Two AST & Science applications published on May 28, 2026 are not about antennas. They are about traffic. US20260149498A1 (“On-Demand Resource Management Satellite Communication System and Method”) describes a base station that can “selectively switch between the first transmission protocol and the second transmission protocol based on an actual amount of data traffic by the UE in the field of view or an expected amount of data traffic.” Its companion, US20260149486A1 (“Satellite Radio Access Network (Sat RAN) Beam and Gateway Seamless Handover”), covers handing a connection from “a first setting satellite” to “a second rising satellite” as their fields of view overlap. Read as a pair, these published applications signal where AST's direct-to-cell program is putting its R&D: into the network-management layer, not the radio front end.

The framing matters because a published application is a delayed receipt for R&D direction, not a product and not enforceable coverage. Applications typically publish about 18 months after filing, so a cluster dated now reflects spending decisions made a year and a half ago. What the AST cluster shows is a company that has already established a deep antenna and phased-array footprint moving up the stack toward the operational problems of running a phone network from low Earth orbit — where calls cannot drop as satellites race overhead and where capacity has to follow demand.

From the antenna to the network

AST's earlier published applications are heavy on the physical layer — the spaceborne phased array that lets an ordinary handset talk to a satellite. US20250183996A1 (“Geolocation of Radio Frequency Devices Using Spaceborne Phased Arrays”) describes locating a handset within a phased-array field of view using “frequency offset (due to Doppler) and signal flight time,” and US20250119203A1 covers a “high throughput fractionated satellite” approach that distributes a spacecraft's function “across many small or very small satellites” flying in formation to synthesize a very large aperture. Those are aperture-and-geometry filings. The May 28 pair sits a layer above them, in the part of the system that decides how traffic is carried and how a session survives a satellite handoff.

The processing device is able to selectively switch between the first transmission protocol and the second transmission protocol based on an actual amount of data traffic by the UE in the field of view or an expected amount of data traffic by the UE in the field of view.— On-Demand Resource Management Satellite Communication System and Method, US20260149498A1

That direction is consistent across the company's recent slate. The handover problem in particular recurs: US20250247144A1 describes transitioning a beam “from inactive to active to provide network access to the user device” and back when the user is out of range, and US20260121738A1 (“Polarization-Transparent Channel Switching in Satellite-RAN Communication”) addresses preventing “dropped connections when communicating with client communication devices” through polarization switching that is “transparent to the users.” The May 28 gateway-handover application extends the same continuity-of-service thread to the satellite-to-satellite handoff itself. The CPC distribution across the cluster is dominated by H04B 7/18513 and neighboring satellite-radio-access codes — the comms-network family, not the antenna-structure family.

The capacity problem the filings address

The on-demand resource-management application is the one to weight, because it speaks to the hardest commercial constraint a space-based cellular network faces: a single satellite covers an enormous footprint but carries finite capacity, and demand inside that footprint is wildly uneven — dense over a city, sparse over an ocean. US20260149498A1 describes switching transmission protocol “based on an actual amount of data traffic… or an expected amount of data traffic,” which is the patent-language version of pointing capacity where the users are. The fractionated-satellite filing US20250119203A1 attacks the same capacity ceiling from the hardware side, distributing function across formation-flying spacecraft to “spatially reuse spectrum.” Two filings, two routes at the same bottleneck — one in software, one in spacecraft architecture.

Stacked against the earlier handover and beam-transition applications, the picture that emerges from the record is of a company filing methodically across every layer a working network needs: aperture and geolocation at the bottom, beam and gateway continuity in the middle, and traffic-aware capacity management on top. That breadth across consecutive filing years is itself the datum. A company solving only the antenna would file only antenna patents; AST's published slate spans the whole stack, which is what a company intending to operate a network, not just demonstrate a link, would be expected to file.

What the signal does and does not say

The grounded reading is that AST's filings point toward operationalizing a standards-compliant mobile network from orbit: traffic-aware protocol selection so capacity tracks demand, and seamless beam-and-gateway handover so a call does not drop as one satellite sets and another rises. Those are the unglamorous network functions a commercial direct-to-cell service has to solve to work for an ordinary subscriber, and the publication record shows AST filing into them across consecutive years. The discipline here is to read that as direction, not delivery: a published application is a statement of where engineering effort went, not evidence the feature is fielded or that any claim has issued. None of these records is a granted patent.

For a reader tracking the company through its filings, the quantifiable point is the shift in subject matter. AST's published applications began on the aperture and geolocation problems and have moved, in the most recent filings, onto traffic management and handover continuity — two of them published in a single week. That progression, from building the link to running the network, is the signal the cluster carries.