The AI Mirage

The euphoria surrounding artificial intelligence has reached a fever pitch. Valuations are soaring, capital is flooding in, and the world’s most powerful technology firms are bound together in a web of overlapping investments so intricate that few can tell where genuine demand ends and financial self-dealing begins. It feels exhilarating, and eerily familiar.

History has seen this movie before. Each technological revolution begins with an idea so transformative that it reshapes not just industries, but imagination itself. Then come the investors, the valuations, the limitless projections ~ until the market begins to believe its own fiction. Artificial intelligence is no exception. The sheer velocity of money moving into the sector suggests not only confidence, but dependence. When chipmakers invest billions in AI firms that in turn spend those billions buying their chips, what results is not organic growth, but a closed circuit of optimism.

Advertisement

At the centre of this cycle is the belief that AI’s potential justifies any price. That sentiment has driven private valuations into the stratosphere and concentrated an extraordinary amount of financial power into a handful of firms. Yet, beneath the glamour of innovation lies a sobering fact: most leading AI companies are not profitable. Their revenue curves are impressive, but they depend on enormous infrastructure investments and spiralling operating costs. When even insiders begin to question whether demand is being inflated through “vendor financing,” the comparison to past speculative bubbles becomes difficult to ignore.

The danger is not confined to Silicon Valley. The modern economy has become so tightly coupled to the promise of AI that a sharp correction could ripple across markets, from hardware manufacturing to cloud computing to real estate. Billions are being poured into data centres in deserts and hinterlands, each of them an immense facility whose environmental and financial costs will endure long after investor enthusiasm fades. If demand slows, these complexes may become the rusting monuments of a technological gold rush gone awry. It is tempting to dismiss these warnings as premature. Optimists argue that even overinvestment has its upside ~ that the infrastructure laid today will serve future generations of innovation, much as excess fiber-optic cabling built during the dot-com era later powered the digital age. That may well prove true.

But unlike the internet, which grew on open protocols and distributed creativity, today’s AI ecosystem is dominated by a few intertwined giants whose fortunes depend on each other’s solvency. If the flow of easy capital tightens, even temporarily, the illusion of infinite growth could collapse with startling speed. The AI revolution is real but its financial underpinnings are fragile. When markets begin to treat hope as collateral, the line between progress and speculation vanishes. The true test of this era will not be how much AI can build, but whether it can stand without the scaffolding of its own hype.