The Fusion Series · Part 3 of 6

The four-tower model wasn’t a mistake. It was built for growing revenue, cheap capital, and a deep bench of veterans. All three have turned — and the structure built on them is the last thing still standing.

The four separate round-the-clock teams I have been describing — the ones Part 1 proposed fusing and Part 2 priced — were not a mistake. They were the right structure for a specific world: average revenue per user rising, capital cheap, and a deep bench of engineers who knew the gear. Change any one of those and the model strains. All three changed at once.

The money changed

5G was sold as a growth engine and did not deliver one. Even as total service revenue grows, blended ARPU is going the wrong way: Omdia projected average revenue per user falling 4.2% between 2022 and 2027, and called the apparent uplift from unlimited 5G bundles an illusion. Capital intensity peaked around 18% in 2022, and capex contracted in 2023 for the first time since 2017. And the debt that funded the whole build-out got more expensive the moment rates rose. The growth era ended and the efficiency era began — not as a slogan, but as a balance-sheet fact.

The cutting already started — bluntly

The largest operators have already accepted that the answer to flat revenue is a smaller cost base. AT&T’s workforce went from roughly 280,000 in 2017 to about 141,000 by early 2025 — nearly 139,000 fewer people. Verizon disclosed about 4,800 separations under a single voluntary program, against a severance charge of $1.7–1.9 billion. Equipment makers cut too. This is not a forecast; it is public record.

What the industry has mostly done, though, is cut with a blunt instrument — across-the-board reductions that take the good people out along with the redundant ones. That approach has a fatal flaw, and it is the third thing that changed.

The industry has already decided to get smaller. The only open question is whether it cuts with a blunt instrument or a scalpel.

The bench is thinning — and you cannot hire your way out

Telecom is hit harder than almost any industry by the retirement wave, because so much of the network is old and so much of the expertise is undocumented, concentrated in one or two veterans per site. By industry estimates, more than half of the professionals who understand legacy systems have already retired, and fewer than one in five new hires can work on them. You cannot hire a twenty-five-year-old who knows your SONET/SDH ring or your TDM switch — the vendor is gone, and you buy spares on eBay. When that veteran retires, the knowledge retires with them.

So the squeeze comes from three sides at once. Revenue will not grow into the cost. Capital will not subsidize it. And you cannot simply hire the people who hold it all together. The four-tower model assumed all three of those conditions would hold. None of them do.

This is where the industry is already heading

The efficiency era is not rhetoric — labor operating expense finally fell 2.2% in 2024 as the first workforce transformations landed. And the direction of travel is not a Rapax idea. Gartner forecasts that by 2026, three-quarters of organizations with operational-technology environments will merge their network and security operations. Splunk has predicted that AI will fold the NOC and the SOC into a fusion center so leaders can redeploy talent instead of maintaining large siloed teams. The question is no longer whether telecom runs its operations leaner. It is how — as a deliberate strategy, or as an annual round of blind cuts.

If the model has to change, and cutting blindly is the dangerous way to do it, then the right version has to start with a reframe most leaders get backwards: the goal is to redeploy your best people, not remove them. That is Part 4.

By Shawn Ennis — Founder & CEO of Citus Technologies and founder of Rapax. Twenty-five years in telecom operations; founder of a service-assurance platform acquired by Oracle; 12+ patents in OSS/BSS software.