Coastal Demand

The Durable Geography of AI Demand: Why Coastal Metros Will Keep Mattering

Cities formed beside water for practical reasons, and those concentrations still shape trade, talent, and enterprise. AI infrastructure should be planned around that durable geography.

· 12 min read

← All Articles

Cities do not emerge randomly.

They form where people can exchange goods, reach markets, access resources, move efficiently, and build institutions. For much of human history, that has meant settling near oceans, rivers, natural harbors, and other navigable bodies of water.

Modern economies have added airports, highways, railroads, electrical grids, fiber networks, and cloud regions. They have not erased the geographic advantages that caused major cities to form near water in the first place.

Trade still moves through ports. International communications still cross oceans. Enterprises still cluster around customers, capital, talent, universities, and commercial institutions. People continue moving toward urban areas where those opportunities accumulate.

AI infrastructure should be planned around that durable geography.

The industry is currently building much of its largest capacity where land, power, and permitting are available. Those factors matter, but they are not the complete demand map. The users, enterprises, data, and applications consuming AI will remain concentrated in metropolitan regions, many of which sit directly beside large bodies of water.

Seabase is developing modular subsea infrastructure to place meaningful AI capacity near those markets without requiring the full terrestrial footprint of another hyperscale campus.

Cities Formed Around Movement

The historical relationship between cities and water is practical.

Before railroads, highways, and aviation, waterways provided the most efficient way to move large quantities of goods. Natural harbors connected local producers with regional and international markets. Rivers connected inland resources with coastal trade.

Commercial activity attracted labor. Labor attracted housing, services, institutions, and capital. Over time, those systems reinforced one another.

The resulting cities became more than places where cargo changed hands. They developed financial markets, corporate headquarters, universities, factories, government institutions, cultural centers, and dense consumer economies.

The original geographic advantage created a lasting economic structure.

Technology has changed how cities function, but it has not removed that structure. Maritime transport still carries more than 80 percent of international trade in goods by volume, and ports remain central to economic integration and supply-chain development. (World Bank)

The largest coastal metros are no longer important only because ships arrive there. They are important because generations of infrastructure, capital, businesses, and people have accumulated around those maritime gateways.

Urban Geography Is Persistent

Cities can rise, decline, expand, or change their economic identity. Their underlying concentration is still difficult to reproduce elsewhere.

A mature metropolitan region contains interconnected systems that took decades or centuries to build:

A company can construct a building in a remote location. It cannot immediately recreate the surrounding metropolitan economy.

This creates geographic persistence.

Once people, businesses, and institutions cluster in one region, each group gains reasons to remain close to the others. Employers need talent. Workers need employment options. Enterprises need customers and service providers. Investors need access to companies. Universities need industry partners. Infrastructure operators need sufficient demand to justify expansion.

These relationships create network effects at the scale of a city.

The global economy will continue changing, but the existing concentration of people and enterprise around major metros is unlikely to disappear simply because computing can technically be placed elsewhere.

Urbanization Continues

The world is becoming more urban, not less.

The United Nations continues to project growth in cities and towns through 2050. Its 2025 urbanization work covers more than 12,000 cities with populations above 50,000 and anticipates continued expansion of the global urban system. (World Urbanization Prospects)

Coastal regions already contain a major share of that population.

Estimates vary depending on the definition of the coastal zone, but United Nations and World Bank sources place roughly 40 percent of the global population within approximately 100 kilometers of a coastline. The United Nations World Ocean Assessment estimates that more than two billion people live in the broader near-coastal zone. (United Nations)

That does not mean every future city will be coastal or that every coastal city will grow indefinitely.

It means that a substantial portion of global population and economic activity is already concentrated near the ocean, and the infrastructure serving those regions will remain strategically important.

AI demand will follow many of the same forces.

Enterprises Follow People, Talent, and Markets

Enterprise demand does not distribute itself evenly across a country.

Companies concentrate where they can reach customers, hire specialized workers, access capital, meet partners, and connect with other businesses.

Large metropolitan regions offer deeper labor markets than isolated sites. An enterprise can hire engineers, salespeople, operators, lawyers, accountants, researchers, and executives within the same economic region. Employees can change jobs without moving their families. Suppliers can serve many customers from one location.

This density creates commercial resilience.

It also creates data.

Financial transactions, media production, healthcare systems, transportation networks, logistics operations, enterprise software, consumer applications, industrial telemetry, and public services all produce information near the places where people and organizations operate.

AI infrastructure can process that data in a distant region, but distance creates a continuing networking requirement. Every interaction must travel to the compute, and every result must return.

As AI becomes more interactive and persistent, the distance between demand and infrastructure becomes more consequential. See The Latency Tax for the network geography analysis, and Beyond Chatbots: The Case for Persistent Regional AI for why persistent workloads intensify that requirement.

AI Will Become Part of Metropolitan Operations

Today, many AI applications are request based.

A user asks a question or submits a task. A remote model processes it. The user waits for the answer.

Future AI systems will increasingly remain active within the operations of cities and enterprises.

They may support:

These applications will not all require ultra-low latency. Many will benefit from regional capacity, predictable networking, dedicated infrastructure, or the ability to process large volumes of local data without repeatedly transporting it across continents.

The concentration of these applications will reflect the concentration of the economic activity beneath them.

AI infrastructure will therefore need more than a small number of enormous centralized regions. It will need a distributed layer positioned around the markets generating and consuming the workloads.

The Central Infrastructure Conflict

Major coastal metros are attractive places to serve AI demand and difficult places to build conventional data centers.

They often combine:

Moving infrastructure farther away can reduce some of those pressures.

It also moves the compute away from the market.

This creates a structural tradeoff.

Build near the users and face land, community, cooling, and development constraints.

Build far from the users and accept the network distance, backhaul requirements, and latency variability.

The AI industry has largely treated that tradeoff as unavoidable.

Seabase is developing another option. Reducing the Community Footprint of AI Infrastructure describes how modular coastal and subsea systems are intended to change the local development tradeoff.

Using the Water Beside the Market

Subsea compute allows infrastructure to move vertically and offshore without moving hundreds of miles away from the people it serves.

A Seabase deployment can be positioned near a major coastal market while keeping the primary compute equipment outside the terrestrial urban footprint.

The system still requires power, fiber, shore infrastructure, marine operations, monitoring, and maintenance. It does not eliminate the physical requirements of computing.

It changes the available site envelope.

Instead of searching only for a large parcel of surface land, the project can evaluate marine areas near ports, industrial zones, coastal power systems, and major network routes.

This creates access to a much larger physical environment beside some of the world’s most infrastructure-constrained cities. The infrastructure overview outlines the modular coastal and subsea deployment model.

Depth Is a New Site-Selection Variable

Terrestrial data centers primarily select sites horizontally.

Developers compare parcels based on land, utility capacity, fiber, water, permitting, taxes, and distance from users.

Subsea infrastructure adds another dimension: depth.

A platform can remain relatively close to shore while moving downward into more stable water conditions. Depending on the location, deeper water may provide lower temperatures, reduced surface-wave effects, and a larger surrounding heat sink.

This does not mean deeper is always better.

Greater depth can increase structural, deployment, inspection, and service requirements. Each site must balance:

The important distinction is that distance from the city and access to a useful thermal environment are not necessarily the same variable.

Infrastructure may be able to move downward without moving far away.

Coastal Fiber Is Already Strategic

International data traffic crosses oceans through subsea cable systems.

Those cables terminate near coastlines and connect into broader terrestrial networks through cable landing stations, carrier facilities, cloud on-ramps, and internet exchanges.

A cable landing station is not automatically a data-center site. It may lack sufficient power, land, heavy-lift capacity, marine operations support, or room for expansion.

The opportunity is broader.

Coastal regions often contain an ecosystem of network and industrial assets:

Seabase can connect those assets rather than requiring every function to exist at one exact point.

The result is infrastructure adjacent to the places where domestic and international networks already converge.

Ports Reinforce the Same Geography

Ports remain important beyond cargo handling.

They can provide many of the operating capabilities required for modular subsea infrastructure:

Ports also sit within or near the metropolitan regions they helped create.

That makes them potential bridges between the digital economy and the marine environment.

Not every port will be suitable for compute. Some will lack power, fiber, depth, environmental compatibility, or community support.

But the combination of port infrastructure, nearby demand, and coastal connectivity creates a repeatable site-selection pattern across many regions.

Coastal Concentration Creates Responsibility

The economic value of coastlines does not justify treating them as empty industrial space.

Coastal regions are environmentally sensitive and increasingly exposed to climate risk, erosion, storms, sea-level change, and development pressure. The same World Bank and United Nations sources that document coastal population concentration also emphasize the need to protect coastal ecosystems and build resilient infrastructure. (World Bank on coastal resilience)

Subsea compute must therefore be:

The geographic opportunity and the environmental responsibility are inseparable. Environmental Accountability for Subsea AI Infrastructure describes how Seabase approaches measurement, thresholds, and independent verification.

Seabase’s objective is not to place equipment wherever water exists. It is to identify locations where regional demand, infrastructure access, marine conditions, environmental requirements, and long-term community value align.

Not Every Workload Belongs Near a City

Centralized hyperscale infrastructure will remain essential.

Large training systems, flexible batch processing, shared storage, and tightly coupled accelerator clusters may benefit from campuses optimized primarily around power, land, and internal network scale.

Regional infrastructure serves a different part of the system.

It is most relevant where workloads benefit from:

The future AI infrastructure map will include both centralized and distributed capacity.

The mistake would be assuming that the location best suited for producing the cheapest raw computation is automatically the location best suited for delivering every AI application.

A Durable Geographic Advantage

Accelerators can be purchased by many companies.

Data-center designs can be copied. Software systems can be rebuilt. Power agreements can be negotiated.

A qualified infrastructure position near a major coastal metro is more difficult to reproduce.

It requires the convergence of:

That convergence creates a geographic moat.

The advantage is not merely that a cluster is located near water. It is that subsea infrastructure can access space beside markets where conventional high-density computing is increasingly difficult to place. Beyond Project Natick explains why commercial subsea compute must be modular to make that position serviceable over time.

AI Infrastructure Should Follow Durable Demand

The global economy will continue evolving.

New cities will grow. Existing cities will change. Some industries will move. Network technology will improve.

The deeper pattern remains durable.

People and enterprises cluster around access, opportunity, trade, infrastructure, and one another. Many of the world’s most important concentrations formed beside large bodies of water and will remain there for generations.

AI infrastructure should not be built as though those concentrations are temporary.

Seabase is developing a platform that follows the actual geography of demand.

It places modular compute near coastal metros while using the subsea environment to address the land, cooling, visual, acoustic, and development constraints that make conventional facilities difficult to build in the same locations.

The ocean has helped shape the geography of trade and cities for centuries.

It can now help shape the geography of AI infrastructure. To discuss coastal placement, regional capacity, or host-jurisdiction requirements, contact Seabase.

Next step

Discuss coastal placement and regional AI capacity

Contact Seabase