AST is building cell towers in space. Instead of covering the planet with ground towers, they launch huge satellites that talk directly to your normal, unmodified phone. No dish, no special handset. Mobile carriers (who already have billions of subscribers) split the revenue with them. The first test satellites work; now it's a race to launch enough of them for real commercial service.
Scale is 0–50%/yr; the bull case runs off the chart. Figures are estimates from the research note. The shape of the verdict is what matters: even the base case clears the 30% growth bar with room to spare.
On October 29, the FCC votes on auctioning 25 MHz of prime spectrum for direct-to-device service, plus a first vote on freeing up another 482 MHz for satellite coverage. The auction is a window into the spectrum arms race: SpaceX already owns 65 MHz outright, Amazon is buying Globalstar's airwaves for about $11.6 billion with Apple as its device partner, and AST works on partner spectrum. Watch the bidder list, and whether AST bids too. An owned-spectrum move would be a strategic shift for a company whose entire model is borrowing carriers' airwaves.
On October 8, SpaceX agreed to buy Grain's 800 MHz airwaves (up to 14 MHz of low-band) for a planned Starlink Mobile carrier. That ends the November 5 partner decision that used to sit in this spot. AST was the other bidder, and its test license on the band runs through October 30. FCC approval of the sale is still pending. This is a real competitive negative: SpaceX chose to become the carrier instead of partnering with one. The counterweight is that a carrier-owning SpaceX gives AT&T and Verizon a reason to back the neutral wholesale partner. Nothing in the investment case broke: no contract lost, no milestone missed.
Two threads: finalization of the J-LEO joint venture (up to about $1 billion, non-dilutive), and Japan's 136-satellite ITU filing under AST's BlueBird name. If Japan funds its own constellation on AST hardware, other friendly governments can copy the template. On October 5, a joint U.S.-Japan technology statement formally welcomed the Rakuten-AST partnership for satellite infrastructure. Political cover arriving before the signature.
Three new US government awards since March through prime contractors, on top of prime seats on missile-defense and space-development programs. Watch for named disclosures and task orders, which is where awards turn into recognized revenue.
AST and Ligado met the FCC on September 30 to push their SkyTerra Next filing, which would put L-band payloads on AST's BlueBirds. Iridium, which opposes the move, met the FCC the day before. The likely outcome is a conditional approval rather than a clean win or a denial, so watch the docket. Separately, Rocket Lab is buying Iridium for about $8 billion, with shareholders approving in September and a close expected in mid-2027, which puts Iridium's L-band spectrum in the hands of a better-funded owner.
On October 1, 2026, AT&T, T-Mobile, and Verizon formally signed their joint venture to pool spectrum for satellite coverage, with industry veteran Paul Roth as interim CEO. Existing carrier deals stay in place, and the venture creates a legal path for T-Mobile subscribers to reach AST capacity down the road. In the same week, T-Mobile quietly removed Starlink branding from its T-Satellite website. The next leg to watch: an actual AST agreement signed through it.
Wayne Thorsen joined as EVP and Chief Commercial Officer in late September. He ran operations at SiriusXM and device partnerships at Google (Pixel, Fi, Nest). That is the hire you make when 60-plus carrier relationships need to turn into purchase orders.
The pending allocation of satellite spectrum to the European carrier joint venture AST is part of. Regulatory and slow, but it decides how big the European opportunity gets.
Most space startups face two impossible jobs: build the technology and win customers one by one. AST only has the first. AT&T, Verizon, Vodafone, Rakuten, and roughly 55 other carriers already own the customer relationships, more than 3 billion subscribers combined. They market the service, they bill for it, and they split the revenue with AST. Several are even funding the ground equipment. That is the whole trick: AST sells network extensions to companies that already have the customers, instead of competing with them.
Those giant antennas turn out to be useful for more than phone calls. AST holds prime positions on U.S. missile-defense and space-development programs, with documented work in positioning, tactical communications, and radar. Government awards to date top $125 million, and the same Block 2 satellites serve both missions. One constellation, two paying customers.
The bigger frame: military advantage is shifting from who owns the best platform to who sees first. AST's antennas are sensor apertures as much as communication arrays, which puts the company inside the fastest-growing budget line in defense, persistent orbital sensing. Think of it as a Cambrian explosion of sensing. When eyes evolved, every creature had to armor up or die. As proliferated satellite sensors come online, every military system built for the blind era has to adapt, and AST sells the eyes. The usual caution applies: most of the government relationship is still positioning and access, not signed production contracts. Price the demonstrated work, not the vision.
In September 2026, Japan's telecom ministry filed paperwork for a 136-satellite constellation at the ITU, filed under AST's BlueBird name. The model on the table: Japan finances its own sovereign constellation, AST supplies the satellites and the platform, and collects economics on hardware it was building anyway. If that template works, other friendly governments can copy it. Same satellites, a second revenue stack funded by taxpayers instead of shareholders.
Two things the market has barely priced in. First, spectrum: usable airwaves are the real bottleneck in wireless, and AST has locked up long-term rights, including 40 MHz of L-band in the U.S., plus access to its partners' bands. Second, starting with BlueBirds 47 and 48, the satellites get onboard AI compute hardware. Orbiting data centers, essentially. A third leg the street models barely mention.
From here the story runs on a launch schedule. About 25 satellites enables a beta with partner carriers late in 2026. About 45 enables full commercial service around February 2027. Management's goal is roughly $1 billion in the first full commercial year. Berenberg, the first big European bank to cover the stock, rates it a buy with a $92 target, and their model needs less than 3.5% of partner subscribers to actually pay. Q3 earnings are estimated for November 9.
Everything hangs on two dates: roughly 25 satellites for test service late this year, roughly 45 for commercial service around February. The 45-satellite campaign already slipped from the end of 2026 into early 2027, and commercial service could slide further. Every delay hands time to Starlink, which launches constantly and already operates at scale. The company also faces regulatory buildout deadlines in 2030 and 2033. A slip does not kill the story, but it pushes revenue outward and tests everyone's patience.
Most of these satellites fly on SpaceX rockets. SpaceX also runs Starlink, the direct competitor, which makes for an awkward dependence: your launch provider is racing you. The backup ride, Blue Origin's New Glenn, has had its own outage. Building an in-house launch option is theoretically possible but unproven and would cost a fortune. For now, the constellation gets built on someone else's manifest. And even if New Glenn comes back, it has never flown the deployment it would need: up to eight plane-change burns over a six-hour sequence to place each satellite into its own orbit.
This is one of the most capital-hungry businesses you can imagine. The company just raised $1.15 billion in convertible bonds, adding to an already large stack of debt-like financing, while burning cash heavily to build. If anything runs slower than planned, more fundraising is likely, and each round dilutes existing shareholders. Until real revenue arrives, the stock is a bet on execution.
Two open questions sit under the revenue story. First, the carriers: nobody knows yet what the wholesale economics look like at scale, what each user is worth, or how the network holds up when millions of phones connect at once. Second, the government: the Pentagon relationship is real, but most of it is positioning and access, not signed contracts with dollars attached, and converting it into booked revenue could be slow. The next proof points are the test service and real pricing, not press releases. The first real-world data point is soft: T-Mobile's CEO said its own satellite service is not getting as much use as anticipated.
This name routinely moves 5 to 10 percent in a single day. Short sellers are heavily involved, which stretches every swing further. On October 1, B. Riley cut its rating to neutral with a $65 target, arguing the shares fairly reflect what is known until subscriber uptake and pricing data arrive. There is also class-action noise around past disclosures: worth knowing, not thesis-breaking. Anyone interested should size it accordingly. This is the volatile seat at the table, not the steady one.
On October 9, Scotiabank moved the stock 13 percent on a framing error, saying the company has "no retail customers" when the whole model is wholesale through carriers. The call was wrong, but the price move was real. That is the lesson: being right about the business does not protect you from the marginal seller not understanding it. A stock can stay cheap on narrative alone, and "the market will figure it out" is not a catalyst.