NATO runs on a handshake

The biggest military alliance on earth has no enforcement mechanism. NATO's Article 5, the clause that's supposed to guarantee collective defence, doesn't actually require anyone to do anything specific. A member state could respond to an attack on an ally by sending troops, imposing sanctions, or, as one analyst put it, "sending flowers instead of armies." It's a handshake dressed up as a treaty. Trump just reminded everyone what that means. His repeated threats to pull the U.S. out of NATO, now escalating against the backdrop of the Iran war, have exposed a truth that's uncomfortable for anyone who builds on top of systems they don't control: every contract that relies on goodwill eventually gets tested. This isn't just a geopolitics story. It's a pattern that shows up everywhere, from military alliances to software architectures to business partnerships. The question isn't whether the handshake will hold. It's what happens when it doesn't.

The treaty that isn't one

NATO's Article 5 is often described as the cornerstone of Western security. Invoke it, and all 32 member states rush to your defence. That's the narrative, anyway. The actual text is far more circumspect. Each member commits to take "such action as it deems necessary, including the use of armed force." The key phrase is "as it deems necessary." There's no minimum response. No penalty for inaction. No enforcement body. The North Atlantic Council can invoke Article 5, but it cannot compel a single soldier to move. As Clara Riedenstein at CEPA noted, the vagueness is by design. It "leaves open the door for less enthusiastic allies to send coal oil rather than armies." The Brennan Center's analysis is even more direct: Article 5 "does not require any member to respond with military force, although it permits such responses." This constructive ambiguity worked for decades because the implicit bargain was clear. The U.S. provided the security umbrella. Europe spent less on defence. Everyone understood the deal even if nobody wrote it down in enforceable terms. Then the terms changed.

When the biggest node threatens to disconnect

Trump has been pulling at NATO's threads since his first term. In 2020, the Pentagon announced plans to cut a third of U.S. troops in Germany. "Germany is not paying their bills," he said. Congress blocked the withdrawal. But the signal was sent. The pattern escalated. During the Greenland dispute in early 2025, Trump refused to rule out leaving NATO entirely. By mid-2025, at the NATO summit in The Hague, the tension was palpable. Then came the Iran war. When the U.S. and Israel launched strikes against Iran in late February 2026, the administration expected allied support. It didn't materialize in the way they wanted. NATO allies largely declined to participate in the Iran operation. Trump's response was to openly threaten reconsidering U.S. membership in the alliance. Congress had already anticipated this. The 2023 National Defense Authorization Act requires a two-thirds Senate majority for any NATO withdrawal. But as the Verfassungsblog's legal analysis pointed out, you don't need to formally leave an alliance to hollow it out. The concept of a "dormant NATO," where the U.S. remains a member on paper while withdrawing meaningful commitment, has gained traction in policy circles close to the administration. The result is an alliance where the most powerful member is actively signaling that the handshake might not hold. And everyone downstream is scrambling to figure out what that means.

One chokepoint breaks everything

The Iran war has provided a brutal stress test, not just for NATO, but for every system that depends on concentrated points of trust. When Iran closed the Strait of Hormuz on March 4, 2026, in retaliation for U.S. and Israeli strikes, it shut down roughly 20% of the world's oil supply in a single move. Brent crude surged past $120 a barrel. QatarEnergy declared force majeure on all exports. Gulf state oil production dropped by at least 10 million barrels per day within two weeks, the largest supply disruption in the history of the global oil market. The World Economic Forum called it a demonstration that "modern power still runs through a surprisingly small number of vulnerable trade routes." CNN reported that the Pentagon and National Security Council had "significantly underestimated Iran's willingness to close the Strait." One chokepoint. One actor deciding not to honour the implicit agreement that shipping lanes stay open. That's all it took. The parallel to technology is hard to miss. Your cloud provider, your payment processor, your critical API dependency, these are all chokepoints. They work because both sides honour the contract. But the contract is often thinner than you think. Terms of service change. APIs get deprecated. Pricing models shift overnight. The vendor can rug-pull you the same way Iran rug-pulled global energy markets.

Alliances as protocols

If you squint, NATO looks a lot like a network protocol. It defines how nodes should communicate and cooperate. It establishes expectations. But like any protocol, it only works when participants choose to follow it. HTTP works because browsers and servers agree on the spec. TCP works because both ends honour the handshake. But there's nothing stopping a node from dropping packets, closing connections, or changing behaviour unilaterally. The protocol assumes good faith. It doesn't enforce it. Article 5 is the same. It's a protocol for collective defence that assumes all parties will act in good faith when invoked. The one time it was actually triggered, after 9/11, the response was broad but voluntary. Each nation chose its own level of involvement. The protocol held because the political will existed. The protocol itself didn't create that will. This is the fundamental fragility of any system built on trust rather than enforcement. It works until it doesn't. And you usually find out which one you're in at the worst possible moment.

Singapore's playbook: never depend on one partner

Singapore has been navigating this reality since independence. A city-state of six million people surrounded by larger neighbours, it has never had the luxury of assuming any alliance would hold unconditionally. The result is a foreign policy built on pragmatic diversification. As Foreign Minister Vivian Balakrishnan put it: "We do not take sides. We uphold principles." Singapore maintains deep defence cooperation with the U.S., trains military personnel with dozens of countries, and was the first Southeast Asian country to join the Global Coalition against ISIS. But it also maintains constructive relationships with China and avoids being locked into any single alliance. At the 2025 Munich Security Conference, Singapore's Defence Minister warned that the U.S. risked shifting "from liberator to great disruptor to a landlord seeking rent." It was a remarkably direct statement from a country known for diplomatic restraint, and it reflected a clear-eyed assessment that the old alliance structures were being renegotiated in real time. Singapore's Total Defence strategy, launched in 1984, extends this thinking beyond military security to encompass economic resilience, social cohesion, and technological self-sufficiency. The premise is simple: if every system that relies on goodwill eventually gets tested, you'd better have fallbacks. It's the same logic that drives good engineering. Don't build on a single point of failure. Don't assume your upstream dependency will always be there. Have a plan for when the handshake breaks.

What sovereignty actually means

The word "sovereignty" gets thrown around a lot, usually in the context of borders and flags. But in an interdependent world, sovereignty is really about optionality. It's about having enough independence that no single actor's decision can break your system. For nations, that means diversified alliances, domestic defence capability, and economic resilience. For companies, it means multi-cloud strategies, avoiding vendor lock-in, and owning your critical infrastructure. For individuals, it means not building your career, your income, or your platform on a single dependency you don't control. The Iran war has made this viscerally clear. Countries that depended heavily on Gulf energy imports are now scrambling. European nations that outsourced their security to the U.S. are rushing to build independent defence capabilities. The Dallas Fed's analysis of the Strait of Hormuz closure paints a picture of cascading failures, each one traceable back to a concentration of trust in a single chokepoint. The pattern repeats at every scale. Every system that concentrates trust in a single node inherits the risk of that node defecting. The Strait of Hormuz, NATO's Article 5, your primary cloud provider, they're all the same architecture. And the architecture has a known vulnerability.

The handshake holds until it doesn't

None of this means alliances are worthless. NATO has kept the peace in Europe for over 75 years. The global shipping system has moved trillions of dollars in goods through the Strait of Hormuz. Your cloud provider probably has better uptime than you could achieve on your own. But the lesson of 2026 is that the cost of misplaced trust is asymmetric. Everything works fine right up until the moment it doesn't, and the failure mode is catastrophic, not graceful. The smart move isn't to abandon alliances. It's to build them while planning for their failure. Honour the handshake, but don't bet everything on it. Diversify your dependencies. Build in redundancy. And pay attention when the biggest node in your network starts talking about disconnecting. Because alliances, like APIs, only work when both sides honour the contract. And the contract was never as binding as you thought.