Cellular outage in U.S. hits AT&T, T-Mobile and Verizon users

Original Article ↗ Hacker News Discussion ↗

A widespread AT&T cellular outage in the U.S. left many users with “SOS only” service, disrupted some 911 and FirstNet emergency communications, and generated confusion as official status pages initially showed no problems while sites like Downdetector spiked with reports. Commenters debate likely causes—from shared telecom infrastructure and botched software or network updates (later cited by AT&T) to less plausible theories like solar flares or cyberattacks—and compare the event to previous large-scale outages such as Canada’s Rogers incident. The incident is used to highlight weaknesses in SMS-based authentication, the opacity and marketing spin of status pages, and the broader fragility of critical communications infrastructure that society increasingly depends on for everyday navigation, payments, and emergency response.

Scope and Symptoms of the Outage

  • Many AT&T customers across the U.S. reported “SOS only,” failed calls, SMS issues, and unusable mobile data; Wi‑Fi calling often still worked.
  • Impact was inconsistent: some phones on the same family plan, in the same car or house, had service while others did not.
  • Reports mentioned both physical‑SIM vs eSIM differences and LTE vs 5G differences, but patterns were anecdotal and inconsistent.
  • Some MVNOs on AT&T were unaffected, others were fully down.
  • Downdetector and similar sites showed spikes for multiple carriers, but several comments argued this likely reflected people unable to call AT&T users rather than true multi‑carrier outages.
  • Timing: many noticed issues starting around 3–4am local time, with partial restoration hours later and staggered recovery by region/device.

Cause and Technical Speculation

  • Early speculation ranged from SS7 problems, BGP misconfigurations, shared hardware/software bugs, to solar flares or cyberattacks.
  • Several telecom‑experienced commenters emphasized that nationwide issues can arise from mundane configuration errors, routing changes, or software updates on core equipment.
  • Later in the thread, AT&T’s public position (via news reports) is cited: a software update / “incorrect process” during network expansion, not a cyberattack.
  • Some references to Cisco and core databases of SIM/authorization data are mentioned as rumors; details remain unclear.

Emergency Services and Critical Infrastructure

  • Some AT&T users reportedly could not reach 911 despite phones showing “SOS only.”
  • Others stated that, in principle, emergency calls should fall back to any available network, but acknowledged this can fail depending on failure mode.
  • Conflicting reports about FirstNet (public‑safety network): some say AT&T claimed it was unaffected, others in 911 centers reported FirstNet devices dropping VPN and losing mobile service.

Status Pages, Monitoring, and Transparency

  • Strong criticism of carrier status pages: often show “all clear” during obvious nationwide incidents or are updated late and vaguely.
  • Some argued third‑party monitors (e.g., Downdetector, ThousandEyes) are now more useful than official pages, while others noted their limitations and bias toward user perception.
  • Desire for better, automated, per‑customer impact dashboards and real postmortems; skepticism that large telcos will publish meaningful RCAs.

Broader Reflections: Reliability, Design, and Policy

  • Discussion of how dependent society is on cellular for 2FA, payments, navigation, logistics, and public safety.
  • Calls for:
    • Less SMS‑based 2FA.
    • More offline‑first app design and offline maps.
    • Stronger regulation around outage reporting and 911 reliability.
  • Some contrast past Bell System reliability and decentralized, fault‑tolerant design with today’s more complex, centralized, software‑driven cores.