Qualcomm to Power Tens of Thousands of SpaceX Starlink Mobile 5G Sites

SpaceX plans to build a terrestrial 5G network in the United States spanning tens of thousands of sites, with Qualcomm silicon and software at the heart of every base station. Qualcomm announced the multi-year collaboration on November 4, 2026. The rollout is set to begin in 2027 and extend over several years, marking a sharp expansion of SpaceX's mobile strategy beyond its existing direct-to-device satellite links for smartphones.

Which hardware will Starlink Mobile use?

Qualcomm supplies two Dragonwing platforms. The QRU100 handles signal processing in the radio unit, the part of a base station connected directly to the antennas. The QDU100 covers baseband processing, executing the physical transmission layer — Layer 1 in mobile terminology — including digital signal processing and the time-critical calculations that keep the link between base station and device stable.

Qualcomm also delivers the matching Layer 1 software, giving SpaceX a common technical foundation across the entire site footprint. The efficiency argument matters at this scale. A small difference in power consumption per unit becomes a significant operating cost when multiplied across tens of thousands of base stations, which is precisely the deployment size SpaceX has signaled.

• Dragonwing QRU100: signal processing in the 5G radio unit
• Dragonwing QDU100: baseband and Layer 1 processing
• Massive MIMO: 64 transmit and 64 receive paths (64T64R)
• Rollout: starting 2027, tens of thousands of sites over several years
• Outlook: long-term development toward 6G

The announcement omits clock speeds, manufacturing processes, power figures and the frequency bands for the final Starlink configurations. It also leaves open which hardware manufacturers will assemble the complete base stations.

What does 64T64R Massive MIMO change?

The planned infrastructure centers on Massive MIMO with 64 transmit and 64 receive antenna paths per unit. Beamforming lets individual antenna paths steer radiated energy toward specific areas, serving multiple devices more efficiently. In suitable radio environments this raises both usable capacity and spectral efficiency.

More antenna paths do not automatically buy more range, however. Achieved network quality still depends on the frequency band, transmit power, site density, local conditions and the number of active users at any moment. Massive MIMO shows its strengths mainly in densely populated regions, where parallel data streams and flexible resource allocation pay off; in rural areas, coverage, site availability and deployment economics dominate the planning calculus.

How does the terrestrial network fit with satellites?

Starlink already operates extensive satellite infrastructure delivering internet across multiple regions, and SpaceX continues to develop direct-to-device connections that let compatible phones reach satellites without a ground station. Starlink Mobile takes the opposite technical path: conventional ground-based base stations provide the radio link first.

The two approaches can complement each other, but their frequencies, network architectures and capacities must be coordinated. According to the companies, SpaceX aims to build a more integrated mobile infrastructure, and the Qualcomm collaboration is intended primarily to accelerate development of SpaceX's own base station architecture. Specific handover mechanisms between satellite and terrestrial networks remain undisclosed, as do the regions to be covered, the timing, and the commercial conditions for end users.

A 2027 start does not mean nationwide coverage on day one. Qualcomm describes a multi-year build across tens of thousands of sites. A full national network also needs antenna sites, power supply, transport networks, spectrum usage rights and connection to a high-performance core — plus real-world qualification of hardware and software.

What remains unknown — and what comes after 5G?

Neither company disclosed financial terms, so the total value of the hardware deliveries to Qualcomm and any minimum purchase volumes committed by SpaceX remain unknown. For practitioners, that gap matters: without committed volumes, build pace stays speculative.

Both companies frame the deal as a step toward 6G, with Qualcomm citing an open, extensible architecture and energy-efficient signal processing, and the partners describing a long-term path to AI-native 6G infrastructure. The current 5G base stations will not support future 6G standards automatically; no specific 6G product features, performance data or timelines have been announced.

For now the agreement positions Qualcomm as a key hardware partner for one of the largest greenfield radio builds in prospect, while SpaceX must still prove it can match established US mobile operators on network quality, coverage and unit economics as the 2027 rollout approaches.