•
Oct 14
Sand Technologies
If you’re a Tier 1 MNO or challenger in the highly competitive U.S. mobile services market—for example, a member of the Competitive Carriers Association (CCA), your approach to network investment and site upgrades is evolving. The traditional model of blanket 5G coverage is giving way to a demand‑driven strategy, where hyper-localized demand profiles justify upgrades and new deployments.
Instead of building where traditional radio frequency (RF) planning tools recommend, primarily driven by coverage objectives and constraints, companies must focus on areas where C‑band or mid‑band 5G upgrades deliver the highest returns, or where competitive and strategic priorities take precedence.
These changes mean that the future of telecom network planning lies in Coverage Intelligence, a new approach that combines real-time, granular competitive benchmarking with sophisticated economic modeling. This strategy redefines how operators think about their networks, from “Is our network good?” to a far more powerful stance: “Where is our network weaker than our competitor’s and what is the ROI of fixing it?”
The evolution of cloud computing capabilities and AI-driven CapEx tools is the key enabler of this transformative strategic shift.
Imagine this scenario: a mobile network operator (MNO) commits hundreds of millions of dollars in capital expenditures (CapEx) to upgrade and/or build sites in a critical market. The RF planning tools and internal dashboards illuminate with green KPIs, indicating improved performance and capacity. Yet, six months later, customer churn in that very area remains stubbornly high. Why? The answer lies just down the street. A competitor, equipped with more precise analytics tools, has densified coverage to deliver a stronger signal and better throughput at locations such as the local grocery store, the community park and the new housing development. The first operator won their internal KPI battle but lost the war for the subscriber experience.
This scenario highlights two key issues: the intense competition in the U.S. mobile services market and a core flaw of conventional network planning. For too long, it has been an inside-out process, focused on meeting internal metrics while overlooking the most crucial context: the subscriber’s experience compared to other available options.
This paradigm shift also means site upgrades or new builds will prioritize outperforming competitors in key markets or points of interest.
Demand for wireless connectivity continues to skyrocket. According to the Cellular Telecommunications and Internet Association’s (CTIA) 2024 Annual Survey Highlights report, Americans now consume over 100 trillion megabytes of data, with volumes growing every year.
Contributing to this data is the growing number of 5G-enabled devices—from smartphones to smartwatches to IoT endpoints—deeply embedding 5G into Americans’ lives. Meanwhile, 5G home broadband fixed wireless access (FWA) is helping to bridge the digital divide, giving households a real choice in connectivity—whether via fiber, cable, or wireless 5G broadband.
This rapid adoption drives intense pressure on networks and, ultimately, the need for more cell sites and more network CapEX to keep up. By the close of 2023, the U.S. had 432,000 operational cell sites, a 24% increase since the 2018 wireless siting reforms, as reported in the 2024 CTIA report.
However, while wireless data consumption continues to grow rapidly, the price per megabyte is declining. Over the past decade, telecom operators have averaged a 2% annual revenue growth, with inflation often erasing real gains. The industry frequently operates below its cost of capital.
Despite large-scale investments into access networks, 5G rollouts have yet to deliver meaningful financial returns. According to Wells Fargo, AT&T, Verizon, T-Mobile and Dish invested over $230 billion in capital expenditures and spectrum from 2020 to 2023; yet, wireless EBITDA grew by only approximately $10 billion, primarily driven by cost-cutting efficiencies rather than new revenue streams.
In any capital-intensive industry, capital expenditure on infrastructure creates a competitive moat: the more extensive the network infrastructure, particularly the radio access network, the harder it is for competitors to catch up. However, the financial realities of the telecom industry necessitate a more strategic approach to capital expenditures (CapEx) deployment.
This objective requires shifting from blanket coverage to demand‑driven deployments. Mobile network operators must:
Given the complexity of C-Band/Mid-Band site upgrades and multi-stage planning cycles, AI/ML-driven analytics are now essential for optimizing network rollouts and making informed, competitor-aware investment decisions.
MNOs are moving toward demand‑driven and competition‑aware deployments, designed not only to cover the map but also to capture market share. This new mindset requires thinking like a competitor, not just a coverage planner. Performance is no longer just about coverage bars—it’s about experience, quality and relative strength.
To operate this way, network strategists must ask the right competitive questions:
These questions represent the shift to a competition‑driven strategy, where winning in strategic markets defines success rather than simply covering ground.
Economic modeling for site upgrades and new site placement in mobile networks involves multiple layers of analysis. For leading MNOs, adding new sites or densifying the existing network primarily serves two purposes: to handle the growing traffic demands of current and new customers, and to improve the overall quality of experience (QoE) for existing subscribers.
Network coverage and quality—measured by signal strength, capacity, latency and bandwidth—directly impact the Net Promoter Score (NPS) and customer value management. In simple terms, better coverage and quality lead to more satisfied customers, higher average revenue per user (ARPU) potential and lower churn.
For challenger carriers with national roaming agreements, such as Dish Network/Boost Mobile (before the recent acquisition announcements), site builds and upgrades also have a direct cost impact. National roaming partners charge per-megabyte traffic fees, and offloading traffic to their sites drives operational expense (OpEx) savings and better unit economics.
As a result, these operators aim not only to improve coverage and QoE, but also to optimize per-MB costs by redirecting traffic from roaming to their networks. In both cases, smarter site investment drives stronger customer loyalty, higher ARPU and better unit economics—the recipe for success. Traditional RF planning tools alone cannot provide the decision‑making depth required, which is why layering AI/ML on top of existing network planning tools is essential to achieve higher ROI.
The Sand Technologies’ AI Smart CapEx telecom network planning suite supports this transformation by:
Sand Technologies’ solutions actively leverage cloud‑native products and services within the AWS ecosystem. All components are fully AWS‑native, enabling seamless scalability, security and integration with modern cloud workflows. The models ingest and process diverse data sources, as described in the architecture below. They also leverage NVIDIA NIMs and NVD‑based acceleration to optimize AI/ML workloads.
The diagrams below illustrate two deployment models:
Chart 1 – Solution design leveraging AWS Bedrock for generative and foundational model integration
Chart 2 – Solution design leveraging AWS SageMaker in combination with NVIDIA NIM for advanced model training and inference
Sand Technologies’ model integrates dozens of internal and external data sources, including:
These datasets train and retrain the AI/ML models to forecast and optimize the supply layer, modeling network capacity for both current and future scenarios, and model the demand layer, capturing existing and potential market demand for data services.
Market‑level demand modeling incorporates ML analysis of current CDRs, market share data and competitor performance, leveraging crowdsourced insights to add a competitive dimension. In parallel, ML models train to map out competitor site locations and accurately model their coverage, providing strategic intelligence for network planning.
Sand Technologies’ AI/ML models and platform employ convolutional neural networks (CNNs) for site performance modeling, inspired by the human visual system. Spatial AI models analyze site-level traffic and payload performance by stacking coverage, demand and performance signals (similar to color channels in an image) to learn spatial patterns.
Finally, the proprietary ML models generate high‑precision competitive coverage layers, calibrated using validated parameters, providing data‑driven insights to guide strategic network investments. With this data‑driven approach, telecoms can target high‑ROI investments and win in strategic markets.
The era of network planning in a vacuum is over. To win in today’s hyper-competitive landscape, MNOs must change from an internal focus to a competitive, customer-centric intelligence model.
The new paradigm for network investment is shifting from a purely internal, performance-based model to an external, competitive-driven strategy. Modern AI-powered tools are enabling operators not just to build a good network, but to build a better network than their competitors in the places that matter most.
In a crowded market, success for a mobile network operator hinges not on the size of its expenditure, but on the intelligence of its investment. Coverage Intelligence is the definitive key to unlocking that strategic and profitable advantage.
Other articles that may interest you
