For more than a decade, the phrase "dumb pipe" has been the telecom industry’s most uncomfortable truth. Operators built the highways of the digital economy, laying fibre, deploying spectrum, and engineering 5G networks, only to watch the value accrue to the platforms riding on top of them. Netflix, Google, Meta, and Amazon generate billions in revenue from content and advertising delivered over networks that operators built and maintain, whilst those operators compete on price for a commodity service.
The data tells the story starkly. Global telecom service revenue is projected to grow at a modest 2.8% CAGR to $1.32trn by 2029, whilst mobile ARPU is expected to tick down marginally to $6.20 in 2029 from $6.32 in 2024 [1]. The structural challenge is real: usage is soaring, but willingness to pay is flat. Something has to change.
The good news is that something is changing. A growing number of operators are successfully executing the transformation from connectivity utility to intelligence platform. The playbook is becoming clearer, and the results are measurable.
Step 1: Recognise the Asset
The first step is recognising that the network itself is a data asset of extraordinary value. Every day, mobile networks generate petabytes of high-frequency, real-time data, including location signals, device behaviour, usage patterns, application preferences, and movement data. This data is consented (subscribers agree to terms of service) and it is network-verified, meaning it cannot be spoofed or fabricated.
Vodafone Analytics, which packages anonymised location data for retailers, event organisers, and city planners, is a prime example of an operator recognising and monetising this asset [2]. Orange Business has gone further, building a suite of big-data and AI-based solutions for enterprise clients across finance, healthcare, and logistics [2].
Step 2: Build the Architecture
Recognising the asset is not enough. The technical challenge of turning raw network data into monetisable intelligence is substantial. Modern telecom networks generate data from Call Detail Records (CDRs), OSS/BSS systems, Deep Packet Inspection (DPI) solutions, location systems, and subscriber-management systems. These streams are complex and often siloed [3]. Building the data ingestion, transformation, and storage pipeline required for effective analytics is a multi-year, capital-intensive undertaking.
We’ve spent years building the foundation for this shift. Our AI-based data-monetisation platform doesn’t just ‘process’ data; it enriches it, categorises it, and monetises it with precision. We are already seeing a 5% ARPU increase for our partners by moving away from indiscriminate billing toward intelligent, AI-driven engagement. The future of telco revenue isn’t in charging more for less; it’s in understanding the intent behind every byte of data.
For operators who want to move faster, the alternative is to partner with specialised platforms that bring pre-built ML infrastructure. The key architectural principle is that raw personal data must never leave the operator’s secure perimeter.
In-network ID" approaches, where machine-learning engines operate within the operator’s environment and generate pseudonymous tokens representing audience attributes, allow operators to participate in the programmatic advertising ecosystem without exposing subscriber PII. This approach satisfies both regulatory requirements and advertiser demand for high-quality, fraud-free audience data.
Step 3: Operationalise the Revenue
The final step is the hardest: turning data intelligence into recurring, scalable revenue. The most successful models are performance-based. Rather than selling raw data, which creates regulatory and reputational risk, operators are delivering targeted campaigns (via SMS, RCS, IVR, or in-app) on a cost-per-acquisition (CPA) or revenue-share basis. This aligns the operator’s incentives with the advertiser’s outcomes and eliminates the risk of data misuse.
Operators using this approach have demonstrated ARPU increases of up to 5%, which is a significant achievement in a market where ARPU is otherwise declining. The key is the combination of machine-learning precision (identifying the right audience at the right time), channel optimisation (delivering via the most effective medium), and continuous feedback loops (using campaign results to improve future targeting).
The transformation from dumb pipe to intelligence platform is not a destination; it is a continuous journey. But the operators who are furthest along that journey are already demonstrating that the data flowing through their networks is worth far more than the connectivity fees they charge to carry it.
Aleksandr Moskotin is Co-Founder & CEO of afina, an AI/ML data-monetisation platform
References
[1] PwC, Global Telecom Outlook 2025–2029 — www.pwc.com/gx/en/industries/tmt/telecom-outlook-perspectives
[2] Avenga, How to Generate Value with Data Monetization in Telecom — www.avenga.com/magazine/how-to-generate-value-with-data-monetization-in…
[3] IBM, AI in Telecommunications — www.ibm.com/think/topics/ai-in-telecommunications
