When people hear the term "Ponzi scheme" or "financial bubble," they usually picture a mustache-twirling villain selling snake oil. They assume the underlying product must be a scam.

But history tells a much more dangerous story.

The most spectacular financial manias in human history don't happen around useless tech. They happen around highly potentially valuable, world-changing technology. Because the technology is so genuinely wonderful, the initial hype is justified. But then, a dangerous shift occurs — and that hype is hijacked, funneled, and magnified by vested interests purely to increase the value of their own investments.

The Internet Didn't Die, It Just Came to Its Senses

To understand where we are with Generative AI today, we must look back at the late 1990s.

When the Dot-Com bubble burst, critics claimed the internet was a fad. They were fundamentally wrong. The internet did change the world. It wasn't destroyed or ended; the market just finally came to its senses. The tragedy wasn't that the technology was a lie, but that the financial scaffolding built around it was entirely unsustainable, built on "eyeballs" and promises rather than revenue.

Today, Generative AI is on a remarkably similar track. But the stakes — and the burn rate — are exponentially higher.

The Gen AI Runaway Train

Right now, major US AI labs are on autopilot. They are locked in a prisoners' dilemma of existential proportions. They are burning billions of dollars on compute power, data acquisition, and talent, racing toward a moving finish line.

They cannot stop, even if they wanted to.

Consider the raw economic reality: Goldman Sachs reports that tech giants and venture capitalists are on track to spend an estimated $1 trillion on AI-related capital expenditure — including data centers, chips, and grid upgrades — over the coming years.

Meanwhile, venture capital analysis from Sequoia Capital reveals a staggering divergence: the AI ecosystem needs to claw in roughly $600 billion annually just to pay for its underlying infrastructure. As of today, actual revenue from generative AI applications is estimated to be only a tiny fraction of that amount.

The moment a single major player blinks — the moment they pause to normalize spending, or admit that the current enterprise adoption curve cannot support a trillion-dollar valuation — the illusion shatters. The house of cards falls.

So, they keep fueling the engine.

Murder on the AI Orient Express

It is easy to look for a scapegoat in this scenario. We love a villain narrative. But the current state of the AI hype cycle isn't the fault of a single "bad guy."

It is, in some ways, exactly like Agatha Christie's Murder on the Orient Express. No one person is solely guilty. They all did it.

Look closely at the ecosystem keeping this train on the tracks. The investors, who need to deploy massive funds and demand narrative-driven growth to mark up their portfolios. The founders, who must project absolute, world-altering certainty to secure the next multi-billion-dollar round. The consultants and influencers, who carve out lucrative careers telling every legacy board of directors that they will go bankrupt tomorrow if they don't integrate LLMs into their workflows today. The media, who trade in clicks generated by either utopian savior narratives or dystopian doom scenarios.

Everyone is feeding at the trough. Barclays recently noted that the capital expenditures of just the top cloud giants have decoupled completely from historic revenue trends, heavily driven by Wall Street's insatiable appetite for AI promises.

The Bottom Line

Generative AI is a profound technological leap. It will redefine industries, automate workflows, and create genuine value for decades to come.

But we must decouple the utility of the technology from the health of the market. A technology can be revolutionary while its current financial ecosystem is a runaway train heading for a wall.

Recognizing the bubble doesn't make you a cynic or a Luddite. It means you can see through the hype and are able to do some basic math.

Sources

• Goldman Sachs Research. "Gen AI: Too Much Spend, Too Little Benefit?" Projected $1 trillion in AI capex and analysis of whether productivity gains will justify the cost.
• Sequoia Capital. "The $600B AI Question." Partner David Cahn's analysis of the gap between revenue required to sustain AI infrastructure investments and actual revenue being generated.
• IoT Analytics / Gartner Hype Cycle Reports. Market analysis on enterprise generative AI pilots versus production-level monetization.
• Barclays Equity Research. Analysis of the unprecedented surge in hyperscaler capital expenditures skewed toward AI hardware ahead of immediate software demand.

Stephen Klein is Founder & CEO of Curiouser.AI, a Reflective AI company. He teaches "Marketing in the Age of AI" at UC Berkeley Extension and writes about the intersection of capital, institutions, and technology. His forthcoming book, The Rogue Entrepreneur, is under development with Georgetown University Press. Curiouser.AI is the only AI designed to augment human intelligence. Curiouser is community-funded on WeFunder.