We are told, incessantly, that we have entered the age of intelligent warfare. Algorithms will outthink our adversaries. Artificial intelligence will see patterns invisible to the human mind. The battlefield will be governed not by the chaos of flesh and blood, but by the cold, precise logic of the machine.
What a lovely fiction. And what a dangerous one.
Let’s be honest about what we’re actually building. The modern warfighting capability—the one being assembled in laboratories from California to Shenzhen—does not rest on silicon valour. It rests on a substance so whimsical, so absurdly fragile, that it escapes from any container not welded to within an inch of its life. Helium. The same stuff that lifts party balloons and makes our voices squeak. It is indispensable to the manufacture of the most advanced semiconductors on Earth. Without it, the extreme ultraviolet lithography machines that etch the brains of our smartest weapons simply stop.
There’s no substitute. There’s no workaround. There’s only the quiet, desperate logic of the supply chain.
Most of the world’s supply arrives as a byproduct of natural gas, processed in facilities that are concentrated in a handful of locations—Qatar being one of the more consequential. To move it, you must cool it to temperatures approaching absolute zero and transport it in vessels so exquisitely sensitive that a delay of days can mean the difference between delivery and dissipation. This alchemy is performed on a global scale, and we have somehow convinced ourselves it is a foundation for strategic superiority.
Now let us talk about the Straits of Hormuz – a ribbon of water through which a substantial fraction of the world’s energy still flows. A missile fired there, a tanker seized, a confrontation escalated—these are not simply geopolitical events. They are interruptions in the metabolism of an industrial civilisation that no longer remembers how to function without them. And those interruptions do not respect the boundaries we imagine between energy, manufacturing, and computation.
Because on the other side of the world, in Taiwan and South Korea, the furnaces of semiconductor fabrication burn ceaselessly. These facilities—particularly those of a single company in Taiwan that produces the overwhelming majority of the world’s most advanced chips—require not only helium but also water so pure it would dissolve glass, energy so reliable it would shame a Swiss clock, and a flow of materials that traverses exactly the same contested waterways. This entanglement is not a condition that favours the technologically ambitious.
What I have come to call recursive fragility is the condition of any system that depends on the conditions it’s actively degrading. A military campaign that destabilises energy corridors doesn’t only threaten oil supplies. It threatens the industrial manufacturers that produce their own computational advantage. A conflict that interrupts the flow of rare gases and exotic metals does not merely inconvenience the economy. It erodes the self-same technologies upon which that conflict relies. The system begins to consume itself. Not with a bang, but with a cascade of failures that propagate faster than any command structure can respond.
Take the much-vaunted promise of artificial intelligence in warfare. Programs like Project Maven, which feed satellite imagery and drone footage into machine learning systems, are presented as the vanguard of a new era. But these systems are not autonomous in any meaningful sense. They are contingent on chips that require helium that travels through straits that can be closed by a single reckless act. They are contingent on fabrication plants that require uninterrupted energy that flows through the same contested arteries. They are contingent on logistics that assume the world remains, against all evidence, stable. Disrupt the chain, and the intelligence dissolves. Not gradually. Not predictably. Like a thought interrupted mid-sentence. Much like Donald Trump’s rhetoric, come to think of it.
This is not a problem to be solved by stockpiling. It’s not a vulnerability to be patched either. It’s a condition of existence for the entire structure of what we call modern warfare. And the truly uncomfortable truth—the one we’re not yet ready to confront—is that the more technologically advanced warfare becomes, the more it depends on conditions that war itself destroys. We have built a warfighting capability that can’t even survive the wars we intend to fight with it.
The dominant worldview of our time—what I have elsewhere called industrial economism, that extractionist, predatory, and ultimately toxic paradigm—has produced a global system of such intricate interdependence that no actor, no nation, no corporation fully comprehends its limits. Taiwan’s semiconductor dominance is not an independent strength; it’s a mutual entanglement with China’s control over critical materials, with the United States’ standing in AI architectures, with Qatar’s helium, and with the Strait of Hormuz. These are threads in a tapestry that any child with a pair of scissors could unravel.
And so we arrive at the paradox that defines the 21st century: the side that prevails in future conflicts may not be the one with superior firepower, nor even the one with smarter algorithms. It will be the one that sustains the integrity of its unseen systems—materials, energy, logistics, fabrication—for one day longer than the other.
Or to be more precise, it will be the one that fails more slowly. Because in a tightly coupled system, collapse is never unilateral. It propagates. It leaps from supply chain to supply chain, from sector to sector, from economy to economy, until the distinction between victor and vanquished dissolves into the same fog of shattered interdependence.
We are engineering systems that cannot survive the conditions under which we intend to use them. We are building a machine that, when activated, begins to eat its own limbs. And we call this progress?
I don’t offer solutions here. The problem-solution dilemma is itself one of the traps of industrial thinking—the assumption that every pathology has a prescription, that every failure can be fixed with more of the same. Perhaps what’s required is not a better supply chain but a different relationship to the idea of war itself. Perhaps the real intelligence we lack is not the kind that fits on a chip but the kind that recognises the futility of building weapons that consume the worlds they are meant to defend.
But that would require a kind of stewardship we have not yet learned to practise: the collective capacity “to turn conversations that matter into actions that make a difference”, not by outcompeting one another, but by recognising that in a system of mutual entanglement, survival is not a zero-sum game. It’s a shared condition. And we are, all of us, in the same fragile vessel, sailing the same unpredictable straits.