Build Warships at Home, Buy Tankers with Allies

A single number explains why Washington’s sudden shipbuilding enthusiasm needs discipline: China builds over half of the world’s commercial vessels. That headline triggers a predictable reaction—“We need to build everything here again.” Most companies (and governments) get this wrong. Capacity is not the same as capability, and the fastest way to waste both money and time is to chase symbolism instead of battlefield utility.

If your priority is deterring a peer competitor in the Pacific, the ships that decide outcomes aren’t generic merchant hulls. They’re attack submarines, unmanned surface vessels, undersea drones, resilient command-and-control nodes, and the maintenance infrastructure that keeps them available. The uncomfortable reality is that the United States can’t afford to treat commercial shipbuilding as a national-security proxy when it directly competes with the specialized yards, workforce pipelines, and supply chains needed for high-end naval systems.

This is where the “AI in Defense & National Security” lens changes the conversation. The question isn’t only where hulls get welded. It’s how quickly the U.S. can design, build, certify, crew, sustain, and fight a force that’s increasingly autonomous, software-defined, and intelligence-driven. If we want a Navy that can actually execute concepts like distributed maritime operations and the much-discussed Taiwan “hellscape,” shipbuilding policy has to prioritize warship throughput plus operational readiness, and AI has to be part of the plan.

Commercial shipbuilding won’t fix the Navy’s bottlenecks

Answer first: Expanding domestic commercial shipbuilding does not materially increase production of the naval platforms that matter most in a high-end conflict, because the constraints are specialized facilities, skilled labor, and certification pathways—not a lack of places to build cargo ships.

The U.S. Navy’s production and readiness problems are happening now, not in a hypothetical future surge. Submarines are the clearest example. Nuclear-powered boats require:

• Nuclear-qualified trades and engineers
• Highly specialized module construction and integration
• Strict quality assurance and regulatory compliance
• A supply chain that can consistently deliver long-lead components

A yard optimized for commercial tonnage simply doesn’t translate into faster submarine output. Treating it as interchangeable capacity is like assuming a factory that makes refrigerators can “surge” to make jet engines.

The near-term priority: readiness and repair capacity

Ship numbers look good in speeches; availability wins wars. Maintenance delays, drydock scarcity, and parts shortfalls reduce effective combat power faster than any shipbuilding press release can offset. If you’re trying to close readiness gaps in 2026–2029 (the window most PacificC-like scenarios focus on), money should go to:

1. More repair throughput (public yard modernization, and yes, additional drydock capacity)
2. Stable demand signals that make private investment rational (multi-year procurement where feasible)
3. Supplier capacity expansion for propulsion, power systems, castings/forgings, and electronics

That’s not anti-industrial policy. It’s targeted industrial policy.

Buy the merchant ships—secure them with smart programs

Answer first: The U.S. can strengthen sealift and tanker access faster and cheaper by partnering with allied shipbuilders while using contracting, flagging, and crewing requirements to guarantee wartime availability.

If the goal is strategic sealift and assured logistics, the U.S. already has workable scaffolding in programs like the Maritime Security Program and Tanker Security Program. The practical play is to expand and modernize these constructs rather than force domestic construction of every merchant hull.

Why? Cost and speed. Allied shipyards—especially in Japan and South Korea—can deliver commercial vessels at scale, and domestic construction has been estimated at roughly four times allied cost in some categories. When you’re staring at submarine backlogs and maintenance deficits, paying a 4x premium for merchant hulls is a direct trade against combat capability.

“But what about coercion?” It’s a real risk—just not the main one

China’s influence over shipping is often framed as a chokehold. It’s not nothing, but it’s commonly overstated for two reasons:

• Merchant ships last ~30 years, so coercion via shipbuilding order books moves slowly.
• Ownership and operations are diffuse globally. By one measure, China owns about 12.2% of the global commercial fleet—less than Greece (~17.6%) and comparable to Japan (~10.9%).

The more relevant vulnerability isn’t that the world “runs out” of ships overnight. It’s that in crisis conditions you can lose access to:

• crews with the right clearances/training
• insurance and reinsurance coverage
• port services and repair slots
• trusted logistics data flows (cargo manifests, AIS anomalies, routing)

That’s not a shipyard problem. It’s a governance, contracting, and data problem—and it’s a place where AI-enabled risk scoring and logistics intelligence can materially improve resilience.

The LNG carrier trap: a case study in costly symbolism

Answer first: Forcing U.S.-built liquified natural gas carriers for exports produces minimal military value and consumes capital, labor, and dock time that should go to naval readiness.

A proposed requirement that 15% of U.S. LNG exports travel on U.S.-built ships by 2047 is a useful example of how good intentions turn into bad strategy. LNG carriers are complex, niche vessels. Building them domestically would demand specialized expertise (like membrane containment systems) that doesn’t map cleanly to building submarines or unmanned combatants.

Using 2024 export volumes and typical carrier capacity, meeting that 15% target implies roughly 20 large LNG carriers. If U.S. build costs are 4x South Korean prices, the additional cost can land around $15 billion—before counting the new infrastructure and training required.

Even if you like the idea of export control and routing resilience, there’s a cheaper way to buy it:

• Incentivize allied-built LNG carriers to be U.S.-flagged or operated under assured charter agreements
• Require wartime availability clauses, pre-negotiated routes, and rapid activation conditions
• Fund port resilience, cyber protection, and logistics analytics that keep shipments moving

If the U.S. wants to spend $15B for maritime security, it should buy outcomes that matter under fire: submarine availability, contested logistics, and autonomous maritime sensing.

Where AI actually fits: shipbuilding as a software-and-data problem

Answer first: AI doesn’t replace shipyards; it increases warship output and readiness by shrinking design cycles, predicting failures, and optimizing constrained resources like drydocks, suppliers, and skilled labor.

“AI in defense” often gets framed as autonomous weapons or fancy drones. That’s part of it, but the bigger opportunity for near-term maritime advantage is quieter: AI for industrial throughput and fleet readiness.

Here are four concrete areas where AI delivers real leverage in naval shipbuilding and sustainment—without pretending it’s magic.

1. AI-assisted design and certification readiness

Warship programs lose years to late-stage redesign and rework. AI-enabled engineering tools can:

• detect interference and producibility issues earlier (before steel is cut)
• flag certification risks tied to materials, weld processes, and QA records
• generate alternative design variants that preserve mission performance while easing manufacturability

The goal isn’t prettier CAD. It’s fewer change orders and fewer surprises on the shop floor.

2. Predictive maintenance for submarines and auxiliaries

Readiness is a math problem: if you can forecast failures and schedule maintenance windows intelligently, you increase effective fleet size without building a single new hull.

AI models trained on:

• sensor telemetry
• maintenance histories
• parts consumption
• operational profiles

can predict likely failures (and their lead times), improving:

• parts pre-positioning
• drydock scheduling
• workforce allocation by trade (pipefitters vs electricians vs welders)

3. Drydock and depot optimization (where the real bottleneck lives)

Public and private yards operate as complex constraint networks: drydock slots, crane availability, hazardous work sequencing, QA inspection staffing, and long-lead parts.

A modern AI planning stack can function like an air traffic control system for ship maintenance:

• optimize schedule under constraints
• simulate “what-if” disruptions (supplier delay, workforce shortage)
• prioritize jobs based on operational value, not just queue order

This matters because an extra submarine deployment next year can be more valuable than a ceremonial commercial launch today.

4. Countering China’s scale with logistics intelligence

China’s industrial scale is real. The U.S. answer shouldn’t be “copy it everywhere.” It should be out-decide and out-coordinate.

AI-enabled logistics and intelligence analysis can:

• detect supply chain coercion attempts early
• map dependencies across tiers (castings, specialty steels, electronics)
• identify “single point of failure” suppliers and propose alternates
• improve contested logistics routing for tankers and sealift in crisis

Scale is one axis of competition. Decision superiority is another.

Snippet-worthy stance: If you’re spending shipbuilding dollars, spend them where AI and specialization multiply combat power—submarines, unmanned systems, and maintenance capacity—not on symbolic merchant hulls.

A practical playbook for policymakers and defense leaders

Answer first: Treat maritime industrial policy as a portfolio: build high-end naval capability at home, buy commercial tonnage from allies, and use AI to raise throughput across design, build, and sustainment.

Here’s what I’d recommend as a crisp, execution-oriented approach for 2026 planning cycles.

Build at home (non-negotiable)

• Nuclear submarines and their supplier base
• Unmanned surface and undersea systems (including rapid iteration lines)
• Naval auxiliaries and repair infrastructure that increase fleet availability
• Cyber-resilient shipboard networks and mission systems integration

Buy from allies (smart and fast)

• Tankers and commercial sealift hulls through allied yards
• LNG carriers (if needed) through allied production, paired with security-focused contracting

Wrap it in “assured access” contracting

• Wartime activation clauses
• Pre-cleared crewing and training pipelines
• Data-sharing requirements for logistics visibility
• Cybersecurity controls for port-call, cargo, and routing systems

Use AI where it pays back quickly

• Maintenance prediction and parts forecasting
• Yard schedule optimization
• Supplier risk analytics and multi-tier mapping
• Design-for-manufacture checks early in the lifecycle

What this means for the AI in Defense & National Security series

Maritime power isn’t only about hull counts. It’s about availability, autonomy, and the ability to sustain operations under pressure. That’s the connective tissue between shipbuilding strategy and AI in national security.

If the United States tries to recreate an entire commercial shipbuilding industry from scratch, it risks starving the exact programs that deter conflict: submarines, unmanned maritime systems, and the industrial sustainment backbone. If it partners with allies for merchant tonnage and uses AI to remove friction from naval production and readiness, the U.S. gets more combat power per dollar—and sooner.

If you’re making decisions in defense acquisition, maritime logistics, or fleet readiness, the next useful question isn’t “How do we build more ships?” It’s “Which ships—and which AI-enabled processes—raise the probability of mission success in the next five years?”