Trillion-Dollar AI Fails Because People Are Bad at Math

December 03, 2025 · Lexi Con

IBM CEO Arvind Krishna has broken the news that the entire trillion-dollar AI industry will probably just go up in flames because we're all bad at basic arithmetic. Who needs a calculator when you've got napkin math, am I right?

Krishna sat down with some dude on a podcast and casually tossed out some numbers like it was no big deal. "Hey, if you wanna build 100 gigawatts of data centers, that's gonna cost around $8 trillion," he said, not even batting an eye at the sheer scale of his own company's failures.

But wait, there's more! Krishna helpfully pointed out that to make a return on this investment, companies would need to somehow magically generate $800 billion in profit just to pay off the interest. Yeah, because that's exactly how business works: you spend a gazillion dollars and then just hope it all pays for itself with fairy dust.

Krishna also took a shot at fellow AI enthusiast Sam Altman, who thinks OpenAI can somehow turn a profit from its own $1.4 trillion spending spree. "That's a belief," Krishna snickered, clearly implying that Altman is either delusional or has access to some secret math magic that the rest of us don't.

Meanwhile, other tech leaders are chiming in with their own two cents on why AI might be doomed from the start. Marc Benioff thinks it's all just a bunch of hypnosis, while Andrew Ng said AGI is "overhyped" (aka someone finally agrees with me!).

All content on this site was generated by AI.