The telephone network you grew up with — the one with physical wires, dedicated lines, and the sound of clicks as calls connected — is being switched off. Not metaphorically. Literally. Carriers around the world are decommissioning the infrastructure that Alexander Graham Bell built, replacing it with something fundamentally different.
Here is what that means, why it is happening now, and why it matters.
The Network That Required Copper and Faith
The Public Switched Telephone Network — PSTN — was an engineering marvel of its era. The Bell System built it over decades, layer by layer, from the first copper wire in 1876 to a continent-spanning web of switches, relays, and physical lines that could connect any two telephones on Earth.
The architecture was elegantly simple, and fundamentally inefficient: every phone call required a dedicated electrical circuit between the two parties for the entire duration of the call. When you were on a one-hour call to Hong Kong, that physical circuit was reserved for you — occupied — for the full sixty minutes. No one else could use those wires, that switch capacity, that path through the network. You paid for that reservation, whether you were speaking or not.
Each voice conversation traveled as an analog electrical signal across copper wiring, converted by codecs into 64 kilobits per second of digital data. That 64 Kbps was a global standard — enough to carry a human voice with reasonable fidelity. The switches that routed these circuits were physical machines: Class 4 switches for long-distance routing, Class 5 switches for local calls, all connected by a signaling layer called SS7 (Signaling System 7) that told the switches where to connect.
It worked. For 100 years, it worked remarkably well. But the architecture had fundamental limits.
The Problem With Dedicated Circuits
Circuit-switched networks scale by adding physical infrastructure. Every new subscriber needs a physical line to a local switch. Every new long-distance route needs dedicated trunk lines between switches. When a major carrier needed to expand capacity, the lead time was months: plan the route, dig the trenches, lay the cable, install the hardware, commission the switches.
This worked fine when phone usage was bounded by the number of physical phones in existence. But data networks changed the comparison permanently.
An IP network — the kind that carries your web traffic and email — is fundamentally different from a circuit-switched network. Instead of reserving a dedicated path for each conversation, IP networks divide every conversation into small packets. Each packet takes the best available route through the network at that moment. Millions of conversations share the same physical infrastructure simultaneously, each taking different paths, each using bandwidth only when actively transmitting.
The efficiency difference is not incremental. It is an order of magnitude. A circuit-switched network at full utilization might carry 10,000 simultaneous voice calls on a fiber link. The same fiber link, carrying packetized voice over IP, can carry 100,000 calls with better quality management and lower latency.
The Costs That Made Change Inevitable
The old telephone network was expensive to maintain in ways that are hard to recover. Consider what keeping PSTN running actually costs:
Physical infrastructure maintenance: Copper cables age. Moisture gets into conduits. Switches require power, cooling, and physical repair. The average age of copper infrastructure in developed markets is now 30–40 years. Replacing and maintaining it is expensive and getting more expensive every year as fewer technicians know how to work with it.
Per-minute billing is a relic: When AT&T charged by the minute for long-distance calls in the 1980s, a call from New York to London cost several dollars per minute. Early VoIP providers undercut that by 90% simply by routing voice as data packets over the internet. The economics of circuit-switched long-distance were never going to survive contact with packet-switched data networks.
Feature asymmetry: PSTN was designed for voice. Adding video, screen sharing, instant messaging, call recording, automatic transcription, and CRM integration to a circuit-switched network requires bespoke engineering — expensive, slow, and limited. On an IP network, all of these features are software problems, not hardware problems.
The Global Shutdown Timeline
Telecom carriers did not switch to VoIP because it was philosophically preferable. They did it because maintaining two parallel networks — the old copper PSTN and the new IP infrastructure — is economically unsustainable.
The shutdown of PSTN is now a documented, scheduled event across most of the developed world:
- United Kingdom: BT announced the PSTN switch-off for 2025, with all-analog services being migrated to IP-based networks. The UK’s transition to digital voice services has been underway since 2020.
- Australia: NBN Co and Telstra announced the PSTN withdrawal program, targeting completion in 2024–2025 for most areas.
- Germany: Deutsche Telekom announced a timetable for shutting down ISDN (the digital PSTN variant) with full IP migration as the replacement.
- Japan, Singapore, South Korea: All have active migration programs from PSTN to all-IP networks, driven partly by the need to retire aging infrastructure before major sporting events.
- United States: While less centrally coordinated, major carriers including AT&T and Verizon have deprecated their PSTN infrastructure significantly. AT&T has stated its intention to phase out copper services in favor of IP-based alternatives.
The common theme: copper PSTN is no longer being invested in. Carriers are not building new circuits. When a business orders a new telephone line today, the carrier is routing it over IP — the old circuit-switched infrastructure is being maintained only for legacy customers, and only until they migrate.
COVID as the Final Accelerant
The pandemic did not cause the PSTN transition. It accelerated it by years.
When offices closed in 2020, desk phones — connected to physical PBX hardware in a server room — became anchors rather than assets. Remote workers needed to make and receive calls from their personal devices and laptops. VPN connections to office networks were slow and unreliable for real-time voice traffic.
Businesses that had already migrated to VoIP systems adapted overnight. Their employees took their softphone apps home, logged into the cloud PBX, and continued making calls as if nothing had changed — because for the IP phone system, nothing had changed.
Businesses still on PSTN scrambled to find workarounds: call forwarding to mobile phones, rushed hardware purchases, manual call routing. The experience demonstrated clearly that an IP-based telephone system was not a nice-to-have modernization but a genuine operational resilience requirement.
Analysts at McKinsey and Gartner both documented accelerated enterprise migration to cloud-based communications platforms in 2020–2021, a trend that has continued as businesses restructured their offices with hybrid work as a permanent model.
What Comes After PSTN
The PSTN switch-off is not an endpoint. It is a transition to something that has already arrived for most users — IP-based voice services collectively called VoIP, delivered through hardware IP phones, softphone applications, or hybrid devices that switch between cellular and WiFi calling.
For enterprise users, the transition has meant adopting platforms like Microsoft Teams Phone, Cisco Webex, Zoom Phone, 3CX, and RingCentral — all of which sit on SIP-based infrastructure. For consumers, it has meant gradual migration to mobile-first communication: WhatsApp calls, FaceTime, and carrier VoLTE (Voice over LTE) services that have already replaced circuit-switched voice for most mobile users.
The telephone network did not disappear. It became data. And becoming data meant it became programmable, scalable, and free from the physical constraints of copper and switches.
The second part of this series examines how that transition actually works technically — and why the protocol that made it possible was designed in 1996, standardized in 1999, and is now the foundation of virtually every voice call made on Earth.