Steel Won the Last War, Software Will Decide the Next
It’s time to trade smokestacks for software and replace industrial nostalgia with strategic innovation.
From speeches on reshoring and supply chain security to op-eds championing "modern arsenals of democracy," policymakers are reviving the language of mid-century industrial might. The message is clear: America must reindustrialize to meet the strategic threat posed by China. And to do that, we must rebuild what worked during WWII.
But there's a problem: China isn’t Kaiser. And this isn’t 1942.
The Myth of Industrial Parallels
In the rush to evoke Rosie the Riveter and Liberty Ships, many in government are missing a crucial distinction: the nature of today’s geopolitical contest and the technologies it turns on are fundamentally different.
The WWII mobilization model was built around centralized, steel-based mass production. Speed came from scale, and scale came from physical industrial capacity. It was a system optimized for predictable, physical warfare, where more ships, tanks, and planes equaled more power.
But today’s battlefield is layered with software, sensors, networks, and asymmetric threats. Strategic advantage lies in adaptability, not just volume. The ability to rapidly iterate, reconfigure, and respond to new forms of attack matters more than the size of any initial inventory.
Trying to copy Kaiser Shipyards with 2020s tools is like trying to fight cyber threats with smokestacks. It risks missing the essence of what defines deterrence and resilience in the digital era.
The Shipbuilding Surge, Not a Strategic Plan
Recent calls from the Department of Defense, Secretary of the Navy, and Secretary of Defense to significantly scale U.S. shipbuilding capacity reflect real concerns. The Navy is stretched thin across multiple regions, and China's maritime build-up continues to outpace Western equivalents in both volume and pace. The response has been a renewed focus on growing the fleet and revitalizing shipyards.
Simply adding more shipyards and tonnage doesn’t equate to strategic advantage in the long term. It may meet near-term readiness metrics, but without a fundamental redesign of how naval platforms are conceived, manufactured, and deployed, we risk repeating the same cycle: years of lead time, limited flexibility, and soaring sustainment costs.
Worse still, this mindset assumes future conflict will mirror past patterns, large-scale fleet actions, predictable domains of engagement, and long production ramps. The reality is messier: distributed maritime operations, contested logistics, rapid theater escalation, and hybrid threats.
If every new destroyer or submarine is treated as a once-per-decade megaproject instead of a node in a modular, upgradable ecosystem, we haven’t modernized, we’ve just made bigger versions of legacy bets.
That’s why the most meaningful evolution in shipbuilding isn’t in tonnage, it’s in tooling. The increasing implementation of AI/ML across the naval enterprise holds the potential to transform:
Logistics optimization: predictive analytics for part failures, smarter inventory distribution, and responsive re-supply that adapts to mission dynamics.
Supply chain resilience: automated anomaly detection, risk flagging, and adaptive sourcing across vendor networks to anticipate and mitigate bottlenecks.
Maintenance cycles: condition-based and automated maintenance scheduling using onboard diagnostics and AI inference to reduce downtime and extend platform life.
Training pipelines: AI-driven simulation, skills assessment, and knowledge retention systems tailored to specific platforms and mission sets, improving warfighter readiness.
This isn’t about replacing humans, it’s about elevating them with better tools. Incorporating these technologies doesn’t just make shipbuilding smarter, it redefines the product itself. A ship is no longer just a steel platform with weapons. It’s a software-defined, data-integrated node in a constantly learning, dynamically updating system.
A Post-Industrial Defense Economy
The U.S. doesn’t need to reindustrialize. It needs to reimagine what defense production looks like in an age of AI, autonomy, and energy transition. It must pivot from heavy industrial centralization to light, modular, distributed capacity that thrives on speed, iteration, and resilience. This means shifting from mass production to distributed fabrication, from single-node supply chains to networked resilience, and from extractive manufacturing to sustainable, modular systems. It means viewing defense production less like a factory floor and more like a software pipeline.
Consider what a truly modern defense economy could look like:
Additive manufacturing nodes spread across the country, capable of printing spare parts and even whole platforms near the point of need, reducing logistical burden and increasing operational tempo.
Digital twins and AI-driven design loops enabling weapons systems to be updated weekly, not every five years. Entire platforms could be simulated, optimized, and stress-tested in virtual environments before a single part is printed.
Zero-emissions microgrids powering forward-deployed operations without requiring diesel convoys, reducing not just carbon output but logistical vulnerability.
Open standards that allow startups and allies to plug into core architectures without needing a decades-long integrator contract. Ecosystem interoperability becomes a force multiplier.
This vision isn’t speculative, it’s latent. The technology exists. What’s lacking is the structural permission to build this way.
The Risk of Looking Backward
The nostalgic push for reindustrialization isn’t just misguided, it’s actively risky. It channels capital into outdated infrastructure, centralizes production in brittle hubs, and locks the defense ecosystem into another generation of inflexible, over-optimized platforms.
It also crowds out alternative models. When the only acceptable demonstration of industrial seriousness is a giant shipyard, a steel mill, or a legacy production line, we lose the ability to imagine lighter, faster, and more resilient models of power projection.
Worse, it sends the wrong signals to the private sector. Startups and commercial tech firms are being told they must align with a 1950s playbook to earn legitimacy. That’s not a pathway to deterrence, it’s a roadmap to stagnation.
It reinforces a procurement culture that rewards compliance over creativity, maturity over maneuver, and sunk cost over strategic flexibility.
Industrial Policy for the Post-Industrial Age
If the U.S. wants to outcompete authoritarian rivals, it must embrace an industrial strategy that reflects this century, not the last. That means:
Funding distributed innovation infrastructure, not just megaplants. This includes regional hubs of excellence, dual-use tech incubators, and deployable fab units.
Rewriting acquisition rules to reward adaptability, not just compliance. The FAR must evolve to make space for iterative builds, off-the-shelf tech integration, and outcome-focused evaluation.
Recalibrating the national industrial base to integrate clean tech, AI-native platforms, and software-defined logistics. Climate security and national security are no longer separable.
Industrial policy can still matter. But it must be fit for the age of autonomy, not assembly lines.
This includes redefining shipbuilding not as an industrial relic, but as a testbed for new workflows: smart yards, modular hulls, AI-optimized supply chains, and deployable logistics networks that move at the speed of relevance.
Our competitors are not just building more, they're building differently. And unless the U.S. shifts from nostalgia to innovation, we’ll lose not because we were outnumbered, but because we were out-evolved.
The Bottom Line
The goal isn’t to rebuild the arsenal of democracy. It’s to build something better. An architecture that is more agile, more sustainable, more secure, and more worthy of the age we live in. We don’t need another Kaiser. We need the post-industrial equivalent of SpaceX and open-source hardware all rolled into one. That’s what the moment demands. Anything less is nostalgia dressed up as strategy.
To get there, we must act to:
Shift investment from monolithic infrastructure to modular, distributed capacity. Prioritize funding for additive manufacturing, mobile logistics, and scalable micro-yards.
Make AI-native logistics and sustainment a requirement in shipbuilding RFPs. Don't just encourage innovation, mandate it.
Stand up acquisition pilots that reward outcomes over process. Focus on cost-per-mission, time-to-field, and lifecycle flexibility, not just compliance.
Empower startups and non-traditional vendors with sandboxed pathways to scale. Create pathways that don't require them to mimic primes in order to survive.
Integrate commercial clean-energy and dual-use platforms into defense planning. If it moves supply faster, trains smarter, or fuels cheaper, it belongs in the toolkit.
Reform how talent is evaluated and rewarded across the defense industrial base. Promote builders, not just proposal writers.
Innovation doesn’t happen through memory. It happens through redesign. If we’re serious about readiness, deterrence, and agility, we must stop building monuments to the past and start constructing platforms for the future. The world won’t wait for us to catch up. Neither should our strategy.