Google is helping build a submarine cable of more than 14,000 kilometers between Chile and Australia.
The project is called Humboldt, and when it goes into operation it will create the first direct submarine connection between South America and Asia-Pacific.
The project is interesting because of its size, but there is a much more important question behind it: why is Google building infrastructure we historically associated with telecommunications companies?
The answer helps explain a transformation that is quietly happening on the Internet.
The Internet needs new routes
Over the last two decades, the amount of information we move across the Internet has grown enormously. First came video and streaming. Then the cloud moved a large share of company systems into massive data centers. Now artificial intelligence is adding another layer of demand.
Networks have to grow with that traffic.
We can imagine a highway that receives more vehicles every year. We can widen its lanes and make it more efficient, but eventually we will need new roads.
Something similar happens with the Internet. If every year we need to move more information between Latin America, the United States, Europe, and Asia, improving the existing routes is not enough. We also need new connections.
And having more paths delivers another benefit: alternatives. If one route has problems, traffic can be diverted through another.
Humboldt does exactly that at a continental scale: it adds capacity and opens a new direct path between South America and Asia-Pacific.
But there is another reason these routes are becoming especially important.
Hyperscalers need more and more infrastructure
Google, Amazon, and Microsoft belong to a category we hear more and more in technology: the hyperscalers.
They are companies that operate technology infrastructure at a gigantic scale. Their data centers run everything from enterprise systems and streaming platforms to cloud services and artificial intelligence models used by millions of people.
And the more they grow, the more they need to control how information moves between those data centers and their users.
That is why Google participates directly in submarine cables. Microsoft and Amazon invest billions in data centers and networks. Meta also takes part in large international submarine systems.
It does not mean Google wants to become a telco and sell us home Internet. It wants something much more strategic for its business: capacity, routes, and greater control over the infrastructure its services depend on.
Artificial intelligence is accelerating that need even further.
Chips get most of the attention, but thousands of GPUs working together also need to communicate with each other and connect to other data centers and millions of users.
That is why even companies like Cisco are seeing billions of dollars in new orders related to AI infrastructure.
The artificial intelligence boom is not only generating a race for better chips. Everything we are building also has to be connected.
Telcos do not want to be limited to carrying traffic
This is where an interesting situation appears.
We have never used telecommunications networks so much. That does not mean, however, that telcos are the ones capturing most of the economic growth that happens on top of them.
A household can pay practically the same for its connection while consuming more Netflix, YouTube, TikTok, cloud, video games, and artificial intelligence every year.
The operator has to expand and modernize the network to support that growth, but a large share of the new value ends up in the platforms that run on top of it.
That is why several operators are trying to move into other parts of the business.
Liberty Latin America is a good example.
In addition to its traditional Internet, television, and phone operations, Liberty Networks operates extensive terrestrial and submarine fiber infrastructure in the region.
That allows it to participate in another market.
A bank needs to connect its data centers. A local operator may need international capacity. A multinational needs to connect operations across several countries. Google, AWS, or Microsoft need to move enormous amounts of information between different points in their networks.
All of them need connectivity. And someone can charge for providing it.
For Liberty, having regional fiber and participating in submarine cables means that part of all that data growth can turn directly into business, instead of simply passing through its networks.
Millicom, through Tigo Business, is following another version of the same strategy. In addition to connectivity, it has developed data centers and cloud, colocation, security, and business continuity services for enterprises.
The logic is similar again: if a company already has fiber, enterprise customers, regional networks, and technical capacity, why limit itself to selling connectivity?
America Movil and Claro have another advantage: scale. They operate mobile and fixed networks across much of Latin America and continue to invest in fiber, network modernization, and 5G. That regional presence allows them to offer connectivity and technology solutions to companies that operate in several countries.
These are different strategies, but they point to the same problem: telcos do not want to become simply the pipe through which the growth of Google, Amazon, Microsoft, or artificial intelligence flows.
They want to capture a larger share of that value.
And Chile wants to become one of the places where everything connects
This is where the story returns to Humboldt.
Chile is not participating only as the place where the cable will land. The Chilean State and Google created Humboldt Connect, a company meant to commercialize capacity from this new infrastructure.
The ambition is larger: to position the country as a relevant digital node for Latin America.
If Chile has good connections to North America, other South American countries, and now Asia-Pacific, plus data centers, cloud infrastructure, and available energy, it becomes more attractive to install new technology capacity in the country.
A cable by itself generates some value. But if data centers, cloud providers, new networks, specialists, and technology companies appear around it, an ecosystem starts to form.
It is similar to what happens around a good port. The business is not only in the ships that arrive. Transport, storage, logistics, services, and jobs appear around that infrastructure.
Chile could try something similar with data.
For an ordinary person, Humboldt will probably not mean opening Instagram and discovering that it now works twice as fast. The impact can be much more structural: investment, specialized jobs, better international connections, greater resilience, and new digital industries operating from Chile.
The opportunity also has a cost
There is something easy to forget when we talk about cloud and artificial intelligence: all of this is still extremely physical.
Data centers need land, cooling, and large amounts of electricity. Chile has an important advantage because of its solar and wind capacity, but generating energy does not necessarily mean being able to take it to the place where it is needed. Transmission and electrical capacity can also become bottlenecks.
There is also an environmental impact. Google already faced questions in Chile over the water consumption associated with the original design of its data center in Cerrillos and ended up rethinking the project.
The conclusion should not be that data centers are bad. Nor that any technology investment is automatically good. What matters is what energy they use, what resources they consume, and how much value they actually leave in the country.
And if Chile becomes more important for regional connectivity, the responsibility to protect that infrastructure also increases. A cable can be damaged, a data center can lose power, and a network can suffer an attack. Having alternative routes and backup systems will be as important as continuing to build capacity.
There is also an inevitable geopolitical dimension. Cables, cloud, semiconductors, and telecommunications are already considered strategic assets by the major powers. For Chile and Latin America, the discussion will increasingly be about how to take advantage of those investments without ending up excessively dependent on a single provider, country, or route.
The value is not only in the cable
For years we mainly looked at the companies that built things on top of the Internet. Google created search, Netflix changed television, Amazon transformed the cloud, and now several companies compete to lead artificial intelligence.
But underneath all of that, a lot more Internet still has to be built.
Google wants better control over the infrastructure its services need. Telcos want to capture a larger share of the value that flows through their networks. And Chile wants part of that infrastructure to connect and operate from the country.
That is, for me, the most interesting opportunity for Latin America.
We do not need to think that having a new submarine cable will automatically turn us into a technology power. Infrastructure by itself guarantees nothing.
The real challenge is something else: what we are able to build around it.
If Humboldt helps attract data centers, companies, talent, new connections, technology services, and knowledge, the value for Chile will be much greater than the data that flows through the cable.
For a long time Latin America has mainly been a consumer of the large digital platforms built elsewhere in the world.
This new race opens a different possibility: also participating in the infrastructure on which the next stage of the Internet will run.
And perhaps that is the most interesting part of Humboldt.
Not only where the cable arrives.
But what we can build ourselves when it does.
Related reading:
- The Invisible Work Behind the Internet
- The Silent Crisis in Telecommunications
- Millicom Landed in Chile
✍️ 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.