'Tokenmaxxing' Fades Amid Concerns Over AI ROI

The practice of 'tokenmaxxing' is fading, say experts, as companies scale back their push for AI use and burn through their annual budgets within months.

Tokenmaxxing refers to companies encouraging employees to aggressively use AI tools, with some tracking token consumption as a proxy for innovation and productivity. However, this approach fell victim to Goodhart's Law, where a metric loses its value once it becomes a target.

A recent report by The Financial Times revealed that Amazon employees generated unnecessary AI workloads to boost usage statistics rather than produce meaningful business outcomes. As a result, Amazon shut down its AI leadership board.

AI researcher Gary Marcus argued that the decline of tokenmaxxing could expose weaknesses in the economics of large language models. He suggested that OpenAI and Anthropic may face increasing pressure as AI models become more commoditized, competition intensifies, and profit margins shrink.