There are copper phone lines in this country billing more than $3,300 a month. Not $33. Thirty-three hundred, per line, documented across ten states in MetTel's July 2026 Copper Retirement Bulletin. That is not a price. That is a carrier telling you to leave, politely, in the only language a rate sheet speaks.

Most of the writing about the copper sunset treats it as a facilities problem — fire alarm panels, elevator phones, NFPA 72, ASME A17.1, all real and all somebody else's budget. I want to talk about the part that lands on the voice and marketing side of the house, and that almost nobody is planning for: copper retirement is a forced carrier migration, and carrier migrations are where phone numbers — and every record keyed to them — quietly break.

What actually changed in March

On March 26, 2026, the FCC adopted Report and Order FCC 26-19. Before it, retiring a copper wire center meant a Section 214 discontinuance application, a public comment period, federal sign-off, and a separate mechanism that let state regulators formally challenge the retirement. That could stretch a timeline by many months. The order eliminated the approval bottleneck, created a uniform 31-day automatic-approval window for discontinuance applications, and asserted preemption of state rules that would delay a federally authorized retirement.

The practical floor is now roughly 90 days of public notice. Ninety days is plenty of time to execute a plan. It is not enough time to write one.

The volume backs this up. AT&T published 16 copper-retirement notices in 2019 and 824 in 2026 — up 51× in seven years and 3.4× since 2024 alone. It grandfathered legacy copper voice across its wire centers effective October 15, 2025 (no new orders, existing lines on climbing rates), was approved to discontinue service across more than 30% of its copper footprint during 2026, and began permanently decommissioning roughly 500 wire centers in June. Its published Discontinued Service Areas workbook lists more than 100,000 location entries across 21 states. Verizon, Lumen, Frontier, and CenturyLink are all filing their own. Meanwhile the FCC counted 15.0 million end-user switched-access lines as of June 30, 2025, down from 27.2 million in mid-2022 — a 17.9% compound annual decline, a base that halves roughly every four years.

The numbers on those lines are not spare parts

Here is the part that gets skipped. A meaningful share of surviving copper DIDs are advertised numbers. They're on the trucks, the yard signs, the door hangers, the directory listings, the ten-year-old print buy somebody still gets calls from. They were provisioned before anyone on the current team was hired, and they've been quietly answering ever since.

When the notice arrives, those numbers get ported under duress. Four things break at that boundary, and I've watched every one of them happen:

  1. History discontinuity. If your call records live in the losing carrier's portal — and for legacy lines they usually do — the number moves and the history doesn't. You keep the digits and lose the baseline. Six months later nobody can answer "did volume on this line drop, or did we just start counting differently?"
  2. Routing logic that exists nowhere. Hunt group order, rollover sequence, after-hours forwarding, the weird Tuesday exception — that behavior lives as undocumented config in a carrier switch. Porting the number ports the number. It does not port the behavior. Somebody rebuilds it from memory, usually the week of cutover.
  3. Record hygiene on the way out. CNAM entries, directory listings, and dispatchable-location records for E911 were provisioned against the old circuit. They need re-provisioning per number, and nothing in the port order does it for you.
  4. Ports rejecting for boring reasons. A simple port is supposed to complete in one business day under FCC rules. Complex and multi-line ports reject over account-number and service-address mismatches against a CSR that was accurate in 2011. And the losing carrier — the one winding the network down — has the least staffing and the least incentive to chase your paperwork.

The Dial800 angle

My bias is structural, so I'll state it as a bias: this gets easier when numbers, routing, and the call record are the same system instead of three vendors with a port order between them.

That's the argument for running business voice on a platform that also owns the number inventory. Routing lives in AccuRoute as configuration you can read, export, and diff — not as switch state you rent and rediscover during a cutover. Reporting lives in CallView360°, keyed to your call record rather than a carrier's billing system, so a transport change doesn't reset the time series. And because the same record carries source and campaign data, you can still answer the attribution question about a number that just changed carriers mid-quarter.

None of that makes the port faster. It makes the port boring, which is the only outcome worth engineering for.

Do this before the notice arrives

Pull your full DID inventory. For each number, mark three things: is it advertised anywhere you don't control, which carrier actually owns it today, and does it have routing behavior that exists only inside a carrier portal. Every number with two or three marks is a migration project with a 90-day fuse already lit.

The copper sunset isn't going to ask your opinion. The only variable left is whether you're migrating on your calendar or theirs.