Mobile Operators Create Cost Savings in Many Areas

Operating and capital expenses are the two major “controllable” costs for any mobile operator, which is why most of the impact of cost control efforts has to come in those areas. There is only so much a mobile operator can do about spectrum licensing fees, taxes or other costs such as content rights or roaming charges. 

Cost Contributor	Estimated Percentage of Total Costs	Key Components
Network Capex	15-25%	Spectrum: License fees, network deployment costs for acquired spectrum. Infrastructure: Cell towers, base stations, fiber optic cables.
Operating Expenses (Opex)	45-60%	Network Opex: Maintenance, energy costs, site leases. Commercial Opex: Marketing, customer service, billing, IT systems.
Content & Licensing Costs	10-15%	Roaming agreements, streaming content rights, mobile VAS (value-added services).
Regulatory Fees & Taxes	5-10%	Spectrum fees, government licenses, corporate taxes.

Nearly every initiative aimed at cost containment, ranging from use of open source to virtualization and software-defined networks, Wi-Fi offload, infra sharing and energy efficiency are aimed at limiting or controlling opex and capex. 

But savings come in several areas. 

Cost Area	Potential Cost Reduction Range	Key Drivers
Network Optimization:	10-20%	Automation, software-defined networking (SDN), AI-driven optimization, cell site modernization.
Operational Efficiency:	5-15%	Streamlining processes, digitization, outsourcing non-core functions, shared procurement.
Spectrum Sharing & Collaboration:	5-10%	Pooling spectrum with other providers, reducing individual license costs, infrastructure sharing.
Personnel Costs:	3-8%	Reskilling workforce, automation, optimizing staffing levels, outsourcing specific tasks.
Content & Network Management:	2-5%	Targeted content delivery, caching mechanisms, optimizing network traffic, data plan differentiation.
Energy Efficiency:	1-3%	Investing in energy-efficient equipment, utilizing renewable energy sources.

What is a Reasonable Growth Rate Expectation for Connectivity Services?

As a rule, utility-type industries and firms (whether formally regulated as public utilities or not) tend to grow at rates just a bit over the underlying inflation rate. So the underlying inflation rate really does matter. 

That might also tend to temper excessive optimism about “telcos becoming techcos” and thereby boosting their growth rates. So long as most of the revenue is earned from supplying the connectivity function, growth rates are unlikely to shift higher. 

But few industries tend to grow at six-percent to nine-percent annual rates, if that is the hope in the connectivity industry. 

Industry	2024 Growth	2025-2028 CAGR
Telecom Services	2.8%	3.2%
Software Development & IT Services	5.8%	6.2%
Applications (Mobile & Web)	8.1%	9.3%
Information Technology (Overall)	3.9%	4.5%
Healthcare	4.2%	4.7%
Retail (E-commerce & Traditional)	4.6%	5.1%
Education (Online & Traditional)	5.5%	6.0%
Renewable Energy & Cleantech	8.7%	10.2%
Financial Services (Fintech & Traditional)	4.9%	5.4%
Manufacturing & Construction	2.1%	2.7%
Transportation & Logistics	3.4%	4.0%

So retail telecom services globally are expected to grow at a compound average growth rate of perhaps 4.9 percent between 2023 and 2028, exceeding inflation rates in most countries (assuming inflation rates decline to around two percent). 

The actual range of forecasts runs from about two percent up to six percent in some cases, with global expectations in the two percent to three percent range, a blend of higher rates in Asia, probably North America and the Middle East and lower rates elsewhere. 

The global electricity market is expected to grow at a CAGR of 3.3 percent between 2022 and 2027, again exceeding inflation. 

The global water utilities market is expected to grow at a CAGR of 3.9 percent between 2023 and 2032, exceeding inflation. 

Likewise, the global waste management market is expected to grow at a CAGR of 4.2 percent between 2022 and 2027, exceeding inflation.

Region	Analyst Firm/Source	Predicted Growth Rate 2024	Predicted Growth Rate 2025-2028 (CAGR)
Global	GSMA Intelligence	2.8%	3.2%
North America	IDC	2.3%	2.9%
Europe	Analysys Mason	1.7%	2.5%
Asia Pacific	Deloitte	4.5%	4.8%
Latin America	Ovum	3.2%	3.8%
Middle East & Africa	Ericsson	4.8%	5.3%
China	China Academy of Telecommunications Research	3.1%	3.5%
India	India Brand Equity Foundation	6.3%	7.1%
Japan	Telecommunications Carriers Association	1.2%	1.5%
Germany	Bitkom	1.9%	2.3%
Brazil	Teleco Síntese	2.7%	3.3%
Nigeria	GSM Association	6.7%	7.5%

AI: Crawl before You Walk; Walk Before You Run

It might be obvious that we are in the “picks and shovels” moment for artificial intelligence. That phrase refers to companies that provide the essential infrastructure or services needed for an industry to thrive.

The phrase originated in the 1850s “gold rush” era, when the people selling picks and shovels to the miners often made more money than the miners.

So “picks and shovels” applies to semiconductors, payment processors and data centers and connectivity networks, for example, that support computing, apps, e-commerce and transactions. 

In the data center business, that means the ability to support the building of generative AI models, notes Raul Marynek, Databank CEO. “Since the second quarter we have seen a tsunami of demand for machine learning and model training,” he said. 

“Even if AI development stopped today, you’d see a 10-fold improvement in coding,” said Armand Musey, Summit Ridge CEO. But professional services and content development will follow, he also noted. 

And even within the infrastructure space, the same customers and partners for cloud, processors and applications are the lead customers for data center AI support as well. “AI is like multi-cloud and hybrid cloud,” said Jim Poole, Equinix VP.                             

To be sure, there is a mad rush to incorporate AI into core business processes and end user experiences, but often without any complete understanding of where such investments will produce meaningful business outcomes, as would be expected whenever a hot new technology is adopted, early on. 

We tend to apply the new technology in support of existing business processes. Only later do we discover how to use the technology to change business processes. And that probably is necessary. We will need to “crawl before we walk and walk before we run.” 

But broadly speaking, the next wave of AI efforts will likely be seen in “software” products including customer relationship management, cybersecurity, business analytics and intelligence, for example. Virtually every other general purpose technology has done so, spreading from core enabling infra to a few key applications and then finding use in many other areas over time. 

So the steam engine originally found use as a method of pumping water out of mines, but later became the means of propulsion for ships and trains. Electrical grids were built to supply home lighting, but later enabled all sorts of powered appliances. 

Computers originally were used for business computations and information processing, but later evolved as media and content consumption appliances; “phones,” “cash registers” and all sorts of other devices. 

Industry	Software Firms	Application Suppliers	Product Suppliers
Software Development	* Machine learning (ML) platforms: Provide tools and frameworks for developers to build and deploy AI models. (e.g., TensorFlow, PyTorch)	* AI-powered development tools: Automate repetitive tasks, generate code, and suggest improvements. (e.g., GitHub Copilot, Tabnine)	* AI-powered testing and debugging tools: Identify and fix software bugs more efficiently. (e.g., DeepCode, Pachyderm)
Image & Video Processing	* Facial recognition and object detection: Used for security, marketing, and customer service applications. (e.g., Amazon Rekognition, Microsoft Azure Cognitive Services)	* Augmented reality (AR) and virtual reality (VR) applications: Enhance user experiences with realistic virtual environments and overlays. (e.g., Google Lens, IKEA Place)	* Self-driving cars and drones: Utilize computer vision and deep learning for autonomous navigation. (e.g., Tesla Autopilot, DJI Matrice 300)
Healthcare	* Medical diagnosis and treatment: Analyze patient data to identify diseases, recommend treatment plans, and predict outcomes. (e.g., IBM Watson Health, Babylon Health)	* Personalized medicine: Tailor treatment plans to individual patients based on their genetic makeup and other factors. (e.g., Foundation Medicine, Invitae)	* Robotic surgery: Assist surgeons with delicate procedures and improve accuracy. (e.g., Intuitive Surgical da Vinci, Medtronic Hugo)
Finance & Banking	* Fraud detection and risk management: Identify suspicious activity and prevent financial losses. (e.g., Featurespace, NICE Actimize)	* Personalized financial advice and robo-advisors: Provide tailored investment recommendations based on individual goals and risk tolerance. (e.g., Wealthfront, Betterment)	* Automated trading systems: Use AI to make investment decisions and execute trades. (e.g., Renaissance Technologies, Two Sigma)
Retail & E-commerce	* Product recommendations and personalization: Suggest products to customers based on their browsing history and purchase behavior. (e.g., Amazon Recommendations, Netflix)	* Chatbots and virtual assistants: Provide customer service and answer questions. (e.g., Sephora Virtual Assistant, Nike Bot)	* Dynamic pricing and inventory management: Optimize prices and inventory levels based on real-time demand. (e.g., Amazon Dynamic Pricing, Walmart Inventory Optimization)
customer Service	* AI-powered chatbots: Companies like IBM Watson Assistant and Microsoft Azure Bot Service provide AI-powered chatbots for 24/7 customer support and automated query resolution.	* Sentiment analysis and personalization: AI tools analyze customer feedback and social media data to understand customer sentiment and personalize marketing campaigns.	* AI-driven product recommendations: Online retailers like Amazon and Netflix use AI to recommend products and content based on individual user preferences and purchase history.
manufacturing	* Predictive maintenance: AI models analyze sensor data from machinery to predict potential failures and schedule preventive maintenance, minimizing downtime and costs.	* Supply chain optimization: AI algorithms optimize inventory levels, routing, and logistics to improve supply chain efficiency and reduce costs.	* Quality control and defect detection: AI-powered vision systems automate product inspection and identify defects with high accuracy, improving product quality.