Wired for Failure: The Undersea Cable Emergency That Could Sink America’s AI Aspirations

The artificial intelligence (AI) dominance the White House called for in its recently released AI Action Plan is not going to happen unless the president, Congress, and the country get serious about protecting the undersea cable system—the 600 or so inch-wide cables over which the world’s internet traffic flows. A combination of natural and human threats imperil the resilience of this critical infrastructure just as AI advances make the cables more essential than ever. Though the plan included 90 recommendations, including several massive infrastructure projects to sustain continued AI development, it also had approximately 600 garden-hose-sized holes—an omission with large political, economic, and technological ramifications. 

A recently announced proposed rule by the Federal Communications Commission (FCC) to expedite review of cable licenses, if finalized, is a step in the right direction. The licensing process is a key bottleneck in laying and retrofitting undersea cables. Private actors rely on predictable and efficient approval to move forward with costly projects, which makes the FCC’s proposed rule all the more important and timely. However, it likely will fall short of the leap in cable development that’s required to match the magnitude of the threat facing this critical infrastructure. Around 100,000 miles of new cables are necessary by 2040 to meet expected internet traffic demands. Prior efforts to streamline licensing have experienced mixed results. Under the current system “a 120-day review often takes closer to six to eight months,” according to one participant. Until the final text of the rule is made clear, it is uncertain whether such delays will become a thing of the past. Moreover, the proposed rule does not significantly address several of the most significant concerns facing the undersea cable system, such as the need for drastically more cables, improved cable quality, and far more monitoring of the ocean floor.

Nearly 100 percent of intercontinental internet traffic travels through narrow undersea cables. Diverting that traffic to space isn’t a viable alternative since information flows five times faster via cables than satellites. Put simply, the cables are the internet plumbing the world has come to rely on. Whether those pipes endure for the next decades and beyond is an open question as they deteriorate due to strong currents, sea creatures, and normal wear and tear and continue to be the targets of bad actors. The president and Congress need to take immediate action if they want to avoid their AI dominance aspirations being thwarted due to an overlooked critical infrastructure.

The Building Threat to Undersea Cables 

Three developments are making the already-brittle undersea cable system all the more susceptible to interference. First, the Trump administration has significantly lowered barriers to mining deep-sea minerals in American waters as well as the high seas. Other countries have either facilitated this unprecedented commercial activity or seem likely to follow in America’s footsteps by initiating projects of their own.

A surge in deep-sea activity—moving rocks, dropping equipment, and so on—will pose a grave threat to the garden-sized hoses that crisscross the oceans. The vast majority of cable breaks occur due to natural causes and human error. Deep-sea mining will presumably make those breaks more common. Cables are not exactly resilient to physical damage. A shark, a fishing net, an anchor, and even a rock moving in the wrong way at the wrong time can sever a cable. Mining promises to introduce a heightened degree of uncoordinated activity on the sea floor, especially considering that the National Oceanic and Atmospheric Administration is on its way to streamlining the mining permitting process and companies have shown a willingness to ignore the guidance of the International Seabed Authority, an autonomous international organization created by the 1982 United Nations Convention on the Law of the Sea.

Second, the explosion in AI use is at once making access to high-speed internet more important and more scarce as increased traffic clogs technical systems suited to a different era. The battle among private and public stakeholders to build out the physical infrastructure associated with AI dominance may soon move under the seas. Private AI labs, such as Meta, are already rushing to lay new cables to keep pace with current and forecast demand. Who builds which cables and for what countries is a hotly contested and highly consequential matter. Adversaries have plenty of reason to attempt to delay or undermine massive cable initiatives such as Meta’s Project Waterworth, which will span five continents and account for approximately 31,000 miles in cable. Setbacks to such resource intensive endeavors may ripple across a nation’s entire tech stack due to diminished high-speed internet access. What’s more, as the undersea cable system itself expands, the institutions and actors tasked with its maintenance and repair will become even further stretched thin. As it stands, no entity or collection of entities meaningfully monitors all 870,000 miles of undersea cables.

Third, there appears to be no end in sight to geopolitical tensions that adversaries have cited as an excuse to disrupt the undersea cable system. In the past year or so, six cable breaks have been attributed to China and Russia. The Houthis may have cut four cables in 2024. Advances in undersea drones and related naval technologies will allow adversaries to commit such acts at greater depths with greater frequency and with even lower odds of attribution. Suddenly the 17 or so cables connecting North America to Europe seems like an awfully low number.

Proposals Reflective of the Value of the Undersea Cable Systems

To be fair to a number of scholars, such as David Opderbeck, and politicians, including former U.K. Prime Minister Rishi Sunak and Sens. Chris Murphy (D-Conn.) and Todd Young (R-Ind.), who have proposed policy ideas, several important stakeholders have recognized and attempted to address the fragility of the undersea cable system. Their solutions, however, have often been too reliant on international law frameworks with low odds of successful enforcement or too meager to result in a substantially more resilient undersea cable system. One of Sunak’s main proposals—an international treaty—is likely a nonstarter in today’s geopolitical environment. What’s more, Congress is currently weighing legislation that would build two new submarine cable-laying and repair ships. That’s akin to the New York City Council touting two new ambulances. It’s just not enough to make a real difference.

Cable repair work poses unique challenges. Bad weather, a shortage of talented workers, and a dearth of boats all mean that in the event of several cables breaking it will take weeks, if not months, to get them back on line. That was the case a few years back when it took six months to repair four cables off the coast of Vietnam. Similarly, in 2006, when an earthquake broke six of the seven cables near the Luzon Strait, it took 11 ships 49 days to bring the cables back on line.

Solutions that have worked in other contexts likewise seem ill-suited to the nature and scale of the crisis facing the undersea cable system. New Zealand, for example, has implemented cable protection zones that limit naval traffic near areas with cable clusters. The government has committed significant resources to enforcing those zones. So far, these zones seem to have worked. But it’s important to note that these zones likely benefit from having a drastically smaller number of cables (just four) in a narrower geographic area than a nation like the United States.

It is time for far more drastic action grounded in two core principles—redundancy and resiliency—and three proposals: 10 new cable repair ships, 100 autonomous undersea drones, and 100,000 miles of new or retrofitted undersea cables—or the “10-100-100,000 initiative.”

On redundancy, the president should apply his “America First” approach to governance by seeking to become the first president to lay 100,000 miles of undersea cables. It’s a big number. He likes setting big goals. Why not aim for the sky? (Or the depths?) Whereas three state-owned Chinese firms are actively extending that country’s ambitions via new cables, the U.S. government—specifically, the Navy—owns just 40,000 miles of cable. The goal would be to lay many more cables between the U.S. and key overseas markets as well as to replace or retrofit cables at risk of diminished capabilities due to age. Ideally, the government would partner with existing cable owners to do so given their expertise and existing infrastructure. However, it may also want to independently build some of those cables given the importance of not relying solely on private entities for the maintenance of this critical infrastructure. As the number of cables grows, the net harm of an attack on any one cable diminishes; traffic can be fairly easily rerouted. This bold endeavor also amounts to good policymaking. Many cables laid near the early days of the internet are reaching the end of their typical life cycle of approximately 20 years. The combined need for a more redundant system and one that is suited to the AI age makes this effort all the more important.

Extensive executive power could aid the president in realizing this aquatic moonshot (dare I say, “sea shot”). In line with several recommendations in the AI Action Plan, the president can lower regulatory hurdles to laying cables and establishing cable landing points on shore. A litany of federal agencies, including the National Oceanic and Atmospheric Administration and the Federal Communications Commission, play a role in determining which individuals can do what in and around the ocean. The cumulative result can bring undersea cable development to a halt. Washington state, despite its proximity to Asia, has not been the site of a new cable connection point in more than two decades; local, state, and federal hurdles may be to blame. The slow and, in some cases, seemingly arbitrary denial of cable licenses by Team Telecom—an advisory body to the FCC made up of the Departments of Justice, Defense, and Homeland Security—deserves particular scrutiny. Team Telecom’s recommendations to the FCC as to whether to approve or deny a license are often determinative, yet commonly turn on ad hoc considerations. The resulting uncertainty has unsurprisingly drawn the ire of cable owners. Proposed FCC rules to accelerate this process may assuage some of these concerns but may stop short of addressing some of the aforementioned state and local barriers.

What’s more, the president can leverage the Defense Production Act (DPA) to ease the burden of securing the materials necessary to lay that many cables. The current supply chain is highly fragmented and involves several scarce, expensive inputs. Cables are the product of parts assembled by dozens, if not hundreds, of companies. The DPA is a tool tailored to remedying those sorts of barriers. Pursuant to its expansive provisions, the president may mandate that federal production and supply contracts receive priority and direct private actors to expand production of certain goods. DPA authorities are contingent on the president acting with an eye toward national defense. That should not pose a problem here given that both commercial and military communications rely on a durable undersea cable system.

On resiliency, the construction and deployment of 10 additional cable repair ships and 100 autonomous undersea drones capable of monitoring adversary ships and drones as well as assessing the durability of cables will go a long way toward helping Americans get back on their feet by getting back online in the event of a sizable attack on the undersea cable system. The value of additional cable repair ships has already been explored and is fairly obvious. Autonomous undersea drones, however, would constitute a novel but overdue investment. New sea drones, such as those created by Germany-based Helsing, can remain underwater for up to four months and clandestinely surveil enemy ships.

Thankfully, the Navy is already soliciting input from the private sector on how to develop and deploy drones with similar capabilities as soon as possible. This effort should include an expectation that the drones be capable of both detecting threats to undersea cables and, critically, pinpointing where a cable has been severed. By championing this nascent effort through the announcement of the 10-100-100,000 initiative, President Trump may be able to scale up the level of congressional support for its continuation as well as to attract more private-sector interest.

***

The undersea cable crisis represents more than a technical challenge—it embodies the tension between America’s digital aspirations and the physical realities that underpin them. Just as the transcontinental railroad required bold federal action to connect a divided nation, today’s digital infrastructure demands similar vision and commitment. The fragility of our current system reflects a broader pattern in American governance: the tendency to build magnificent superstructures while neglecting the foundations that sustain them.

The 10-100-100,000 initiative offers more than redundancy and resilience—it presents an opportunity to reclaim American leadership in the infrastructure that will define the next century of global competition. History suggests that nations that control the arteries of communication wield disproportionate influence over the flow of information, commerce, and, ultimately, power itself. The British Empire’s telegraph cables, America’s satellite networks, and now China’s Digital Silk Road initiative all demonstrate this enduring truth.

Yet the path forward requires acknowledging an uncomfortable reality: America’s adversaries have recognized the strategic value of undersea cables while the U.S. government has treated them as utilities rather than a key feature of our national defense. The garden-hose comparison is apt not merely for its physical dimensions, but for how policymakers have conceptualized these vital arteries—as mundane infrastructure rather than the nervous system of American digital dominance.

The president’s opportunity is clear. By framing undersea cable expansion as both economic necessity and national security imperative, he can marshal the same political energy that built interstate highways and put Americans on the moon. The ocean floor awaits America’s next great infrastructure project. The question is whether the United States will seize this moment or allow others to write the rules of our digital future from the depths below.