From mobile phones to Starlink: RF chips are becoming the commanding height of the industry in the satellite internet era.

From Mobile Phones to Starlink: RF Chips Are Becoming the Strategic High Ground of the Satellite Internet Era

In early 2026, an application for frequency and orbital resources covering 203,000 low-orbit satellites was submitted to the International Telecommunication Union. At the same time, SpaceX officially launched preparations for its IPO, with a target valuation as high as $1.5 trillion and plans to raise more than $30 billion. These two events may seem to belong to the narratives of aerospace and capital markets, but their impact is rapidly transmitting along the industrial chain to RF chips—a semiconductor category with high technical barriers and an entrenched market structure.

European chipmaker STMicroelectronics has delivered more than 5 billion RF antenna chips for Starlink to SpaceX, and its executives have said that deliveries under this collaboration could double over the next two years (by 2027). The competitive space for RF chips is shifting from smartphones to satellite internet. This article will analyze, from the perspectives of the industrial chain, technology roadmaps, competitive landscape, and regional map, how the wave of low-orbit satellites is reshaping this industry sitting atop the "crown of analog chips."

1. RF Chips: The "Throat" of Analog Chips and the Ceiling of Concentration

An RF communication system includes the antenna, RF transceiver chips, baseband chips, and the RF front end. The RF front end consists of power amplifiers (PAs), low-noise amplifiers (LNAs), filters/duplexers, RF switches, and antenna tuners. It processes high-frequency analog signals and must be developed using special processes based on GaAs, SOI, bulk silicon, and piezoelectric substrates. Because of the high frequencies, wide bandwidths, and stringent requirements on power consumption and linearity, the RF front end is regarded as the field in analog chips with the highest threshold and the greatest demand for accumulated expertise.

Over the past decade or more, smartphones have defined the iteration logic of RF chips. The market landscape is also highly concentrated: Skyworks, Qorvo, Broadcom, Qualcomm, and Murata together account for 84% of the global market, with Skyworks ranking first with a 21% share. However, consumer electronics demand has leveled off, and RF vendors need to find a second growth curve in new infrastructure.

2. Low-Orbit Satellites Press the "Launch Button": RF Demand Will Move from the "Millions" to the "Hundreds of Millions"

Low-orbit satellite constellations are the core prerequisite for the 6G space-air-ground integrated network. Unlike terrestrial networks, satellite internet requires each satellite's RF system to have wide coverage, large capacity, low latency, and high reliability. The manufacturing, launching, and networking of satellites first stimulate investment in front-end hardware. Referring to industry data, global satellite industry revenue had already reached $277.4 billion in 2018, of which satellite manufacturing revenue was $19.5 billion, with a growth rate of 28%. The structure of that year shows that satellite network construction always drives hardware manufacturing first.The figure of 203,000 satellites now far exceeds the total currently in orbit. Whether this is a long-term vision or a claim-staking move, the scarcity of spectrum and orbital resources means countries must “file first and build later.” For the RF chip supply chain, this means orders will not grow linearly; instead, they will rise in a stepwise manner driven by “large-scale launches.” SpaceX’s Starlink has already become one of the world’s largest buyers of RF antenna chips, and ST’s 5-billion-chip delivery volume validates the enormous consumption capacity of satellite terminals for RF chips.

III. Technology Route: 3D Heterogeneous Integration Is the Watershed in the RF Sector

LEO satellites impose constraints on RF systems not only in parameters, but also in size, weight, and power consumption. Conventional planar-assembled RF modules struggle to meet the lightweight requirements of satellite synthetic aperture radar (SAR) and phased-array communication payloads. The reference content clearly states that 3D heterogeneous integration technology for RF micro-front-end systems is becoming the key to breaking through the performance bottlenecks of traditional RF front-ends.

This technology stacks chips or bare dies with different materials, process nodes, and functions at high density in the vertical direction using TSV (through-silicon via), micro-bumps, RDL (redistribution layer), and hybrid bonding. For example, GaN/GaAs and other III-V compound semiconductor devices can be integrated with silicon-based CMOS control/digital circuits, passive devices, and MEMS into a tiny system, thereby shortening transmission paths, reducing parasitic effects, and improving system efficiency.

In aerospace scenarios, the most direct value of this technology is miniaturization and weight reduction. The weight and volume of T/R modules in phased-array antennas were previously difficult to reduce, whereas RF microsystems can vertically interconnect RF, power, and control modules through micro-scale integration, dramatically reducing satellite payload volume and lowering launch costs. Meanwhile, replacing a large number of discrete solder joints with integration helps improve system reliability in high-/low-temperature and radiation environments.

The competitive barrier for 3D heterogeneous integration lies not in a single process, but in the synergy of material supply chains, wafer-bonding equipment, advanced packaging and testing production experience, and system design tools. This also explains why advanced packaging is becoming a high-frequency term in discussions of the RF chip industry.

IV. Industry Chain Analysis: Linked Restructuring of Upstream, Midstream, and Downstream

Upstream: Specialty Materials and Equipment Are the “Ballast” of the Supply Chain

Demand from satellite-grade RF chips for GaAs, GaN-on-SiC, SOI substrates, and piezoelectric materials (lithium tantalate, lithium niobate, etc.) will rise rapidly. Aerospace-grade certification cycles are long, and once suppliers of materials and epitaxial wafers enter the aerospace procurement list, they gain a strong customer lock-in effect. Similarly, equipment for TSV deep reactive-ion etching, wafer bonding, and temporary bonding will benefit from RF microsystem capacity expansion.

Midstream: IDMs and Foundries Dance Together, but Aerospace-Grade Capacity Faces a “Crowding-Out Effect”Traditional RF front-end markets are led by five major IDMs, but the satellite track has introduced diversity. European IDMs represented by ST have already secured their position in the Starlink supply chain, proving the strategic value of in-house wafer fabs for aerospace-grade devices. At the same time, the silicon-based digital and interface portions still rely on advanced process nodes from leading foundries, while the GaAs and specialty SOI portions rely on specialized foundries or IDMs in the United States, Japan, and Europe. When handset RF orders have not yet declined and satellite demand begins to ramp up, specialty process capacity may face periodic squeezes. Traditional OSAT majors and high-end SiP packaging houses will absorb the demand for 3D integration; advanced packaging equipment and materials are essential infrastructure for the next stage.

Downstream: Satellite Operators Translate Needs into “System Specifications”

SpaceX and Starlink have formed a closed loop of “self-built satellites + self-operated ground networks,” and their influence over RF chips is no less than Apple’s influence over the mobile phone supply chain. ST’s delivery record shows that, once a supplier and an operator complete co-design, the collaboration lifecycle is sufficient to support chip orders on the scale of tens of billions. China is accelerating the first flight of its reusable rocket. Once constellation deployment enters the batch-launch cycle, the ability to define requirements for local terminals and phased-array antennas will become an important bargaining chip for domestic brands.

V. Competitive Landscape: New Growth Points Outside the Big Five’s Walls

The Big Five in mobile RF front-end have long competed for share based on the “device index,” whereas satellite internet values “high reliability + system integration” more. The latter requires companies to understand aerospace systems, master compound-semiconductor processes, and be able to execute collaborative design for advanced packaging—none of which is friendly to traditional consumer-grade suppliers. Therefore, the incremental opportunity brought by LEO satellites will not necessarily be fairly distributed across all of the Big Five. IDMs with aerospace-grade delivery experience and self-owned production lines (such as ST), as well as system-level suppliers with advanced packaging capabilities, are more likely to receive a share of the “satellite dividend.”

On the other hand, domestic RF vendors will gain a rare entry point in China’s independent satellite constellation construction. Compared with the mobile RF track, the technical specifications for LEO satellite systems have not yet been fully fixed, and supply-chain dominance is still being contested. China’s continued investment in localized materials, compound-semiconductor production lines, and system-level packaging is expected to create an opportunity for a “lane-change overtaking” within the defining window of domestic satellites and 6G networks. But this window will not stay open forever; the next three years are especially critical.

VI. Regional Industry Impact: The U.S. Defines the Rules, Europe Drives Manufacturing, China Sets the Tempo- United States: The Starlink ecosystem centered on SpaceX holds the power to define satellite system architecture, and a blockbuster IPO in the capital markets will bring financing spillovers to U.S. aerospace semiconductor startups. The U.S. RF industry's strength lies in system architecture and advanced design capability, while manufacturing and packaging remain highly dependent on overseas regions. - Europe: ST's cooperation with SpaceX proves Europe's irreplaceability in space-grade RF manufacturing. Going forward, Europe may leverage its quality systems accumulated in automotive and industrial grades to further consolidate GaN and RF-specific manufacturing capacity. - China: Low-orbit satellite constellation construction is a national-level demand. China is advancing recoverable rockets and satellite internet infrastructure in parallel, giving it a more urgent timetable in the race for orbital frequency and slot resources. The localization rate of RF materials, specialty manufacturing, and advanced packaging will rise rapidly driven by "SatNet orders." - Japan, South Korea, and Southeast Asia: Japan holds advantages in piezoelectric materials for filters and upstream compound semiconductors; South Korea's advanced packaging and interposer capacity can create value in the back-end of RF microsystems; Southeast Asia is expected to absorb part of the module final-test capacity, forming decentralized global supply nodes.- Next 3 years: LEO constellations shift from experimental networking to large-scale deployment, and satellite RF front-ends enter a period of rapid iteration. ST's expected doubling of deliveries implies that satellite applications will substantially divert consumer-grade RF capacity. After the maiden flight of China's recoverable rocket, domestically produced RF chips will receive their first batch of "space-bound" orders. - Next 5 years: Frequency band planning for 6G space-air-ground integrated communications will become clearer, and the potential applications of Ka/Q/V and terahertz communications will pose challenges at the materials level. RF microsystems will mature, three-dimensional heterogeneous integration will become the default option for high-end systems, and the boundary between traditional packaging and wafer foundry will further blur. - Next 10 years: If constellations on the scale of hundreds of thousands of satellites become a reality, total industry demand for RF chips could surge to a volume that consumer electronics cannot cover. The new type of communications infrastructure enabled by satellite internet will transform an industry formerly "attached" to mobile phones into an independent, strategically significant major semiconductor track.

Conclusion: RF chips are the "hard currency" in the competition for space infrastructure

LEO satellite constellations are pushing the competition for communications infrastructure from the Earth's surface into space. The RF front-end is the supply chain chokepoint on this new battlefield with the least room for retreat. The migration of market share will not be evenly distributed among all traditional giants. What truly captures incremental growth will be companies that can integrate device process technology, three-dimensional heterogeneous integration, and space-grade system certification into a cohesive whole, while embedding themselves deeply in the satellite ecosystem. The competitive coordinate system of the RF chip industry is shifting from "number of phones" to "number of satellites"—a transformation far more profound than people imagine.