The era of throwing unlimited computing power at artificial intelligence may be giving way to something more disciplined. According to CNBC, OpenAI and Anthropic are facing a new reality as their customers shift away from what the outlet calls "tokenmaxxing" — loosely, the practice of running AI models as much and as freely as possible — toward squeezing more value out of every dollar spent.
The driver is money. CNBC reports that companies are tightening their AI budgets to focus on getting a return on their investment. Rather than asking how much AI they can deploy, buyers are increasingly asking what they actually get back for the cost.
That change in mindset has direct consequences for the two best-known names in the field. CNBC notes that this shift toward efficiency could dampen growth rates at both OpenAI and Anthropic, whose businesses have ridden a wave of rapid, expansive adoption.
In plain terms, "tokens" are the units of text that AI models read and generate, and they are what customers ultimately pay for. When businesses prioritize efficiency over sheer scale, they consume fewer tokens for the same goal — good for their own bottom lines, but a potential headwind for the providers whose revenue grows with usage.
Why it matters: AI companies have been valued on the assumption of relentless, ever-expanding demand, so a customer-led pivot toward cost discipline could test whether that growth story holds.