Interview: Why this Taiwan mmWave maker chose to sidestep SpaceX and look to niche markets for demand

SpaceX has collapsed the mmWave value chain—phased-array silicon, satellite bus, ground terminals—into one vertically integrated entity. A Taiwan, China mmWave fab choosing to sidestep isn't retreating; it's the only rational capital allocation against an asymmetric player controlling 70%+ of active LEO terminals. The technical implication is a demand bifurcation at the 28/39/47 GHz GaN/SiGe node: volume fabs serve constellation operators, while niche foundries feed OEMs building certified ground equipment for regulated verticals (maritime, aviation, enterprise backhaul). The binding constraint isn't process maturity—it's export classification. Components above 100 GHz increasingly sit in EAR/ITAR adjacency, and a Taiwan, China-based fab qualifying for both restricted and unrestricted markets simultaneously faces a 12–18-month certification treadmill that inadvertently entrenches incumbents. SpaceX won't react; its playbook is absorption. The real competitive pressure arrives from Qualcomm's NTN modem roadmap and Samsung's 6G NTN chipset timeline. The niche player's window closes when 3GPP R19 reference designs lock in, roughly 2026. Over the next 12–24 months, expect 2–3 M&A events in mmWave ground-terminal silicon as Ericsson and Nokia seek constellation-agnostic NTN capability without single-vendor lock-in. The 'sidestep' is not concession—it's the only non-capital-destroying strategy against a vertically integrated, state-subsidized competitor.