North Americans are really uneconomical shipbuilders, and their navies should demand better.

Ford announced this week that the company will stop building small cars in Michigan, or anywhere in the States, as price pressure precludes paying workers what’s worth their while. Ford will build its new Lincoln Continental in Michigan, but that’s because the profit margins on luxury sedans can support the prevailing wages. As those twin reports in the Wall Street Journal reminded us, not all companies in North America are equally adept in all tasks, so not everything should be built in the US and Canada. That’s just Ricardian comparative advantage. So why are North American military forces trying to buy everything there? The cost penalties are particularly dire in shipbuilding, but they could be otherwise, with a really serious effort at innovation.

To understand the problem, let’s start with another announcement this week—that the Royal Canadian Navy will be renting two resupply ships this year, from Chile in the Pacific and Spain in the Atlantic. The RCN is still aiming to get its two new supply ships—HMCS Queenston and HMCS Châteauguay—from Vancouver Shipyards, sometime in 2019. That source was selected in October 2013, with construction to begin in 2016. In the interim, HMCS Preserver has been laid up in Halifax as uneconomical to repair, and HMCS Protecteur suffered her fatal fire underway between Pearl Harbor and Esquimalt. Those ships were, after all, 45 years old.

With those commercial arrangements filling two gaps, one might wonder why the DND doesn’t just buy commercially-derived tankers now? Britain’s Royal Fleet Auxiliary is sourcing its four new Tide-class tankers from Daewoo Shipbuilding and Marine Engineering. The contract was signed in 2012, the keel of the first ship was laid in January, and the first ship is scheduled to be delivered next year. As I estimated back in November 2013, the two British Columbian ships are forecast to cost Canada over 14 times per ton what the South Korean ships are to cost Britain.

That’s insanity, but it doesn’t prevail everywhere. In contrast, the Tide tanker is sufficiently fetching that the Norwegian Navy is ordering one too. The Royal Australian Navy is interested in a Tide from Daewoo, but perhaps instead a Cantabria from Navantia. Either way, as the defense minister put it last year, “Australia is not in a position to manufacture those vessels—20,000 tonnes and above—and accordingly we will see those ships produced either in Spain or in South Korea.” There’s no point in building a shipyard just to build one ship.

But another notable fact about the Tide-class is that the ships were designed in England, by BMT Defense Services. And why not? Supply chains are pretty globalized today. Like its predecessors, Apple’s latest A9 processor was designed in-house in California. Actual manufacturing is said to be underway in Texas by arch-rival South Korean firm Samsung, and on Taiwan by TSMC. Final assembly of Apple’s phones and computers is mostly in China, though that process adds but a slight percentage of the value.

Meanwhile, constrained by law, the US Navy has released its RFP for new oilers to but two shipyards, in California and Mississippi. By my earlier estimate, these will cost only slightly more than twice as much per ton as a South Korean ship. Fairly though, in a massive naval war with China, sourcing new ships from yards in South Korea could be challenging—if the yards are not bombarded out of business, they may be building at full clip just for the home fleet. Ships aren’t cell phones, so something remains to be said for security of supply.

So between madly overpaying and sourcing overseas, what are the alternatives? There is that innovation thing. Out in San Diego, NASSCO at least has been working with Daewoo to build its ships more cost-efficiently. But is something more radical possible? In aviation, Boeing and Northrop Grumman are individually rumored to be working on seriously new manufacturing technologies in advance of that Long-Range Strike Bomber project. Lieutenants Cheney-Peters and Hipple of the US Navy have written about their eagerness to “print a cruiser”—harnessing additive manufacturing advances for shipbuilding. Whatever the means, figuring out how to build modern cargo ships as fast as Henry Kaiser once built Liberty ships would be very useful in wartime. It would also get around the need for bumming rides from otherwise gracious Chileans and Spaniards.

Manning is a related question. As Military Times recently told it, Michael Jones of Google Earth fame is bemused by labor-intensive American carriers: why can’t the aircraft be refueled and repositioned by robots? For decades, aircraft aloft have refueled with a handful of aircrew on each side—and now drones can refuel drones. Ships shouldn’t be much harder. Indeed in Norway, Rolls Royce Marine is working on the concept of wholly unmanned cargo ships. Habitation is a big part of ships, so what would one do with all that freed-up space? Missiles and marines, I say. Distributed lethality and alternative at-sea basing are the naval watchwords this year, and should be for years to come.

More should be possible here. For whatever Ford’s examples, Tesla never seriously considered building cars anywhere but stateside. Now ensconced at the old NUMMI plant in California, Elon’s third outfit is making work what GM and Toyota couldn’t. Saliently, it’s with a radically different approach to what constitutes a car. Just maybe that’s what’s needed in shipbuilding as well.

James Hasík is a senior fellow at the Brent Scowcroft Center for International Security.