You’ve probably noticed it by now. That creeping anxiety every time a tech CEO steps on stage to announce the next phase of the “AI revolution.” They tell you this changes everything. They tell you resistance is obsolete. And if you dare to look at the actual financial numbers, they tell you you’re just a doomer.
But let’s be honest: the math isn’t mathing. The AI boom is being driven less by proven economic value and more by a coordinated hype cycle. Companies, investors, and hardware suppliers all benefit from sustaining the narrative. When the math doesn’t work, the industry doesn’t change the math—it changes the marketing.
Look at the numbers Ed Zitron recently laid bare. We’re told AI is a booming, cash-printing machine. Microsoft, the poster child of the AI era, reported around $34 billion in AI Annual Recurring Revenue (ARR). Sounds incredible, right? Until you look at their capital expenditures. Microsoft’s total depreciation and amortization in a recent quarter was $11 billion—and that’s before we even calculate the massive, unprecedented infrastructure costs being poured into data centers just to keep these models running.
The revenue might look shiny, but the cost to generate it is astronomical. We are pouring hundreds of billions of dollars into a technology whose profit margins are being propped up by amortization tricks and blind faith. Yet, if you point this out, you get dogpiled. The AI consensus demands absolute loyalty.
Supporters dismiss critics as doomers because engaging with the numbers would force a choice: either the AI consensus is based entirely on faith, or the people pushing it are knowingly selling an overhyped promise. Neutrality in the face of a hype cycle isn’t objectivity; it’s complicity.
Here is the terrifying tension nobody in Silicon Valley wants to talk about: The more money poured into AI, the harder it becomes to stop the hype without admitting the entire premise was overvalued. But continuing the hype only enlarges the eventual correction. They are building a plane in mid-air, and they can’t afford to land it because they’ve already sold tickets to Mars.
This isn’t just a Silicon Valley problem. If AI is a bubble, its deflation won’t stay contained in venture capital boardrooms. It will hit every company that has bet its five-year strategy on AI integration. It will hit every worker who has pivoted their career to become a “prompt engineer.” And it will hit every user who was told that adapting to this technology was the only way to survive the future.
We’ve been here before. We were told Web3 was the future. We were told the Metaverse was inevitable. The difference this time is the sheer scale of the capital at risk. You aren’t crazy for questioning the ROI. You’re just one of the few willing to look up from the hype and read the balance sheet.
The tech industry is gaslighting you, insisting ‘this time is different’ while the disclosed financials suggest otherwise. Don’t look away. Demand the numbers. Because when the hype finally buckles under its own weight, the only people who survive will be the ones who refused to buy the lie.
FAQ
Q: Aren't AI revenues actually growing fast? Microsoft reported billions in AI ARR.
A: Yes, but revenue doesn't equal profit. When you spend tens of billions on capex and amortize it against that revenue, the math gets terrifyingly thin. Growth isn't the issue; the unsustainable cost of that growth is.
Q: So what should my company do right now regarding AI?
A: Stop treating AI as a magic bullet. Demand hard ROI metrics before integrating it into your core strategy. If your vendor can't prove the technology saves more than it costs to run, don't buy the hype.
Q: Is this just another 'tech bubble' prediction that will look stupid in 5 years?
A: Maybe. But bubbles don't pop because the technology is useless; they pop because the capital structure collapses. The tech might survive, but the companies betting the farm on unproven economics might not.