On May 12, 2026, the FCC did something it has not done in three decades: it tore down the regulatory architecture that governed how loud satellites could transmit. The decision was announced alongside approval of SpaceX's $17 billion purchase of 65 megahertz of mid-band spectrum from EchoStar, but the spectrum deal is the easier story to tell. The real rupture is the replacement of the Equivalent Power Flux Density (EPFD) framework, which since the late 1990s served as a hard ceiling on energy transmitted by non-geostationary orbit systems, with a voluntary coordination model that lets operators negotiate their own limits. This is not a tweak. This is the FCC removing a 30-year-old constraint that every NGSO satellite company has had to design around.
The landscape context matters here. Satellite connectivity has been fragmented into two camps: geostationary operators (like EchoStar, Viasat, Intelsat) that controlled mid-band and Ka-band spectrum but moved slowly, and NGSO constellations (SpaceX's Starlink, Amazon's Project Kuiper, OneWeb) that moved fast but operated under strict power limits meant to protect the incumbents. The EPFD framework was that protection, it literally defined maximum power density over any given point on Earth. Scrapping it transforms the regulatory basis of competition. SpaceX and Amazon have already received individual EPFD waivers (January 9 and February 20, 2026, respectively), but those were exceptions. The May 13 Federal Register publication of FCC 26-26 makes the exception the rule.
The 65 megahertz SpaceX acquired spans AWS-4, AWS-H Block, and unpaired AWS-3 licenses. The FCC projects this will enable a capacity increase of more than 100 times compared to first-generation Direct-to-Device systems. That is not exaggeration for a press release, it reflects the enormous difference between operating under EPFD constraints and operating under negotiated power agreements with neighboring GSO operators. AT&T acquired 50 megahertz of the same spectrum for terrestrial 5G use. The rules become effective July 13, 2026, with equipment certification waivers already granted to accelerate device compatibility. Commercial D2D services, messaging and emergency alerts, are targeted to begin in the final quarter of 2026. This is not a roadmap. It is a regulatory green light paired with hardware already in orbit.
What enabled this moment was the combination of three factors colliding at once. First, SpaceX's Starlink constellation actually works, it is transmitting usable signal to consumer terminals in real time, which means the company could credibly propose a D2D service without it being theoretical. Second, Amazon's Project Kuiper forced the FCC's hand by demanding equivalent EPFD relief, creating pressure to move from exception-granting to rule-making. Third, FCC Chairman Brendan Carr's stated priority was spectrum efficiency and competition, not spectrum scarcity. The old EPFD framework was designed to prevent interference by limiting NGSO power. The new framework trusts operators to coordinate. That philosophy shift explains why the deal closed fast and why the Federal Register publication happened just one day after the spectrum approval announcement.
Who wins and who loses is asymmetric. SpaceX gains spectrum and regulatory relief simultaneously, a combination that unlocks Direct-to-Device as a genuine consumer service, not a novelty. Amazon gains certainty that its similar waiver will hold and that the operating environment will not suddenly change. Existing satellite operators like Viasat and Intelsat lose the artificial power ceiling that protected their legacy spectrum from newer competitors. Terrestrial wireless providers do not lose anything here, 5G operates in a completely different part of the spectrum, and AT&T's acquisition of 50 MHz actually strengthens its midband portfolio. The real question is whether the removal of EPFD limits leads to more interference than the FCC's coordination model can handle, and whether voluntary negotiation between competitors actually prevents conflicts or just pushes them into litigation.
Three signals will tell whether this unfolds as planned. First, whether SpaceX actually deploys D2D messaging services on unmodified iPhones and Android phones by December 2026, not early access, not beta, but advertised commercial service. Second, whether Amazon achieves parity on D2D capability by 2027, which would confirm that the regulatory change benefited new entrants symmetrically. Third, whether the FCC's coordination model prevents actual interference complaints within the first 18 months of the new framework, or whether disputes between NGSO and GSO operators escalate to the point where ad-hoc waivers are replaced by new hard rules. The architecture is now in place. The question is whether it actually works.