Why telcos can’t change into tech corporations (Analyst Angle)

It’s an concept that sounds proper — however is it attainable?

Each few years, one other telecom technique presentation declares that “our purpose is to change into a tech firm.” The phrase sounds inspiring. It alerts ambition, innovation, and a want to flee regulatory gravity. However it’s fallacious and probably value-destructive.

The distinction between a telecom firm and a know-how firm shouldn’t be tradition, capital, or management. It’s physics. Telecom belongings behave in a different way from software program belongings. They obey completely different price curves, scale legal guidelines, and threat profiles. Whenever you ignore these variations, you get methods that sound visionary however destroy worth.

If telecom leaders need to perceive the place actual transformation lies, they should begin from first rules, not PowerPoint analogies.

Scope

This dialogue focuses on cell community operators. Fastened-line fiber networks face completely different constraints: greater per-premises capital prices, even higher geographic fixity, and easier operations as soon as deployed. The rules of asset fungibility, working leverage, and price scaling apply to each. Wi-fi networks dominate international subscriber and income development, and the interaction between spectrum, protection, and mobility creates the sharpest distinction to hyperscalers.

I. The physics of hyperscalers versus telcos

Hyperscalers and telecom operators each construct massive networks of infrastructure. However the best way these networks generate return on capital is basically completely different.

In cloud infrastructure, capability might be redeployed wherever on the planet with virtually no friction. Spare compute in Oregon can serve prospects in Singapore. The power to reassign belongings dynamically, what I’m calling fungibility, turns mounted price right into a lever for exponential development.

Telecom networks, against this, are anchored to geography. A radio website in Saskatchewan can’t serve a buyer in São Paulo. As soon as constructed, it’s immovable. Each cell website, trench, and spectrum license is a sunk price outlined by location. Utilization is bounded by native demand.

That’s the first legislation of telecom economics: telecom is spatially constrained infrastructure.

II. 4 legal guidelines of telco economics

Telecom networks are ruled by bodily legal guidelines that outline how capital turns into capability, and capability into money circulate. These legal guidelines will not be strategic decisions. They’re structural realities that form each funding resolution, pricing mannequin, and transformation effort within the trade. Understanding them separates technique from storytelling. Every legislation captures a constraint that software program economics can bend however not break: geography limits utilization, spectrum scales in steps, protection expands by duplication somewhat than pooling, and integration determines whether or not these prices compound or cancel. Collectively, they clarify why telecom returns flatten as networks develop whereas hyperscaler returns speed up with scale.

Legislation 1: Geography defines utilization

A cell website price is generally mounted, whereas the income it could seize is determined by native density and on the fraction of every consumer’s visitors that passes by way of that website. Busy-hour throughput usually runs 4 to 6 occasions the day by day common, producing low time-averaged utilization. In lots of mature markets, time-averaged RAN utilization sits within the 15–30% band; hyperscaler knowledge facilities usually attain 70–90% by pooling uncorrelated workloads.

Labored instance, traffic-weighted attribution (clear assumptions)

• ARPU = $50 per thirty days = $600 per yr.
• Visitors fraction per website = 10%. This displays customers traversing many websites per day.
• City website customers = 300. Rural website customers = 30.

Step math

1. Per-user annual ARPU = $600.
2. Website-attributable income per consumer = $600 × 0.10 = $60.
3. City website income = 300 × $60 = $18,000 per yr.
4. Rural website income = 30 × $60 = $1,800 per yr.

Typical direct website OpEx (lease, energy, upkeep, discipline visits) runs $80,000–$200,000 yearly, relying on market and website complexity. The arithmetic reveals why particular person websites can seem loss-making even when the community as an entire is worthwhile. Altering assumptions strikes the image. Elevate visitors share to twenty% or ARPU to $100, and per-site numbers enhance. Nevertheless, the structural level holds: per-site economics are extremely delicate to visitors share, consumer density, and busy-hour skew, and spare capability in low-density areas can’t be redeployed to higher-value places.

Concrete operator examples

• City macro in London or New York: excessive subscriber density and heavy daytime focus produce a lot greater utilization than distant rural websites in northern Canada or central Australia.
• Stadiums and transit hubs: quick length excessive depth. These websites require provisioning for peak however have low time-averaged utilization, driving price per helpful byte a lot greater than in data-center-like environments.

Legislation 2: Spectrum economics observe step features

Spectrum is purchased in lumpy chunks. Examples: the U.S. C-band public sale raised roughly $80 billion in 2021. T-Cellular paid about $9 billion for huge 2.5 GHz holdings in earlier mid-band acquisitions. As soon as bought, spectrum yields near-zero marginal price per further consumer till the community hits congestion. At that time the operator faces a discrete set of funding choices: add carriers, densify with small cells, improve backhaul, or purchase extra spectrum. These choices include six-figure to multi-million-dollar value tags per website or license.

Concrete illustration

• A metro sector could function at 60% busy-hour utilization and add customers with negligible further CAPEX. When utilization passes an empirical threshold, for instance, 80–85%, the operator should densify. Small-cell deployments can price $50k–$200k per website put in in city settings when mounting, fiber, energy, and permits are included; a macro website densification or a brand new service deployment can exceed $200k–$500k in complicated environments — the economics soar, not slope.

Investor implication

• The step-function creates lengthy plateaus of engaging marginal economics adopted by sudden reinvestment cliffs. Forecasting returns requires modeling these items, not assuming easy marginal price curves.

Legislation 3: Networks scale by protection, not concurrency

Hyperscalers add nodes to extend pooled capability. Telcos add websites to extend attain. The economics differ as a result of incremental protection prices rise as density falls.

Concrete examples

• Rural FTTH prices: $2,000–$5,000 per premises in low-density territories versus $500–$800 in city areas.
• Nationwide protection: the primary 90% of inhabitants protection usually captures the densest, most cost-effective customers. The final 10% usually incorporates the costliest to serve per buyer. In a number of OECD nations, reaching the ultimate 5–10% can price three to 5 occasions extra per consumer than the prior protection tranche.

Operational impact

• Workforce geometry issues. Ten thousand urban-adjacent websites might be serviced effectively from a handful of depots. Ten thousand broadly dispersed rural websites require many extra crews, regional depots, and better journey time overhead, rising OpEx per website.

Aggressive distinction

• Hyperscalers select areas with demand focus. Telcos should serve geography as a result of shoppers are geographically distributed and regulators usually mandate common service.

Legislation 4: Integration creates worth; isolation exposes price

Networks are systemic. Integration throughout spectrum, RAN, core, and backhaul creates operational synergies. Disaggregation can reveal hidden asset worth, however it additionally transfers worth to 3rd events and introduces coordination friction.

Examples and mechanics

• Tower monetization. Many operators bought towers to TowerCos like American Tower or Crown Fortress to free capital. This improves stability sheets and converts CAPEX into predictable lease prices. It additionally transfers upside from tower appreciation to tower buyers and imposes recurring hire. The web impact shouldn’t be uniformly useful to working economics; it is determined by how effectively operator leasing phrases, upkeep coordination, and improve home windows are negotiated.
• Impartial-host fashions. Airport DAS or subway neutral-host programs scale back duplication and decrease complete price for carriers, however they require a governance mannequin and infrequently a third-party operator to run the shared system. The positive aspects are actual in dense venues, however they don’t make rural websites fungible.

Operational warning

• The extra you modularize the stack, the extra you commerce systemic optimization for discrete accounting readability. Integration permits you to optimize truck rolls, spare components swimming pools, and grading of capability throughout layers. Isolation turns these levers into contractual processes and tends to scale back the efficient utilization advantages that built-in programs seize.

Collectively, these legal guidelines clarify why telcos and tech corporations stay on reverse sides of the return-on-capital spectrum.

Hyperscalers flip compute elasticity into rising returns (α > 1): Every unit of infrastructure helps ever-greater workloads. Telcos, constrained by physics, face diminishing returns (α

III. Why the “tech firm” analogy fails

When executives say, “We have to assume like a tech firm,” they normally imply “we want recurring income and scalable margins.” Software program economics don’t map onto community physics.

Software program scales by way of replication: each new buyer makes use of the identical code.

Networks scale by way of duplication: each new buyer requires extra {hardware} in a selected place.

You possibly can virtualize components of the telco stack, however you can not virtualize geography. Digital transformation, comparable to AI in operations, APIs for publicity, and data-driven pricing, can enhance effectivity, however it doesn’t change the denominator of the enterprise: kilometers, kilowatts, and capital depth. The telco drawback shouldn’t be digital immaturity; it’s structural non-fungibility.

IV. Strategic playbook for telcos

Telcos can’t change into hyperscalers. Their belongings are sure by geography and physics. However they will evolve into hybrid infrastructure operators with platform formed traits. By this, I imply companies that exploit leverage the place scale physics enable it (shared infrastructure, AI-enabled effectivity, spectrum reuse) and innovate digitally the place abstraction is feasible (API publicity, service orchestration, enterprise integration).

The purpose is to not imitate cloud economics, however to industrialize the airwaves: turning protection, capability, and connectivity into programmable infrastructure that others can construct on. Telcos won’t ever scale just like the cloud, however they will make the community behave extra like software program. That is sufficient to change the sport.

1. Monetize utilization development on mounted belongings

• Fastened Wi-fi Entry (FWA): Makes use of present spectrum and RAN so as to add 20–40% incremental income. Instance: T-Cellular FWA serves ~5 million subscribers, producing $3–4 billion yearly. Marginal price is primarily CPE and modest backhaul upgrades.
• Enterprise Personal Networks: Larger ARPU for devoted slices of spectrum on shared infrastructure. Relevant to manufacturing, logistics, campuses.
• Wholesale and MVNOs: Promote extra capability to resellers and digital operators. Smooths utilization, generates income with near-zero marginal price throughout off-peak hours.

2. Pursue shared infrastructure aggressively

• Lively RAN sharing: Deploy one community to serve a number of operators, reducing price per busy-hour Mbps by 15–25%. Examples: Cornerstone (UK), MBNL (UK).
• Impartial-host preparations: Pool rural protection or indoor programs (airports, stadiums, subways) to amortize deployment prices.
• Tower divestments: Convert owned towers to sale-leaseback preparations. Improves stability sheet and reduces capital depth, although long-term appreciation is forfeited.

3. Leverage AI and convergence applied sciences

Some applied sciences genuinely shift telco economics towards hyperscaler traits, although solely partially.These improvements make operations smarter inside mounted physics. AI-RAN, as an illustration, can enhance utilization 10–20% and scale back working bills 15–25%, primarily by higher capability allocation and predictive upkeep. However they don’t change the truth that protection belongings can’t transfer.

Convergence, in different phrases, principally reshuffles worth somewhat than creates new elasticity. It might probably enhance margins and optionality however can’t convert mounted protection into international scalability.

4. Competing on integration, not infrastructure

If each operator pursues the identical hybrid construction, differentiation can’t come from proudly owning capability alone. It should come from orchestrating expertise. Integrating mobility, compute, edge-services, and reliability into propositions that prospects worth.

Enterprise differentiation is turning into seen in current outcomes. For instance, Verizon Communications reported its Enterprise section working earnings rose 12.7 % year-over-year in Q3 2025, and section EBITDA margin elevated to 23.4 %. These figures counsel that enterprise companies (which embrace personal 5G, edge compute, and managed community choices) are contributing extra profitably, despite the fact that general Enterprise section income declined 2.8 % in Q3. In the meantime, Singtel’s Digital InfraCo and its Paragon platform acquired a 2025 “Aggressive Technique Management” award from Frost & Sullivan for its GPU-as-a-Service providing leveraging edge, compute and community infrastructure. These milestones assist the thesis that integration of community, compute and ecosystem is gaining traction. Nevertheless, publicly-disclosed proof nonetheless falls wanting exhibiting full platform-scale economics (for instance, incremental EBITDA per greenback of latest capex stays undisclosed).

Client differentiation stays extra nascent however is supported by current outcomes. Verizon’s Q3 2025 Client wi-fi service income rose 2.4 % year-over-year, and ARPA (common income per account) reached roughly $147.91, up ~2.0 %. These outcomes point out that service high quality and bundling could also be contributing to modest ARPA development. On the Singtel facet, whereas detailed shopper ARPU uplift is much less publicly disclosed, launches comparable to its 5G+ Cellular Workspace and integrations with enterprise gadgets sign a method to increase past connectivity.

Placing this collectively, the rising sample is that operators who can reliably combine connectivity, edge compute, companies and ecosystem partnerships are starting to outperform segments of their enterprise. However the thesis have to be calibrated:

• The outcomes are stronger in enterprise/personal community segments than in consumer-bundle segments.
• Most operators nonetheless don’t publish the complete set of transformation indicators (e.g., price per busy-hour Mbps, income per deployed website, incremental EBITDA per capex greenback).
• Shared infrastructure (towers, RAN, spectrum) stays a value burden; the differentiation lever shouldn’t be new belongings however new experiences constructed on present belongings.

Operators that grasp integration and reliability can have a aggressive edge even when underlying infrastructure is shared. The win doesn’t come from proudly owning extra towers however from making connectivity invisible and outcomes seen.

5. Measure what drives transformation

Cease monitoring transformation by agile squad counts or innovation lab sq. footage. Measure outcomes:

If these metrics enhance, transformation is actual. If they don’t, you might be working theater. Tradition applications and agile coaching don’t matter if economics don’t change.

V. The true constraint and the true alternative

Telcos serve geography. Hyperscalers serve workloads. Geography is mounted. Workloads are cell. That’s the boundary.

The chance is to industrialize the airwaves the place obligatory and innovate digitally the place attainable:

• Spectrum: Excessive mounted price, near-zero marginal price, logical flexibility
• Core: Centralized, scalable, robust economies of scale
• RAN: Distributed, coverage-bound, shareable, incrementally monetizable
• Providers: Asset-light, software-defined, platform-enabled

Hyperscalers scale code. Telcos scale protection. Technique fails while you confuse one for the opposite.