Companies question cost of AI as tokenmaxxing spending adds up
Companies are increasingly questioning the high costs associated with artificial intelligence, as spending on practices like "tokenmaxxing" and tokenomics rises. The financial burden of AI investments is prompting businesses to reassess their strategies and evaluate the return on their expenditures.
Background
- **"Tokenmaxxing"** (or "tokenomics") is a sarcastic industry term for the frenzy of spending on AI tokens — the computational units used by large language models like ChatGPT. Companies rack up enormous bills running experiments, fine-tuning models, or simply querying AI services, often without clear ROI.
- This article reports a growing backlash among businesses that piled into AI spending over the past two years. CFOs and tech leaders are now questioning whether the costs of inference (using a model) and training justify the gains, especially as AI-hype stocks have cooled and budgets tighten.
- Major cloud providers (Amazon AWS, Microsoft Azure, Google Cloud) have seen AI-related revenue surge, but some enterprise customers are hitting "bill shock" and starting to curb usage or demand more transparent pricing.
- The piece reflects a broader cycle: after the initial gold-rush phase of any transformative tech (cloud, mobile, now generative AI), the "hangover" period sets in, where companies must separate actual productivity gains from speculative over-investment.