative with expected project value of up to approximately JPY148 billion, presented by AST as around US$1 billion July filing · June report subsequent events. The selected entity is RAST Co., Ltd., named as an indirect subsidy recipient, and the filing states that there is no assurance that the joint venture will be finalized or that government financing will be secured. This is not cash already received by AST, nor a new parent-company quarterly balance. July filing · June report subsequent events.

Rakuten’s August 10, 2026 investor presentation describes its selection as an indirect subsidy operator for domestic J-LEO infrastructure and separately targets commercial service through AST for the fourth quarter of 2026 Rakuten presentation, page 49. These are related ambitions with distinct contracting and infrastructure steps. Rakuten says the J-LEO details will be disclosed once finalized, and no AST filing through October 5, 2026 reports definitive J-LEO terms. Rakuten presentation, page 49.

The evidence that would change this chapter is straightforward: operator activation with defined customer access, separately reported commercial service receipts, or finalized project contracts identifying the recipient and timing of funds. Partnership counts remain useful context, but the stronger evidence connects the relationship to usable service and money that has actually been earned or received.

Canada supplied a concrete example on October 5, 2026. TELUS and AST SpaceMobile announced that they had completed the first integration test between TELUS’ wireless network and AST’s space-based network, showing broadband data, voice calls and text messages passing directly between smartphones customers already own, with no special equipment, and low-Earth-orbit satellites; the companies expect availability to customers within the next year. Source The release names no test location, spectrum or launch date. An integration test is the step that moves a partnership from paper into a working network connection, which is the kind of evidence this chapter looks for. It is not yet the activation of paying customers, and the release discloses no commercial terms or revenue.

The Rakuten venture also received a political signal on October 5, 2026. In the joint statement on the U.S.-Japan Technology Prosperity Deal published by the U.S. Department of State, both sides “welcomed Japan’s Rakuten Group and the U.S.-based AST SpaceMobile cooperation for the development and deployment of Low Earth Orbit (LEO) satellite infrastructure.” Source The sentence names no program, amount or date. It can be read as support for the strategic standing of the Rakuten venture described above, but it adds no contract, funding or revenue to it.

06 Recognized revenue, backlog, government awards and the guidance bridge

The June 30, 2026 quarterly report records US$31.520 million of second-quarter revenue, compared with US$1.156 million in the corresponding 2025 quarter Quarterly report. For the quarter ended June 30, 2026, product revenue was US$24.428 million, and services revenue US$7.092 million Quarterly report. The filing identifies gateway equipment and software and government performance obligations as the drivers. It explicitly states that SpaceMobile Service revenue had not yet been recognized. Growth in reported revenue is real; its composition remains essential. Quarterly report.

For the first half ended June 30, 2026, revenue was US$46.255 million, versus US$1.874 million a year earlier Half-year accounts · First-quarter results. The May 11, 2026 release reports first-quarter revenue of US$14.735 million in its financial statements (US$13.406 million product, US$1.329 million services), so the second quarter increased by US$16.785 million, calculated by subtraction Half-year accounts · First-quarter results. A sequential increase supported by deliveries and milestones can show progress in execution. It does not by itself establish a recurring consumer revenue stream, because the timing of equipment acceptance and government obligations can produce uneven quarterly recognition. Half-year accounts · First-quarter results.

The annual report for the year ended December 31, 2025 records revenue of US$70.918 million, made up of US$44.389 million product revenue and US$26.529 million services revenue Annual report. This historical comparison helps explain the direction of the business: the recognized base already included equipment and government work before the consumer network reached scaled service. Treating all revenue growth as satellite subscription growth would therefore distort both the prior year and the current quarter. Annual report.

Backlog figures come with different definitions. At June 30, 2026, the filing reports approximately US$1.2 billion of remaining performance obligations and expects approximately 6.6% to be recognized in the following twelve months Revenue note · Update. Variable amounts tied to future operator activity can be outside that measure. The August 10, 2026 release reports approximately US$1.30 billion of aggregate contracted revenue and government awards Revenue note · Update. The first is a footnote measure in the unaudited June 30 accounts; the second is a press-release figure dated August 10 that includes government contract awards, so the two are not interchangeable. Revenue note · Update.

Contract liabilities of US$266.933 million at June 30, 2026, compared with US$226.980 million at December 31, 2025, represent obligations associated with amounts received or due before revenue recognition; the June balance splits into US$14.840 million current and US$252.093 million non-current Quarterly report. They are not additional revenue to be added to the income statement. The filing also discusses a US$45 million Verizon commercial prepayment due as of August 10, 2026. The filing states that the payment was due at the filing date, not that it had been received; separately, it reports that certain regulatory approvals for the SpaceMobile Service were received on April 22, 2026. The timing of invoices, collections and fulfilled obligations is part of the funding and revenue story. Quarterly report.

Government work provides a second commercial route. The Space Development Agency announced a US$30 million firm-fixed-price HALO Europa prototype award on February 23, 2026, with demonstration planned for completion by December 2027 SDA announcement · Issuer update. The August 10, 2026 AST release describes aggregate government awards exceeding US$125 million SDA announcement · Issuer update. These announcements cannot simply be added together because their scope may overlap, and an award is not the same as recognized revenue or immediate cash collection. SDA announcement · Issuer update.

On August 10, 2026 the company said it was on track to achieve full-year revenue guidance of US$150–200 million, requiring the calculated second-half contribution set out in the free scenarios. The useful question is which deliveries and performance obligations can bridge that requirement, and whether receipts follow recognition. A result driven mainly by one-time equipment orders would have different persistence from recurring service usage. A missed guide would require explaining whether the problem was acceptance timing, a delayed government milestone, fleet deployment or weaker underlying demand; these causes have different implications.

07 Cash, working capital and the cost of building the constellation

The June 30, 2026 balance sheet reports US$2,288.253 million of cash and cash equivalents Quarterly report · Release. At June 30, 2026, restricted cash totals US$434.581 million, calculated by adding the current and non-current restricted balances, producing combined cash of US$2,722.834 million Quarterly report · Release. The August 10, 2026 release’s cash figure exceeding US$3.7 billion is pro forma for the subsequent July financing Quarterly report · Release. It is not the actual unrestricted cash balance on June 30, and it is not a newly reported October balance. Quarterly report · Release.

This distinction changes the practical interpretation of financial capacity. Restricted cash can support a secured loan while remaining unavailable for ordinary operating needs. Gross proceeds from new debt create repayment and interest obligations. A pro forma balance can explain a financing event without updating every subsequent expenditure. The documents support a substantial funding position, but they do not support treating every announced facility or restricted dollar as simultaneously free to spend.

For the six months ended June 30, 2026, operating cash flow was negative US$145.212 million, compared with negative US$72.024 million in the prior-year period Cash-flow statement. For the six months ended June 30, 2026, cash purchases of property and equipment were US$859.215 million Cash-flow statement. For the six months ended June 30, 2026, subtracting those purchases from operating cash flow gives negative US$1,004.427 million, calculated Cash-flow statement. This measure is a simple cash bridge, not management’s definition of free cash flow and not a forecast that the same pace will recur indefinitely. Cash used in investing activities totalled US$979.723 million in the half, against US$430.622 million a year earlier; the MD&A attributes the increase to US$428.6 million more property and equipment purchases, the US$100 million Ligado advance and US$42.1 million more in L-band, revenue-share and Crown Castle payments, partly offset by US$21.6 million of BlueBird 7 insurance proceeds. Financing activities provided US$1,068.3 million. Cash-flow statement.

The build is visible in the assets as well as in cash spending. At June 30, 2026, net property and equipment were US$2,069.336 million, with satellite construction and related advances forming a major part of the capital base Balance sheet and property note. Inventory was US$28.417 million, compared with US$12.007 million at December 31, 2025 Balance sheet and property note. These balances represent resources committed to production and deployment; they do not establish that the whole balance is already generating service revenue. Balance sheet and property note.

Working capital also matters as equipment revenue expands. Receivables increased to US$79.297 million at June 30, 2026, from US$37.726 million at December 31, 2025, and the cash-flow statement records a US$41.571 million first-half cash use from the receivables movement Accounts. Revenue recognition can therefore move ahead of cash collection. It would be incomplete to compare only sales growth with operating expenses and ignore the money tied up in customer balances. Accounts.

The June 30, 2026 report estimates long-run direct material and launch cost of US$21–23 million per Block 2 satellite across a constellation exceeding 90 units, with initial launches higher and improvements dependent on scale, supplier economics and future launch terms Capital requirements. This is a management estimate with exclusions and assumptions, not the complete cash cost of the entire company divided by its fleet. Ground assets, spectrum, overhead, financing and replacement risk sit elsewhere. Capital requirements.

The financing question is consequently not just whether AST can pay its next bill. It is whether spending converts into working capacity before another major funding need, and whether that capacity produces enough durable revenue to support the capital structure. A useful next report would connect receivable collections, deployment spending and the fleet’s operating status. Cash on hand buys time; the return on that time remains dependent on execution.

08 Ligado access, recurring charges and the conditional October financing window

The June 30, 2026 report describes agreements for access to up to 40 MHz of L-band spectrum and up to 5 MHz in the 1670–1675 MHz band, with closing dependent on approvals and other conditions Spectrum and commitments notes. AST had paid US$520 million of capital advances by quarter end, including the US$100 million payment made on or about March 31, 2026, which a Bankruptcy Court order of April 2, 2026 placed in escrow pending a further court order Spectrum and commitments notes. In the June 30, 2026 transaction discussion, the definitive agreements provide for a US$550 million contingent payment from AST to Ligado, with the remaining obligations treated as an off-balance-sheet commitment conditioned on approval and closing Spectrum and commitments notes. These are contractual payments and access arrangements, not unrestricted licenses already usable everywhere. Spectrum and commitments notes.

The transaction also creates recurring economics. The June 30, 2026 report describes an L-band annual payment of at least US$80 million, plus revenue-sharing obligations, and separate Crown Castle usage payments for up to 5 MHz in the 1670–1675 MHz band, carrying a 30% premium payable at the company’s discretion in cash or Class A shares Quarterly report. Certain payments had begun before final closing. Buying access can strengthen the capacity strategy while creating costs before the service has reached commercial scale. Quarterly report.

A backstop commitment can support a refund of the US$520 million advanced if the relevant regulatory approval and closing conditions fail, subject to the agreement’s requirements as described at June 30, 2026 Spectrum transaction discussion. That is conditional protection against a specific outcome. It is not cash on demand that can be added to the operating balance, and it does not remove the time or opportunity cost of funds committed to the transaction. Spectrum transaction discussion.

The Sound Point delayed-draw facility is another conditional item. The June 30, 2026 report says the US$550 million facility was undrawn, with availability through October 5, 2026 and an option to extend by 180 days on payment of an additional fee of 1% of the loan amount Credit facility. Approval, bankruptcy-related events, security documents and other funding conditions remain relevant. The availability date is a contractual checkpoint, not a scheduled FCC decision or proof that cash will automatically arrive. As of October 5, 2026 no SEC filing reported a draw, an extension, a termination or FCC approval of the Ligado transaction. The approval path also depends on Inmarsat: under the June 2025 mediated agreement it must provide regulatory support, but in December 2025 it sued in New York State court instead; in January 2026 the Bankruptcy Court ordered it to comply, a District of Delaware stay of that order on February 27, 2026 was vacated by the Third Circuit on March 2, and Inmarsat’s appeal remains pending, which the 10-Q says could impair approval if the ruling is overturned. Credit facility.

The pricing shows why availability and usable proceeds differ. The June 30, 2026 report describes a 2% commitment fee already paid, a monthly 0.15% ticking fee, a 3% fee on drawing, and interest based on SOFR plus 8% or an alternate base rate plus 9% Debt and commitments. Extension carries the 1% fee, and early termination a fee of 1% to 5% of the loan amount, payable in cash or Class A shares at the borrower’s option; drawn loans would mature 48 to 60 months after funding. The non-recourse structure limits certain creditor claims against the parent and AST LLC, but it still places subsidiary assets and associated revenues within a secured financing arrangement. Debt and commitments.

Separately, the UBS bridge loan had US$420 million principal outstanding at June 30, 2026 and was collateralized by US$428.4 million of restricted cash; it bears interest at Term SOFR plus 2.0% Debt note. The agreement’s stated maturity is October 31, 2028, subject to earlier repayment events Debt note. It is therefore misleading to add the borrowing to available liquidity while ignoring the cash pledged behind it. The loan and the restricted balance are connected parts of the same financing structure. Debt note.

The next meaningful evidence is an approved and completed transaction, an actual draw or documented extension, or a change to the access economics. An approaching deadline can matter without being a binary stock catalyst. The uncertainty lies in the combination of permission, financing conditions and future commercial returns; it cannot be reduced to a single spectrum headline.

09 Convertible debt, share counts and the cost of future capital

At June 30, 2026, debt principal totaled US$3,022.152 million Debt note. The balance sheet’s net carrying amounts differ because of issuance costs and other accounting adjustments. Principal describes the contractual debt amount; carrying value describes the reported liability under the accounting treatment. Both are useful, provided that they are not mixed in an unexplained net-debt calculation. The total excludes the US$1,150 million of 1.625% convertible notes due 2034 issued in July; adding them gives about US$4,172.2 million of principal, a Merlintrader pro forma calculation that ignores any repayment or conversion after June 30. Debt note.

The June 30, 2026 debt schedule includes remaining 2032 notes and the newer 2036 issues, together with the UBS and equipment financings: US$3.514 million of 2032 4.25% notes, US$325.0 million of 2032 2.375% notes, US$1,150.0 million of 2036 2.00% notes, US$1,075.0 million of 2036 2.25% notes, US$48.638 million under the Trinity equipment loan and US$420.0 million under the UBS loan; the filing puts the fair value of the debt at about US$3.5 billion Debt note and cash flows. Earlier convertible repurchases involved issuance of Class A shares, and the half-year cash-flow statement records US$88.654 million of induced conversion expense. The cash-flow statement records related share proceeds and debt-repurchase payments; counting the proceeds without the linked payments would overstate money raised for deployment. A conversion or share-funded repurchase can reduce debt while increasing the common-share base. Debt note and cash flows.

The July financing adds a separate layer. The July 20, 2026 filing records closing of the base offering; the later June-quarter subsequent-event note reports US$1.150 billion of 2034 notes including the fully exercised option, with a 1.625% coupon and maturity on February 1, 2034 Closing filing · Subsequent events. For the July 2026 notes, the initial conversion price is approximately US$79.57, while the capped-call transactions have an initial cap price of US$149.20 Closing filing · Subsequent events. The latter is not the notes’ contractual conversion price. The initial conversion price was set at about a 20% premium to the July 15, 2026 last reported sale price. Closing filing · Subsequent events.

The June 30, 2026 report’s subsequent-event note gives July net proceeds of approximately US$1,131.2 million and capped-call cost of approximately US$111.4 million, leaving US$1,019.8 million, calculated, before other uses Subsequent events. Capped calls can offset dilution or cash settlement exposure within their terms; they do not permanently eliminate dilution at every stock price or under every future corporate action. The August 10 release does not define its “effective conversion price of US$149.20” or its “effective dilution of less than 2%”; in Merlintrader’s reading they reflect the notes and the capped calls combined, since US$149.20 matches the cap price reported in the 10-Q, and they are not the notes’ own terms. Subsequent events.

Share counts need similar care. At June 30, 2026, the accounts show 299,731,073 Class A shares, 11,215,111 Class B shares and 78,163,078 Class C shares Equity note. Class B and C voting shares are linked to economic interests in AST LLC and do not independently carry the same economic rights as Class A stock. At June 30, 2026, the fully exchanged economic count is 389,109,262, calculated from Class A plus the linked LLC interests; each LLC interest and its paired voting share count once, not twice. Equity note.

The quarterly cover gives 299,789,305 Class A shares as of August 6, 2026, with the other voting-class counts unchanged, making the comparable fully exchanged count 389,167,494, calculated Cover and earnings-per-share note. These are dated measures, not a live cap table. Potential awards and convertible shares are separate from outstanding stock. Equal basic and diluted loss per share in the June 2026 accounts reflects anti-dilutive treatment during a loss period; it does not establish that the securities cannot dilute holders later. Cover and earnings-per-share note.

The old October 2025 at-the-market program was terminated in March 2026 after virtually using its capacity, according to the June 30 report. Its original headline amount is no longer an untouched current funding line. The investment question is how future cash needs will be met and what claim existing shareholders retain after that financing. Neither a low coupon nor a large cash balance settles the cost of capital by itself.

10 Voting control, insider transactions and the market snapshot

The company has an Up-C structure rather than a single uncomplicated share class. The June 30, 2026 report distinguishes the listed company’s shares, AST LLC interests and noncontrolling ownership Equity and noncontrolling-interest notes. This affects the denominator used for market value and the relationship between economic ownership and voting control. A third-party figure for Class A shares cannot be assumed to represent all exchangeable economic interests or the total votes. Equity and noncontrolling-interest notes.

In the June 23, 2026 Schedule 13D amendment, Abel Avellan reports beneficial ownership including shares issuable on exchange of his LLC interests Founder filing. The June 23, 2026 filing distinguishes its 20.8% beneficial Class A percentage from a 71.6% voting interest under the stated denominator and Class C structure Founder filing. These are dated definitions, not interchangeable ownership percentages. The same filing describes a June 22 prepaid forward covering up to 2.5 million shares, with settlement in March 2028 and retained voting rights over the pledged interests. Founder filing.

That forward arrangement creates a personal financing and contingent settlement exposure. It is not equivalent to an immediate ordinary sale of the entire referenced block, and its effect cannot be inferred from the largest share count alone. What matters for governance is the rights that remain, the settlement conditions and the founder’s continuing control. Economic alignment and minority-holder influence are related questions, but they are not the same question.

The September 17, 2026 Form 4 for chief technology officer Huiwen Yao reports a 40,000-share sale on September 16, with weighted-average price US$58.93, under a trading plan adopted on June 5, 2026 Yao filing · Gupta filing. The September 17, 2026 Form 4 for chief operating officer Shanti Gupta reports a 12,000-share sale on September 16 at weighted-average US$58.89; unlike Yao’s, it does not tick the Rule 10b5-1 trading-plan box Yao filing · Gupta filing. These are disclosed transactions. They do not establish either an imminent operational problem or a guaranteed positive outlook. Other recent filings add context: on August 17 the company granted restricted stock units to six officers, 150,000 to Abel Avellan, 200,000 to Christopher Turco, 75,000 each to Andrew Johnson, Shanti Gupta and Scott Wisniewski and 20,000 to Maya Bernal, 595,000 in total by Merlintrader’s sum; on August 19 Yao exercised 40,000 options at US$0.0641; on August 31 director Adriana Cisneros reported indirect purchases of 10,822 shares at US$57.00–58.87; and on September 28 the new Chief Commercial Officer received 200,000 units. Forms 4 · Cisneros filing · Thorsen filing. Yao filing · Gupta filing.

Nasdaq records closes of US$58.86 on September 30 and US$58.45 on October 2, 2026. Multiplied by the 389,167,494 shares as exchanged on the August 6 cover, the October 2 close gives a market value of about US$22,746.8 million, a Merlintrader calculation; on the 299,789,305 Class A shares alone it would be about US$17,522.7 million. Nasdaq prices · Accounts. The two denominators answer different questions in an Up-C structure, and dividing a value computed on one by the other creates an apparent inconsistency. Free float is available only as a third-party estimate: Finviz shows 187.70 million shares at September 30. Finviz.

Nasdaq reports short interest of 65,261,118 shares at the September 15, 2026 settlement, equal to 6.9 days of average volume, up from 62,487,152 at June 15; that is about 21.8% of the Class A shares on the August 6 cover, a Merlintrader calculation. Nasdaq short interest. Nasdaq’s 13F aggregation shows institutional ownership of 65.55% across 876 holders, with holder report dates that vary; Finviz reports insider ownership of 37.39% under its own definition. Nasdaq holdings · Finviz. These are third-party data measures with their own definitions and observation dates, not the founder’s voting power. A large short position can contribute to volatility and crowded trading, while also reflecting hedging or skepticism. It does not itself prove a future squeeze, fraud, undervaluation or a direction for the share price.

Market liquidity makes the stock tradeable; it does not make a large gap between current revenue and market value harmless, and the operating record is separate from the mechanics of the market snapshot.

11 Competition, launch dependence and the limits of the AI opportunity

AST’s competitive case rests on direct connectivity to ordinary phones, operator integration, large arrays and access to suitable spectrum. The annual report for the year ended December 31, 2025 describes competitive and technical risks, including alternatives in satellite communications and terrestrial coverage Annual report. Those differentiation claims are tested by delivered capacity and service economics. A company can have a distinct design while still facing competition for the same customers, spectrum and operator budgets. Annual report.

A competitor’s messaging or narrowband product need not match AST’s intended broadband performance to affect demand. Some users may need only a low-cost emergency connection, while others may value higher capacity. Operators decide how the satellite service fits into their own offers. The relevant comparison therefore includes reliability, usable coverage, pricing and distribution, rather than asking only which network advertises the largest peak speed.

Launch access is both a technical dependency and a financing issue. The June 30, 2026 report discusses multi-provider launch arrangements and third-party risks Quarterly report. A booked launch reduces one uncertainty but leaves readiness, weather, vehicle performance and deployment steps. BlueBird 7’s loss demonstrates the practical consequences Quarterly report. Manufacturing more rapidly can shift the bottleneck to transport or orbital activation; buying additional access to orbit can consume cash before the network’s return becomes visible. Quarterly report.

The July 15, 2026 offering material describes possible partnerships or acquisitions intended to improve vertical integration and mitigate launch-provider risk, while stating that no understandings or agreements existed for such strategic transactions at that time July filing · September filing. It is an intended use of funds, not an announced acquisition. The September severance policy also does not establish that a transaction is underway. Merlintrader has found no filing through October 5, 2026 that names a target, a buyer or a counterparty. July filing · September filing.

The August 10, 2026 update discusses opportunities in government communications, radar, emergency response, IoT and AI edge computing Business update · Quarterly accounts. These can broaden the platform’s potential uses. The quarterly revenue disclosures do not isolate a recurring AI service business or establish how much of those opportunities will be commercialized. On the disclosed record they are development opportunities, not an existing revenue franchise. Business update · Quarterly accounts.

AI-enabled spectrum management can potentially improve scheduling or resource allocation within the network. It does not remove radio physics, the need for spectrum permission, spacecraft reliability or partner adoption. A useful future disclosure would explain deployed functionality, performance under operating conditions and the customer contract that pays for it. Until then, multiplying an addressable-market estimate by an assumed market share would create apparent precision without verified inputs.

The counterargument to this cautious reading is that capital investment and testing must precede service revenue in a network build. That is true, and it explains why current revenue alone cannot describe the platform’s full opportunity. It does not eliminate the need to test the eventual economics. The stronger case emerges when technical milestones, operator usage and cash receipts advance together, rather than when a new application expands the narrative.

12 Losses, controls, litigation and the September change-of-control policy

The June 30, 2026 accounts report second-quarter operating expenses of US$329.097 million and net loss attributable to common stockholders of US$230.909 million Quarterly income statement. Consolidated net loss before allocation to noncontrolling interests is a different measure. The loss includes the satellite incident and non-cash costs, but a large accounting loss is not the same as operating cash burn: first-half operating cash outflow was US$145.212 million (chapter 07). Quarterly income statement.

The August 10, 2026 release reconciles adjusted operating expenses of US$119.055 million for the June quarter against US$91.179 million for the March quarter Results and reconciliation. The adjustment removes stock-based compensation, depreciation and amortization and the satellite-loss item. Excluding adjusted cost of revenue gives another measure, not the same expense base. The reconciliation helps isolate spending trends, while also removing costs that remain economically relevant to shareholders or to asset replacement. Results and reconciliation.

Stock-based compensation in the June 30, 2026 accounts is material: US$63.467 million expensed in the quarter and US$118.820 million in the half year, with a separately disclosed capitalized amount Stock compensation note. It can reduce immediate cash expense while increasing potential share issuance. Engineering expense can support the platform’s future capacity while recurring administrative costs absorb resources now. Neither category becomes irrelevant simply because an adjusted measure excludes part of it. Stock compensation note.

Management concludes that disclosure controls were effective at June 30, 2026 and reports no material change in internal control over financial reporting during the quarter Controls and legal proceedings. These are management’s dated conclusions, not an independent guarantee against every future error or legal challenge. Under legal proceedings, the 10-Q describes only ordinary-course matters and says management considers a material loss remote; it predates the September securities complaint. The Ligado bankruptcy-court process, including the April 2, 2026 order placing AST’s US$100 million payment in escrow, is described separately in the spectrum note. The same report states that Inmarsat’s appeal of the Bankruptcy Court ruling requiring it to provide regulatory support for the Ligado transaction remains pending and that, if the ruling were overturned, AST’s ability to obtain regulatory approval for the transaction could be impaired. Controls and legal proceedings.

The complaint filed September 14, 2026 in Hunter v. AST SpaceMobile, case 7:26-cv-00378, pending before the U.S. District Court for the Western District of Texas, Midland/Odessa Division, against the company, Abel Avellan and chief financial officer Andrew M. Johnson, alleges misleading statements about financing needs, dilution, competition and adoption. It seeks to represent purchasers between March 4, 2025 and July 15, 2026 under Sections 10(b) and 20(a) of the Exchange Act, and cites among other things the SpaceX–EchoStar spectrum agreement and the July 15, 2026 convertible offering. Filed complaint. Rosen Law Firm’s September 30, 2026 notice reiterates the November 13, 2026 deadline to seek appointment as lead plaintiff; it is a reminder about the existing lawsuit and announces no court ruling or operational development. Rosen notice. The plaintiff’s allegations are unproven. Neither the complaint nor the law firm’s reminder establishes liability or a loss estimate, and no AST filing through October 5, 2026 addresses the suit; the company’s most recent legal assessment, in the June 10-Q, predates it. SEC filings.

The change-of-control severance policy adopted September 25, 2026 and filed September 28 provides benefits following a qualifying termination tied to a change of control, subject to its conditions September 28 filing. The policy generally covers qualifying events within one year after a change of control and certain specified circumstances before it. It covers the chief executive, the president and all executive and senior vice presidents. A qualifying termination without cause, or for constructive discharge, on or before the first anniversary of a change of control, or in certain cases within 180 days before it, entitles the executive to a lump sum of 2.0 times (chief executive) or 1.5 times (others) annual base salary plus target bonus, a pro-rata target bonus and 24 or 18 months of health-premium differential, subject to a release of claims; equity awards granted after the policy’s adoption vest in full on such a termination. This is a contingent compensation framework, not an announcement of an executive departure or takeover. September 28 filing.

These issues affect different parts of the assessment. Adjusted expense trends inform the cost of deployment. Stock awards affect existing holders’ claims. Control conclusions describe reporting processes. Litigation introduces allegations and legal uncertainty. Severance terms describe contingent obligations. Combining them into a single “good” or “bad” signal would hide the mechanisms behind each one.

13 The next evidence that can strengthen or weaken the operating case

The closest contractual checkpoint in the June 30, 2026 report is the Sound Point facility’s October 5, 2026 availability end, with the extension option and funding conditions already described Credit facility terms. A documented extension, draw, amendment or transaction closing would be informative. As of October 5, 2026 no SEC filing had reported any of these, so the outcome of the window remains undisclosed. Credit facility terms.

For the fleet, the next useful evidence is actual shipment, launch, deployment and activation of the spacecraft that the August 10, 2026 update placed in the production pipeline Quarterly planning · Business update. The company’s target of approximately 45 BlueBirds by early 2027 remains a target Quarterly planning · Business update. The timetable has already moved more than once: the annual report planned approximately 45 to 60 Block 2 launches by the end of 2026; the April 20 filing and the May 11 release targeted approximately 45 satellites in orbit by the end of 2026, a count that includes Block 1; since the July 15 filing the target has been approximately 45 BlueBirds in early 2027. Production capacity and parts completed cannot fill a missing deployment milestone. Quarterly planning · Business update.

For revenue, the key test is the transition from equipment and milestone payments to service usage. The June 30, 2026 quarterly report’s absence of SpaceMobile Service revenue is a dated fact, not a claim that the business can never generate it Report · Beta description. Separately disclosed revenue, usage, partner activation and collection would allow a more concrete assessment. A beta program described as non-commercial can validate the technology without yet satisfying that monetization test. Report · Beta description.

For Japan and Europe, final agreements and country-specific operating evidence matter more than an aggregate partnership count. The decisive documents would be a definitive funding agreement naming the entity that receives the funds, the obligations attached and the schedule, and service announcements stating what customers can access and under what conditions. These disclosures would narrow the gap between a distribution opportunity and a functioning business without requiring a speculative valuation model.

For cash, the third-quarter accounts will show the effect of the July funding after financing costs, deployment spending and working-capital changes; no reporting date had been announced as of October 5, 2026, and last year’s third-quarter results came on November 10, 2025 Investor relations. For the capital structure, new share issuance, conversions and awards matter alongside debt changes. For spectrum, permission and access costs need to advance together. Each of these tests follows a mechanism already visible in the current documents; none depends on guessing the share price.

The editorial assessment recognizes substantial funding and meaningful engineering progress, alongside a large deployment program and unproven recurring service economics. A favorable technical demonstration does not settle capital efficiency. A weak quarter of revenue does not alone disprove the network. The strongest evidence is a sequence in which deployed assets, useful service, recognized recurring revenue and collected cash reinforce one another. The assessment will change when those facts change, keeping the dates and conditions that support each conclusion.