Anyone who thinks of Starlink exclusively in terms of satellite dishes and internet from space will have to get used to a broader picture. Under the Starlink Mobile brand, SpaceX plans to build a terrestrial 5G mobile network in the United States in the future. To that end, Qualcomm announced a multi-year collaboration with the space company on November 4, 2026. The focus is on the Dragonwing QRU100 and QDU100 platforms, which are expected to take on key radio and baseband processing functions in the future. The network rollout is set to begin in 2027 and cover tens of thousands of sites over several years. This significantly expands SpaceX’s mobile strategy beyond the direct satellite communications with smartphones offered to date.
Qualcomm is supplying two different hardware platforms for the planned mobile network sites. The Dragonwing QRU100 handles signal processing within the so-called Radio Unit. This is the part of a mobile base station that is directly connected to the antennas and processes the necessary radio signals. The Dragonwing QDU100, by contrast, is intended for baseband processing. Key tasks of the physical transmission layer, referred to in mobile communications as Layer 1, are performed there. These include digital signal processing and various time-critical calculations required for reliable communication between the base station and the end device.
Qualcomm is providing not only the processor platforms but also the corresponding Layer 1 software. This is intended to give SpaceX a technical foundation that can be deployed across numerous sites. The platforms are designed to maximize processing efficiency, since even small differences in power consumption can have a significant impact on operating costs across tens of thousands of base stations.
The current announcement does not include specific information on clock speeds, manufacturing processes, power consumption, or the frequency bands used in the final Starlink configurations. Nor has it been conclusively explained which hardware manufacturers are expected to supply the complete base stations. A key component of the planned infrastructure is Massive MIMO with a 64T64R configuration. This designation refers to 64 transmit and 64 receive paths that can be used within a corresponding antenna unit.
The technology makes it possible to direct radio signals more precisely in space and serve multiple end devices more efficiently. With beamforming, individual antenna elements or antenna paths can be controlled so that the radiated energy is preferentially directed toward specific areas. In suitable radio environments, this can improve both usable capacity and spectral efficiency. However, a higher number of antenna paths does not automatically mean correspondingly greater range. The network quality that can actually be achieved still depends on the frequency band, transmit power, site density, local conditions, and the number of active users at any given time.
Massive MIMO can make full use of its advantages particularly in densely populated regions, since multiple data streams can be processed in parallel and radio resources allocated more flexibly. In rural areas, by contrast, coverage, site availability, and the economic cost of network deployment are often more important. The new agreement is part of a broader SpaceX strategy. Starlink already has extensive satellite infrastructure that provides internet services in various regions. The company is also working on Direct-to-Device connections, which allow compatible mobile phones to communicate directly with satellites.
The terrestrial mobile network announced now takes a different technical approach. Here, the radio connection is initially provided through conventional ground-based base stations. Satellite-based and terrestrial communications can complement each other, but their frequencies, network architectures, and available capacities must be coordinated. According to the company, SpaceX aims to build a more integrated mobile infrastructure. The technical collaboration with Qualcomm is primarily intended to accelerate the development of SpaceX’s own base station architecture. Specific handover mechanisms between the satellite and terrestrial networks have not yet been disclosed in detail. Binding details are also lacking on which regions will receive coverage and when, and under what conditions end users will be able to use the services.
The announced start in 2027 does not mean that Starlink Mobile will already be able to offer nationwide mobile coverage at that point. According to Qualcomm, the plan is instead to roll out the network over several years across tens of thousands of sites. A nationwide mobile network requires numerous additional components beyond the base stations themselves. These include suitable antenna sites, power supply, transport networks, spectrum usage rights, and connection to a high-performance core network infrastructure. In addition, hardware and software must be tested and qualified under real-world conditions. The financial terms of the agreement have not been disclosed. It is therefore not yet known what the total value of the hardware deliveries to Qualcomm could be or what minimum purchase volumes SpaceX has committed to.
Although the current agreement focuses explicitly on building a 5G network, both companies are already discussing long-term development toward 6G. Qualcomm points to an open, extensible architecture and energy-efficient signal processing in this context. In the longer term, such platforms could also enable additional AI functions for optimizing mobile networks. In this context, the manufacturers describe a long-term development toward AI-native 6G infrastructure. However, this does not mean that the 5G base stations announced now will already support all future 6G standards. Specific 6G product features, performance data, and schedules have not been announced.
Conclusion
What I find particularly interesting about this agreement is that SpaceX is now pursuing its mobile ambitions on a much broader scale. The combination of satellite communications and conventional terrestrial infrastructure could open up new possibilities in the long term, even though the two technologies have different technical and economic requirements.
With its Dragonwing platforms, Qualcomm is positioning itself as an important hardware partner. The combination of Massive MIMO, baseband processing, and proprietary Layer 1 software could be particularly interesting for large-scale deployment. At the same time, the announcement should not be confused with an already existing nationwide network. There is still a lot to do before tens of thousands of sites are actually operational. Whether SpaceX can seriously compete with established mobile network operators will ultimately depend on network quality, coverage, and economic viability.
