For years, improving the Internet was relatively easy to explain: more speed, more capacity, better Wi-Fi, new modems, fiber, and new television platforms.
When I started working in telecommunications, going from 10 to 100 Mbps really changed a family’s experience. I later worked on projects for new modems, set-top boxes, video platforms, and network evolution where the technology improvement was also fairly easy to turn into customer value.
Today we keep building better networks. But there is a problem. For most households, going from 500 Mbps to 1 Gbps probably does not change their life much. Going afterward to 2 Gbps, even less.
That does not mean telcos can stop investing. Quite the opposite. Traffic keeps growing, and networks need more capacity, fiber, spectrum, resilience, and technology. The problem is something else: it is getting harder and harder to charge much more simply for delivering more speed.
So what do you sell next? I think that is one of the most interesting questions the industry faces today.
Maybe we are looking at the wrong asset
A telco does not only have an Internet connection. It has fiber, backbone, spectrum, antennas, data centers, central offices, distributed infrastructure, equipment inside millions of homes, and interconnections with practically every major content and cloud provider.
It is also in a fairly particular position: it sits between the user and a large share of the digital services they use. It can authenticate devices, knows the conditions of its network, can bring content closer, connect enterprises directly to the cloud, and in mobile offer network information to fight fraud. With technologies like slicing, it can even guarantee certain connectivity characteristics.
So maybe the question is not how to sell another gigabit. Maybe it is: what else can a telco sell using the infrastructure it has already built? And some answers are already appearing.
Selling enterprises something more than connectivity
This is probably the least futuristic opportunity, and one of the most real: cloud, cybersecurity, IoT, private networks, data centers, managed services, and direct connections to AWS, Azure, or Google Cloud.
Vodafone Business, for example, already gets a meaningful share of its revenue from digital services. Something similar is happening in Latin America. Claro offers cloud infrastructure and direct enterprise connections. Tigo has been expanding its B2B business for years. And Liberty Latin America has a particularly interesting business: Liberty Networks.
In 2025, Liberty Networks generated approximately US$471 million in revenue selling enterprise and wholesale services over its terrestrial and submarine infrastructure.
That shows something important. A telco does not need to invent the next Netflix to create a new business. It can find new ways to monetize the network it already has.
It can also sell the network to applications
This is where Open Gateway appears. Claro, Entel, and Telefónica are already working with these capabilities in Chile. One of them is SIM Swap.
Imagine someone tries to transfer several million pesos from a bank account. Before authorizing the transaction, the bank could query the mobile network directly to check whether the SIM associated with that number was recently changed. That is worth money because it can prevent fraud. There are also APIs to verify numbers, devices, or location.
It does not sound as spectacular as talking about 6G or smart cities, but that is probably where the lesson is. The first network capabilities finding customers are the ones that solve concrete problems. Then come more ambitious ideas like Quality on Demand or network slicing. And here Chile offers an easy example to understand, though the problem is not only Chilean.
What is a connection worth when an earthquake hits?
Under normal conditions, a video taking a few extra seconds to load probably does not matter much. During an emergency, the situation changes completely.
Chile has earthquakes, wildfires, floods, and other situations where thousands of people try to use the networks at the same time—exactly when firefighters, ambulances, hospitals, Carabineros, or Senapred also need them. Something similar happens across much of Latin America: hurricanes in the Caribbean and Central America, floods in Brazil, storms, wildfires, or critical infrastructure crises in different countries. In all of those cases the network saturates right when it is needed most.
Chile has already made progress on regulation and special infrastructure for emergency communications. But technologies like network slicing open an even more interesting possibility: creating logical parts of a 5G network with specific characteristics for certain services. That would make it possible, for example, to guarantee certain connectivity conditions for critical services even when the network is under heavy pressure.
And it does not have to stay inside one country. A regional operator—or several working together—could offer those capabilities in Panama, Brazil, Chile, or wherever they operate, and even consider multi-country scenarios when the emergency or the critical operation crosses borders.
Who pays? The government? Hospitals? Critical infrastructure companies? Does it simply become a regulatory obligation? There are still questions. But at least here slicing stops being a technology demo and starts solving a problem that anyone who has lived through an earthquake in Chile—or a similar emergency elsewhere in the region—can understand.
From CDN to bringing processing closer too
This part interests me especially because for years I have worked around video platforms and CDN. CDNs start from a fairly simple idea: if millions of users want to watch the same content, it does not make sense to keep fetching it from thousands of kilometers away. You bring it closer.
Netflix, Google, and other providers have been placing caches inside or near operator networks for years. Last year at IBC I even saw a solution where users’ own devices could store small video segments and share them with other peers. The idea stuck with me.
Because for years we have brought content closer to the user. With AI we may also start bringing processing closer. Not all AI needs to be at the edge: training large models will still require enormous centralized data centers. But some applications can benefit from running inference closer to the user for latency, privacy, data sovereignty, or simply efficiency.
And here operators have something very hard to build from scratch: distributed locations connected directly to their networks. Deutsche Telekom already launched a sovereign AI infrastructure in Germany with around 10,000 NVIDIA GPUs. SK Telecom turned AI Data Centers into a business line that generated approximately US$350 million in 2025. Telefónica already has 17 edge nodes in Spain. Liberty Global is developing distributed data center infrastructure through AtlasEdge and exploring with Google Cloud how to use part of that capacity.
Maybe a telco does not need to become AWS. Maybe it can sell infrastructure to AWS.
And that is where Humboldt appears
I recently wrote about Humboldt, the submarine cable that will connect Chile directly with Asia-Pacific. But there is another way to look at that project. The value does not necessarily end when capacity on the cable is sold. A new international route can attract data centers, interconnection, cloud, AI infrastructure, and companies that need to connect Latin America with Asia.
That creates an opportunity around the infrastructure. And also a warning. Google does not only need telco networks. It is investing directly in submarine cables, cloud regions, data centers, and AI infrastructure. Hyperscalers can be customers, also partners, and increasingly competitors in some layers of infrastructure.
Even television can become another business
There is another possibility I know closely from my experience in video: advertising. I am not talking only about the classic TV commercial. Modern platforms enable addressable advertising: different households can receive different ads even while watching the same content. VodafoneZiggo already uses this model in the Netherlands.
And for a telco with linear TV, replay, VOD, apps, and millions of customers, that opens another monetization opportunity. The network stops generating revenue only by transporting the content. It can also participate in the value created around the audience that consumes it—always within the limits of privacy and regulation.
It is another example of the same idea: look at infrastructure that already exists and ask what else can I do with this.
The risk is repeating the history of the Internet
Here I see the greatest danger. Telcos have been chasing efficiency for years. I have seen it directly: teams that get centralized, reduced, outsourced, and sometimes later brought back in-house depending on strategy. Today I work precisely from that model, delivering technology services for large operators.
AI will probably accelerate that search for efficiency even further. But cutting costs does not answer the growth question. And outsourcing too much also has a risk. If the telco buys the hardware from a vendor, uses a hyperscaler’s cloud, consumes its AI models, outsources the integration, and lets another platform control the developers and the applications, what remains under its control? Mainly the infrastructure.
The history of the Internet should serve as a warning. Telcos built extraordinary networks. On top of them grew Google, Netflix, Amazon, Meta, and some of the most valuable companies in the world. Operators built a fundamental part of the infrastructure. Others captured a large share of the new value created on top. With AI, edge, and cloud, it could happen again.
The opportunity for Latin America
That is why I think this discussion is especially interesting for Chile and Latin America. We have operators with national networks, submarine cables, data centers, spectrum, millions of customers, and enterprise relationships. We have mining, energy, ports, agriculture, banking, and governments with problems where connectivity, security, edge, and AI can have fairly concrete applications. And now we have hyperscalers increasing their physical presence in the region.
I do not think the next telco business is a single technology. It will not simply be slicing, or edge, or APIs, or AI. It will probably be a combination of many things: B2B services, infrastructure for hyperscalers, data centers, AI compute, security, identity, APIs, private networks, advertising, and network capabilities that today we still deliver as an invisible part of a connection.
For decades, telcos mainly sold access to the network. Now they also have the opportunity to sell what that network knows, where it is, and what it is capable of doing.
Because after decades of building one of the hardest-to-replicate technology infrastructures in the world, the next battle may not be making it even faster. It will be discovering how much more value it can generate before others do.
Related reading:
- Google Is Connecting Chile to Asia
- The Silent Crisis in Telecommunications
- AI Data Centers in Latin America
✍️ Claudio from ViaMind
“Dare to imagine, create and transform.”
FEATURED POSTS
ABOUT ME
ViaMind Ecosystem
- ViaMind — The brand hub for ideas, technology and strategy.
- NeuraPRO — Practical software for managing real operations.