OpenAI Is Taking the “Crack Cocaine” Approach to Pricing
TL;DR
Futurism argues that OpenAI and other AI vendors trained users on cheap or flat-rate access, then are now moving toward usage-based token pricing because compute costs are catching up. The article cites alleged OpenAI losses of 39 billion dollars in 2025, while noting that part of that figure may be accounting-related rather than pure operating burn. The financial pressure still looks hard to ignore.
Nauti's Take
The drug metaphor is harsh and engineered for attention, but the stronger point sits underneath it: AI has been sold to many users as magically cheap infrastructure. That was always a mix of investor money, scale hopes and aggressive growth.
Anyone building workflows blindly around flat rates may be buying a costly surprise later. The sane response is not panic but measurement: which AI tasks actually save time, revenue or error costs, and which are just convenient luxury?
Briefingshow
This is not just an OpenAI pricing story; it is about the business model behind the AI boom. If vendors subsidize usage for years, teams build habits, workflows and dependencies on prices that may not last. For companies, AI shifts from a cheap experiment into a cost risk that belongs in budgets, product margins and governance.
Sources