I wrote recently about the international order coming apart . About how America is trading structural influence for transactional leverage, and how the uncertainty alone — whether or not the commitment was ever formally broken — has already changed the strategic landscape. The question I ended on was what the rest of the world does next.

I’ve been thinking about that question from where I sit: a small European country, inside the system that’s being destabilized. The more I look at what needs to happen, the more it resembles a problem I recognize from my own work — and the more I suspect the hardest part isn’t figuring out what to build. It’s that the type of problem demands exactly the kind of work we’re worst at.

The headlines are full of the flashy stuff. New fighter jet contracts. Drone programs. Billion-euro defense spending announcements. These matter. But in engineering, the systems that fail rarely fail at their most visible components. They fail at the joints nobody thought about. The monitoring that wasn’t set up. The capacity planning that got deferred. The documentation that didn’t exist when the person who understood the system left. The glamorous feature ships. The infrastructure underneath it doesn’t.

Deterrence works the same way. It isn’t the fighter jet that deters — it’s the manufacturing capacity to produce ammunition for years, the logistics network that can move supplies under fire, the reserve training infrastructure that can mobilize a population in weeks, the stockpiles that mean you don’t run out in the first month. It’s the boring stuff. And boring stuff doesn’t make headlines, doesn’t win elections, doesn’t photograph well at summits. It gets funded last and cut first.

The obvious question is whether any of this can be done alone. For most small and medium-sized countries, the answer is no — not all of it. But the instinct to network everything, to solve the problem collectively, runs into a tension that took me a while to sort through: you can’t outsource the things you need to survive if the network fails. Machiavelli wrote about this five centuries ago, in the context of mercenaries. A state that relies on others for its fighting capacity has outsourced something more fundamental than labor — it has outsourced the political will that comes from a population defending its own. The insight still holds.

Sovereign what you need to survive alone. Networked what you need to compete at scale. The dividing line follows the economics of the threat. Where the attacker’s cost per unit is low — drones, loitering munitions, the kind of cheap attrition warfare Ukraine has been living through since 2022 — the defender needs sovereign production capacity. You can’t afford to run out of interceptors because a supplier had a delayed shipment. You need to be able to build these yourself, operate them yourself, sustain them yourself. The recent conflict in the Middle East illustrated what happens when you use expensive interceptors against cheap drones — you run through your stockpile at a rate that’s economically and strategically unsustainable.

Where the attacker’s cost per unit is high — strategic missiles, advanced platforms, precision-guided munitions at scale — networked defense makes more sense. The adversary can’t sustain an infinite barrage. Integrated air defense, shared early warning, joint procurement of complex systems — these benefit from scale in ways that individual countries can’t replicate. Finland and Sweden are already building this kind of complementarity: Finland provides the mass land defense capability it has spent decades preparing, Sweden provides strategic depth and air basing. Neither planned this division of labor from the top down. Their threat environments shaped it.

The Ukraine-Gulf exchange is perhaps the sharpest illustration. Ukraine has developed world-class drone defense expertise under existential pressure. Gulf states have Patriot interceptors suited to a different threat profile. They’re trading capabilities: Ukrainian expertise for Gulf hardware. This isn’t institutional allocation. It’s emergent complementarity driven by mutual need. Countries develop what their environment forces them to develop, then find partners whose gaps complement theirs.

The formal institutions — the EU, what remains of functioning NATO coordination — provide the interoperability standards, the financing instruments, the procurement frameworks. Not central planning. Infrastructure that enables the market to function.

Then there’s nuclear deterrence. The western world has been living under America’s nuclear umbrella for seventy years. That umbrella is degrading. CSIS assessed that the French and British arsenals — roughly 290 and 225 warheads respectively — would not constitute a viable replacement in the event of an abrupt American withdrawal. France’s nuclear doctrine is built around retaliation against existential threats to France, and Macron’s recent moves toward extending coverage are presidential decisions that a successor could reverse.

The uncomfortable lesson is the Budapest Memorandum. In 1994, Ukraine gave up its nuclear weapons in exchange for security assurances from Russia, the US, and the UK. Those assurances proved worthless. The lesson every non-nuclear state is forced to draw: the bargain of the Non-Proliferation Treaty — you don’t develop weapons, we guarantee your security — is only as strong as the guarantee.

I don’t have a clean answer here. My sense is that conventional deterrence has to come first, because it’s faster to build and provides the baseline security within which the harder nuclear question can be addressed. The existing umbrella — degraded but not gone — buys time. But the gap is real, and pretending otherwise is a luxury that’s running out.

Energy is the other domain where independence and security converge. The framing has to shift: energy infrastructure isn’t just a climate project, it’s a defense project. Decentralized production, distributed storage, resilient grids — these are security investments that happen to align with the energy transition. When the Baltic states disconnected from the Russian-era BRELL grid in February 2025, they succeeded without blackouts. But they also exposed a new vulnerability: their connections to the European grid are limited, the undersea cables are targetable, and the redundancy is thin. They eliminated one chokehold and uncovered another. Strategic resilience isn’t the same thing as autonomy. It’s distributing your vulnerabilities so that no single point of failure is fatal.

Ukraine is the live case study. Repeated Russian strikes on centralized power infrastructure have accelerated a shift toward distributed generation — batteries, solar, modular cogeneration units creating energy islands for hospitals, water treatment, critical industry. The investments overlap almost entirely with what the energy transition demands anyway. That convergence makes the politics easier, which matters for sustaining investment across election cycles.

All of this — the defense infrastructure, the energy resilience, the industrial networks — rests on whether people will actually fight for what they have.

Deterrence requires convincing the adversary not only that you can fight, but that you will. Herman Kahn, the Cold War strategist who made a career of thinking about the unthinkable, analyzed thermonuclear war scenarios against three criteria: the impact on population, the impact on the economy, and whether the survivors would envy the dead. That third criterion is the one that translates here, in a less extreme form: are the people willing to fight for what they risk losing? Do they see the system as worth preserving?

A fractured society where large segments of the population don’t feel they have skin in the game is a society with a credibility problem. Not a moral one — a strategic one. I’ve written elsewhere about how systems accumulate accidental complexity until they serve the people who shaped them, not the people who live in them. The same dynamic applies here. A society that’s been maintained — kept legible, kept fair, actively refactored — is one whose members will fight for it. One that has accumulated unchecked complexity and inequality is one where the adversary finds fractures to exploit. Healthcare, education, the rule of law, some basic sense of fairness — these aren’t ideological luxuries.

The question that actually keeps me up is different: what makes any of this happen in systems that are structurally bad at sustaining boring, comprehensive, long-term work?

My answer is anxiety. The countries that are actually doing the boring work — Finland’s total defense doctrine, Estonia’s conscription system, Poland’s East Shield, the Baltic energy disconnection — are the ones closest to the threat. Proximity concentrates the mind. And the circle of anxiety is widening. Countries that could afford complacency a year ago are recalculating.

But anxiety is a double-edged instrument. The same force that drives Finland to maintain total defense readiness can, channeled differently, produce nationalism, fortress mentality, the kind of social fracturing that weakens exactly the cohesion deterrence depends on. Europe in the 1930s is the obvious cautionary tale — collective anxiety producing resilience in some countries and catastrophe in others.

What determines the direction, I think, is whether the anxiety is anchored in something worth defending. Not any particular political position, but the underlying conditions — the freedoms, the institutional norms, the capacity for disagreement and cooperation — that make different political positions possible in the first place. These are specific enough to fight for and universal enough to network across countries with different cultures and political traditions. They’re what distinguish an alliance of the anxious from a collection of fortresses.

There’s a circularity here that I can’t resolve. You need social cohesion to sustain the defense investments. The anxiety and sacrifice required for defense can erode social cohesion if the values aren’t strong enough to hold. Everything depends on the thing that everything else is supposed to protect.

Everything depends on the thing that everything else is supposed to protect.

I don’t know whether the values will hold. I don’t know whether democratic systems can sustain the kind of comprehensive, boring, decades-long investment that the situation demands — given short electoral cycles, fractured attention, the constant pull toward the flashy and the visible. I’m not even sure the timeline works. Russia may be capable of testing NATO within a few years. Europe needs five to ten years to rearm meaningfully. The gap is real.

But the alternative to starting is not starting. And the type of problem this is — one where the difficulty is the point, where the solution requires doing many unglamorous things simultaneously for a long time — doesn’t have a shortcut. You don’t get to skip the boring stuff because it’s boring. You pour the concrete and hope it sets before the storm arrives.