The global digital economy depends on infrastructure that most businesses and consumers never see. Beneath the world’s oceans, fiber-optic cables connect data centers, cloud regions, financial markets, telecommunications networks, and billions of internet users.

These systems carry more than 99% of intercontinental data traffic. TeleGeography counted 596 in-service submarine cable systems in November 2025, while the International Telecommunication Union’s 2026 resilience report described a network extending more than 1.5 million kilometers.

Satellites remain important for remote connectivity, navigation, broadcasting, and emergency services. However, undersea fiber provides the capacity and cost efficiency required for most data exchange between continents.

Investment is rising as cloud computing, video streaming, social media, enterprise software, artificial intelligence, and data-center expansion increase the amount of information moving between regions. The construction cycle is also being shaped by a second priority: creating more diversified routes so that economies and digital platforms depend less heavily on a limited number of cables and landing points.

TeleGeography estimates that the aggregate construction cost of submarine cables built during the past nine years averaged more than $2 billion annually. The value of announced systems expected to enter service between 2026 and 2029 exceeds $16 billion.

With the exception of several unusually active years, the industry has not experienced comparable investment levels since the technology and telecommunications boom of 2000 and 2001. A separate TeleGeography forecast projects that annual submarine cable investment could average approximately $5 billion from 2026 through 2035.

The $16 billion figure should be treated as a project pipeline rather than guaranteed completed spending. It represents estimated construction costs for announced international and domestic systems but excludes later capacity upgrades, operating costs, and ongoing maintenance. Some projects will be postponed, redesigned, or abandoned before entering service.

Even with that qualification, the pipeline indicates a substantial expansion. Investment is being supported by technology companies with large internal bandwidth requirements, telecommunications consortia seeking additional capacity, and governments financing routes considered important for economic development, security, or national resilience.

At mid-2025, international internet bandwidth stood at approximately 1,835 terabits per second, representing a 23% year-over-year increase. Global bandwidth had more than doubled since 2021, while Africa recorded the fastest regional expansion, with a 38% compound annual growth rate between 2021 and 2025.

Demand is expected to remain strong. TeleGeography forecasts that used international bandwidth will grow at a 24% compound annual rate from 2025 through 2035, meaning used bandwidth could double roughly every three years.

AI will contribute to this expansion by increasing data movement among training clusters, cloud regions, storage systems, inference platforms, and users. Large models and datasets may need to be transferred or replicated between data centers, particularly when computing capacity is distributed across countries with different power availability, operating costs, and regulatory requirements.

However, AI is only one part of the demand picture. Video, cloud applications, social networks, software distribution, data backup, enterprise communications, and expanding broadband access continue to generate substantial traffic. Attributing all future cable demand to AI would therefore overstate its role.

One of the most consequential changes is the growing role of hyperscale technology companies. Content and cloud providers accounted for approximately 75% of used international bandwidth in 2025. Their international bandwidth requirements are forecast to increase ninefold between 2025 and 2035.

Submarine cables were historically financed primarily by groups of telecommunications carriers. Each participant acquired part of the system’s capacity and shared its construction and operating costs. That consortium model remains important, particularly on routes serving several national markets.

Companies such as Google, Meta, Amazon, and Microsoft, however, now generate enough internal traffic to justify investing directly in cable infrastructure. Owning fiber pairs can provide several advantages:

Privately funded systems are not necessarily reserved entirely for their principal owners. Technology companies may sell, exchange, or grant long-term rights to unused fiber pairs, while telecommunications operators can finance branches connecting additional countries. These arrangements can distribute construction costs and extend new capacity to other network users.

Several announced projects illustrate how undersea cable investment strategies are evolving.

Meta’s Project Waterworth is particularly notable. At more than 50,000 kilometers, the planned system will be longer than Earth’s circumference. Meta says it will create three new oceanic corridors and use a 24-fiber-pair design, compared with the 8 to 16 pairs commonly used in other modern systems.

AWS’s Fastnet demonstrates a different strategy. Instead of following established transatlantic landing corridors, the cable will connect Maryland with County Cork. Its alternative route and landing points are intended to strengthen network diversity while adding substantial new capacity between the United States and Europe.

Asian and trans-Pacific corridors are attracting some of the largest planned investments. TeleGeography estimates that approximately $2 billion was invested in new intra-Asian cable systems during the preceding three years, while trans-Pacific routes attracted about $1.7 billion.

For systems scheduled between 2026 and 2029, the announced pipeline rises to more than $3.7 billion within Asia and over $3 billion across the Pacific. Google- and Meta-supported projects represent a significant portion of planned trans-Pacific construction.

Several factors explain this concentration. Asia contains some of the world’s largest internet markets, expanding cloud regions, growing AI infrastructure, major semiconductor and technology industries, and established data-center hubs such as Singapore, Tokyo, Mumbai, Chennai, and Sydney.

Investment is also creating alternative routes among North America, Asia, Australia, Africa, and South America. These connections can reduce dependence on concentrated corridors through Europe, the Mediterranean, and the Red Sea while bringing new landing points closer to emerging data-center markets.

Capacity growth alone does not explain the current construction cycle. Network operators are increasingly willing to invest in cables that provide geographic diversity, even when existing routes still have available capacity.

The ITU reports approximately 150 to 200 cable faults worldwide each year. More than 80% are attributed to human maritime activity, particularly fishing equipment and ship anchors.

Other risks include natural hazards, structural or component failures, abrasion, and intentional damage, although the cause of an incident is not always conclusively established.

Modern systems therefore incorporate route diversification, additional landing stations, reinforced nearshore cable, deeper burial in high-risk areas, and branching units that can support future connections. Operators also distribute traffic across multiple cables so that capacity can be redirected when one system fails.

This changes the financial calculation. A new cable may be justified not only by the capacity or revenue it creates but also by the losses it helps prevent. For cloud providers, financial institutions, telecommunications companies, and online platforms, network availability is directly connected to service quality, customer retention, regulatory compliance, and business continuity.

Commercial investment naturally concentrates on routes with large and predictable traffic volumes. Smaller island economies and less-developed coastal markets may struggle to support a cable through user demand alone, even when improved connectivity could create substantial economic benefits.

The ITU has recommended blended-finance structures combining private investment with government funding, development-bank capital, guarantees, or long-term public capacity purchases. A government can act as an anchor customer by committing to purchase bandwidth, converting uncertain future demand into revenue that can support project financing.

Public investment is also increasing for security and resilience reasons. By October 2025, the European Union had invested €420 million through the Connecting Europe Facility Digital across 51 Digital Global Gateway projects supporting strategic connectivity.

In February 2026, the European Commission announced a further €347 million for strategic submarine cable initiatives. The package included a €20 million call for adaptable cable-repair modules, alongside funding for new cable projects and measures intended to improve monitoring, repair capacity, and infrastructure security.

Public financing can help create routes that commercial investors might otherwise reject, but governance remains important. Open-access landing stations and nondiscriminatory capacity arrangements may be necessary if public funding is expected to encourage wider competition rather than reinforce a single network operator.

Expanding the cable network requires more than financing. Projects depend on specialized fiber, repeaters, branching units, cable-laying vessels, marine surveys, landing stations, regulatory approvals, and technicians capable of joining and testing cables at sea.

Maintenance capacity is an emerging concern. A SubOptic-commissioned study estimated that approximately $3 billion may be needed by 2040 to acquire 15 replacement cable ships and five additional vessels. Without that investment, the expanding network could face longer repair times as existing vessels age and the number of cables requiring maintenance increases.

Regulation can produce additional delays. The ITU found that environmental approvals in complex coastal areas can take 24 to 36 months or longer. Projects may also require coordination among telecommunications regulators, port authorities, environmental agencies, customs departments, national-security bodies, and local governments.

In November 2025 testimony, TeleGeography said some cables then scheduled for completion between 2025 and 2027 would probably be delayed by one to three years, while others might fail to finalize financing. This warning applied specifically to that group of projects rather than every system in the wider investment pipeline.

The record pipeline should therefore not be treated as a precise forecast of the number or value of cables that will enter service on schedule.

Rising investment creates opportunities throughout the digital infrastructure industry. Cable manufacturers, marine contractors, landing-station operators, network-equipment suppliers, data-center developers, and terrestrial fiber providers can all benefit from additional construction.

New landing locations may also attract internet exchanges, cloud infrastructure, content-delivery networks, and data centers. The economic benefit is not automatic, however. A cable that lands in a country without competitive terrestrial backhaul, reliable power, open access, or an effective regulatory framework may transport data through the market without creating a substantial local digital ecosystem.

The ownership shift also creates challenges for traditional carriers. When hyperscalers own the infrastructure supporting more of their traffic, they may purchase less capacity from conventional wholesale providers on certain routes. Carriers can respond by operating landing stations, financing regional branches, providing domestic distribution, or joining technology-led consortia.

For governments, cables have become strategic infrastructure rather than ordinary telecommunications assets. Decisions involving landing permissions, suppliers, ownership, and routing increasingly reflect national-security and geopolitical considerations alongside commercial economics.

The expansion of undersea cable investment is supported by measurable demand rather than speculation alone. International bandwidth continues to grow at a double-digit rate, content and cloud providers account for most used international capacity, and the largest platforms are committing capital to systems linking their data-center regions.

The investment cycle is also about more than moving additional data. Companies and governments are paying for greater control, lower unit costs, route diversity, economic security, and network resilience.

Not every announced cable will enter service on schedule, and substantial investment will still be required in repair vessels, permitting systems, landing infrastructure, and terrestrial networks. Nevertheless, the underlying direction is clear: the physical network supporting the global digital economy is expanding, becoming more heavily influenced by hyperscalers, and spreading across a wider range of routes and landing points.

Data and project information are current through July 2026. Pipeline figures represent announced or estimated construction costs and may differ from eventual completed spending.