The April every AI plan broke (thefinancialengineer.substack.com)

April 2023 marked a tumultuous period for the AI/ML community as major players like Anthropic, OpenAI, and GitHub implemented rapid changes in subscription models and pricing structures. Five significant announcements were made in just three weeks, indicating a collective realization that existing plans could no longer accommodate evolving product capabilities and usage patterns. For instance, Anthropic's abrupt decision to restrict third-party tools like OpenClaw from drawing on subscription quotas highlighted the unsustainable economics of their offerings. Users suddenly faced exorbitant costs that far exceeded the original intended usage, prompting backlash and a series of corrective measures. The significance of these developments lies in the broader implications for unit economics in AI services. As observed, the shift from flat-rate pricing to per-token billing reflects an industry-wide transition driven by the need to align revenue models with actual usage patterns. OpenAI and Anthropic's adjustments—such as OpenAI doubling its API prices and Anthropic restructuring enterprise agreements—underscore a critical moment of reckoning in the financial engineering of AI products. This trend indicates a fundamental restructuring of how AI services will be sold and consumed, ultimately impacting developers and businesses reliant on these tools. As companies grapple with the financial realities of running complex AI models, the landscape of AI subscription services is poised for significant change.