Companies are grappling with the challenge of pricing AI services due to the unpredictable nature of "tokens," the fundamental units used by Large Language Models (LLMs). The rapid evolution of AI technology makes long-term cost models difficult to establish.

• The cost of individual tokens has plummeted, but token consumption is skyrocketing, projected to increase 24 times between 2026 and 2030.
• Key individuals like Simon Gooch at Saviynt and Will Venters, Associate Professor at the London School of Economics, highlight the difficulties in managing and pricing AI costs.
• Companies are experimenting with various pricing strategies, such as flat fees or bundled services, but these can be easily disrupted by changes in LLM providers' pricing.