The rising costs of artificial intelligence (AI) are forcing companies to rethink their strategies as they transition from a subsidized model to one focused on profitability. Initially, AI firms offered low prices to attract customers, but as demand grows, particularly for AI agents that handle complex tasks, expenses have soared. This increase has led to a phenomenon known as 'tokenmaxxing,' where companies overspend on AI usage, sometimes exceeding employee costs. Major corporations, including Meta and Uber, have expressed concerns about the lack of productivity gains from their AI investments. In response to these rising costs, many businesses are exploring alternatives, such as free open-source AI models or specialized smaller models designed for specific sectors like finance or real estate. The cost difference can be substantial, with large models priced at $15 per million tokens compared to just five cents for smaller models. This shift indicates a trend towards treating AI more like a commodity, where the choice of model becomes less critical than finding the most cost-effective solution.