Semiconductors have long been a cornerstone of the space industry—a highly technology-intensive sector requiring a wide range of power and radio-frequency components to operate. This dependency is becoming even more pronounced as satellites evolve to deliver high-speed broadband and, increasingly, direct-to-device connectivity. Conversely, the space industry is now emerging as a major growth engine for select semiconductor companies, supported by LEO constellations, lunar missions, and next-generation defense initiatives such as the U.S. Golden Dome.
STMicroelectronics illustrated this deepening relationship in recent days, highlighting the scale of the Low Earth Orbit (LEO) opportunity. The company expects its LEO-related revenue to grow from ~$600 million in 2025 to ~$1 billion in 2026 (+60%), and to exceed $3 billion cumulatively over 2026–2028. It also emphasized its role as a key partner to Starlink, supplying chips for both satellites—where content can reach several thousand dollars per unit—and user terminals. Beyond its largest disclosed program, STMicro indicated it has won a second global LEO customer.
The outlook is indeed compelling, supported by mega-constellations such as Starlink, Amazon Leo (which just acquired GlobalStar for $12 billion), and OneWeb. In broadband alone, STMicro estimates an addressable market of ~$650 million in 2025, expected to expand to ~$2 billion by 2028 and ~$2.9 billion by 2030.
Importantly, semiconductor content per satellite is set to increase materially, gradually shifting the revenue mix from user terminals toward satellites. This reflects rising bandwidth requirements, the expansion of direct-to-device connectivity, and the need for more advanced modems, RF components, and power management to handle massive device volumes. At the same time, constellations are evolving into true mesh networks in space, requiring high-speed optical interconnects (laser links between satellites) as well as increasingly sophisticated routing and switching capabilities.
Looking further ahead, space-based data centers could represent an additional leg of growth. The concept has gained traction as a potential solution to some of the key constraints facing terrestrial AI infrastructure—namely energy, land, and water. While significant challenges remain (launch costs, cooling, maintenance), near-term applications are already emerging. Space AI servers—essentially satellites dedicated to computing—appear particularly well suited for onboard data processing, scientific research, and defense use cases. Satellites generate vast amounts of data from Earth observation, climate monitoring, astronomy, and communications, and processing this data in orbit could reduce bandwidth needs while improving latency and responsiveness.
Against this backdrop, we see a substantial opportunity for STMicro and other semiconductor players with strong exposure to the space ecosystem. In our view, the company’s guidance (>$3 billion cumulative over 2026–2028) appears overly conservative. We believe LEO-related revenue could reach $1.5–2.0 billion annually by 2027–2028, implying that space could account for more than 10% of group revenue (vs. ~5% in 2025) and, more importantly, emerge as the second-largest driver of earnings after data centers.
