The Unseen Threat to the AI Bubble: A Financial Time Bomb Waiting to Detonate

The recent warnings about the dangers of uncontrolled AI have dominated headlines in recent weeks. However, beneath the hype and hyperbole surrounding potential existential risks, a more pressing concern lurks in the shadows – the financial instability of the entire industry.

The tech giants and their investors tout the supposedly limitless potential of AI while building a house of cards on shaky ground. The $132 billion in debt issuance this year alone to fund datacentre rollouts by hyperscalers like Google, Amazon, Microsoft, Meta, and Oracle is staggering, especially when considered against fragile bond markets and rising borrowing costs.

The unit economics of AI are a major concern. Prices have collapsed while costs continue to rise. OpenAI’s repeated fee cuts in an effort to retain customers only mask the deeper issue: that the maths simply doesn’t add up without epic revenue growth. When the inevitable downturn comes – as it inevitably will – the consequences could be catastrophic.

A recent research note by Groundbreaker has shed light on a third, even more insidious threat: the $1.5 trillion “compute commencement wall” facing AI labs over the next couple of years. This is not just debt; it’s a financial promise that underpins the entire boom, threatening to unleash a perfect storm when obligations can’t be met in full.

The comparison with the 2007-2008 global financial crisis is apt. Teaser mortgage rates were unsustainable, and so too are the “take or pay” contracts being used to fund datacentres. When these costs come online, they will be a rude awakening for an industry that has grown accustomed to propping itself up with leverage.

Calls for AI regulation and limits on technological progress are necessary and timely, but they should not distract from the more pressing issue of financial stability. The tech giants’ hubris and recklessness pose the greatest threat to humanity – at least in the short term.

The risks are not confined to Silicon Valley; a collapse of the AI bubble would have far-reaching consequences for the global economy, affecting supply chains and employment markets. Cory Doctorow astutely observes: “These companies are claiming that they are so cool that their profitability can only be measured using a novel, secret form of mathematics.”

The truth is both more prosaic and more terrifying. The AI bubble may not take over the world just yet, but its implosion would have a far greater impact on our daily lives than any hypothetical robot uprising.

A House of Cards Built on Sand

The financial structures underpinning the AI boom are anything but robust. Investors cheer on companies like OpenAI and Anthropic as they book expected future revenue from “take or pay” contracts, creating a ticking time bomb waiting to detonate. The maths only works if revenue continues to rocket – but what happens when it doesn’t?

A Global Financial Crisis Redux

The parallels with the 2007-2008 global financial crisis are too striking to ignore. Just as then, the masterminds behind the AI boom have created a system that is both opaque and unsustainable. They tout their supposed “novel, secret form of mathematics” to obscure the basic fact: that the economics of their businesses don’t work.

The Unseen Threat to Humanity

The stakes are high indeed, extending far beyond the tech giants themselves. A collapse of the AI bubble would send shockwaves through supply chains, employment markets, and economies worldwide. It’s a sobering thought that should give pause to policymakers and regulators alike.

In the end, it’s not the hypothetical dangers posed by uncontrolled AI that we should fear most – but rather the financial instability and hubris of the tech giants themselves. For in their reckless pursuit of profit and power, they may just be building a house of cards that will ultimately bring them crashing down.