Just over two years after I wrote about Optus going quiet, it was America’s turn.

From around midday Eastern on 14 January, Verizon customers across the US saw “SOS” where their signal bars should have been. Voice, text and data were down for roughly ten hours, until service was restored shortly after 10pm Eastern. Verizon apologised, offered a $20 credit, and said this was “a software issue”, with no indication of a cyberattack. The FCC has opened an investigation, and many telecoms commentators suspect a bad update in the 5G standalone core.

“Software Issue” Isn’t an Answer

I understand why companies are cautious in the first week. Facts are still being established, there are regulators to brief, and lawyers have opinions. But “software issue” tells customers almost nothing. Nearly every modern outage is a software issue. The questions people actually care about are:

  • What changed? An update, a configuration push, a certificate expiry, an automated process?
  • Why did it spread so widely? Where were the blast-radius controls?
  • Why did it take ten hours? What slowed detection, diagnosis or rollback?
  • What stops it happening again?

The best post-incident reports I’ve linked on this blog — AWS, Cloudflare, Google — answer all four. That openness builds more trust than any credit on a bill.

The Part That Worries Me Most

When phones show SOS, they can usually still reach emergency services over another carrier’s network, but many people don’t know that. Telcos aren’t just another app. When they go down, people can’t call for help. Having served with the Irish Coast Guard, I know how much depends on people being able to raise the alarm. Network resilience is ultimately a public safety issue.

More Information