A tethered ROV slowly descends toward a pipeline that was last thoroughly inspected maybe two years ago while a crewed vessel idles off the coast for about $100,000 per day. The crew is waiting. The information gradually becomes available. And an analyst somewhere in an Aberdeen or Houston office completes the task, files the report, and moves on. This is the subsea industry in 2025: a massive, vital, almost ridiculously outdated business that everyone wants a piece of.
The figures being discussed are astounding. Undersea infrastructure contributes more than $3 trillion to the world economy. Fiber cables installed on the ocean floor carry about 95% of all internet traffic worldwide. One of the biggest categories of capital expenditures in the energy industry is offshore energy, which includes both wind and fossil fuels. However, in many ways, the technology stack handling all of this is still more akin to the 1980s than it is to any realistic future vision. The maintenance of the modern global economy using tools and procedures that wouldn’t be out of place in a documentary about the Cold War creates an odd cognitive dissonance.
The appetite has recently changed. Inspired by the AI boom and eager to explore the next frontier, venture capital has begun circling the globe. Budgets for defense are growing. Offshore wind is being forced into deeper, more difficult waters by energy transition pressures. The argument for a subsea revolution seems almost obvious on paper. It’s much more difficult in practice, and the discrepancy between the pitch and reality needs to be examined more closely than it has been.
Oceaneering, Kongsberg, Teledyne, Saipem, and a few other incumbents did not establish their dominance through constant innovation. They used capital intensity to build it. Sensors that require years to qualify, shipyards, foundries, and exclusive testing facilities. Selling a limited number of very expensive systems to a captive base of defense ministries and oil majors who had nowhere else to turn was their business strategy, which was essentially intended to stifle competition. There was little pressure to change, costs remained high, and margins were safeguarded. It was effective until the logic of that model began to resemble a trap rather than a moat.
The issue is easily identified by anyone attempting to construct a new underwater vehicle today using high-quality parts. Expenses soar to $2–$3 million per unit. Supply chains are managed by sluggish contractors under monopoly control. Lead times can take years. Underwater GPS and radio frequency communication are nonexistent. At scale, acoustic positioning is costly and unreliable. Each deep-sea pressure-resistant battery costs hundreds of thousands of dollars. Over weeks, biofouling silently deteriorates sensors. The ocean may be more difficult than space, as some engineers like to say, and space has already destroyed a number of well-funded businesses.
Subsea narratives are beginning to reflect the circular capital logic that is currently driving the larger tech industry. Investors appear to think that new battery chemistry, inexpensive sensors, and AI-enabled autonomy will drive down these costs in the same way that smartphones drove down computing costs. The analogy is appealing. Additionally, it is primarily theoretical, at least thus far. The physical surroundings are unchanged. Software iteration cycles don’t matter to the ocean. Furthermore, deploying autonomous systems in international waters still presents significant operational, regulatory, and certification challenges that cannot be overcome by a GitHub push.
Vendor financing loops, circular equity arrangements, debt pledged against future contracts, and other financial engineering dynamics that are now apparent in the AI data center buildout seem to be starting to emerge at the periphery of the subsea investment thesis as well. Capital is flowing quickly. The pace of due diligence is slowing. The expression “build it and they will come” has been replaced by “lease before it’s built,” but the underlying reasoning behind securing capital before sound economics is the same, and it should cause some concern.
This does not imply that the subsea boom is entirely made up. There is a genuine need. In actuality, the infrastructure is getting older. For a civilization this reliant on the underground, the blind spots—we still know more about the lunar surface than the ocean floor—are truly embarrassing. The first businesses to find a dependable, cost-effective, long-lasting solution for subsea sensing and autonomy will probably be in a very advantageous position. That result is conceivable. Simply put, it’s not inevitable yet. It’s difficult to ignore how frequently the pitch’s urgency outpaces the engineering when observing this area. For 150 years, the ocean has been referred to as a frontier. Though mostly in the brochures thus far, the tide may finally be turning.
