My AI bill just went way up

I’m a technologist with a software studio. AI is in nearly everything my agency builds and most of how we work. But when our Anthropic bill recently skyrocketed, I was happy about it. The price we pay for AI is a lie. The longer the lie continues, the more dangerous AI is to the future of human expertise and human teams. The sooner businesses start paying up, the better. THE PRICE IS FICTION Let’s start with the math. In 2025, Open AI booked around $13 billion in revenue and posted a roughly $21 billion operating loss, spending close to $1.60 for every dollar it took in. That isn’t an accident or a rough patch. The frontier labs are deliberately pricing below what it costs to serve you, to win the market before the market figures out what it’s worth. Venture capital and the cloud giants are covering the gap. They will, eventually, want it back. Most analysts expect API prices to climb once the subsidy race winds down. So when AI feels absurdly cheap, that’s because it is—temporarily, and on someone else’s dime. Here’s why I’m glad the bill is coming. A subsidized price lies to you about one of the most pressing calls we’re all making right now: human or machine? At today’s fake-low rate, the machine wins arguments it has no business winning. Move the price to where it actually sits, and a lot of those “obviously cheaper to automate” decisions quietly flip back to the human side. PERMANENT DECISIONS BASED ON TEMPORARY PRICES What worries me isn’t AI getting expensive. It’s what we’ll do before it does. We’re being invited to make permanent decisions based on a temporary price. If you clear out your writers, your support team, and your analysts now, while AI is on sale, you don’t just trim a line item. You lose the people, the relationships, the institutional memory, and the human judgment that AI frankly still can’t touch. You can re-sign a software contract. You can’t always rehire a relationship. We’ve seen this movie enough times to know how it ends. Amazon undercu