DeepSeek cut prices on its V4-Pro model by 75%, but many enterprise AI vendors and developers are finding that cheaper models do not automatically translate into healthier margins because agent systems consume tokens faster than prices decline. While a chatbot turns one user question into one model call, an agent turns it into a chain of planning, retrieval, tool use, verification, summarization, and follow-up decisions, creating a "100x problem" where the same user-visible request can cost much more to serve as an agentic workflow. In a single-turn chatbot, the input-to-billed ratio is about 1:5, but in a multi-step agent, that ratio routinely lands at 1:700 or higher. For example, a "simple" agent query like “What did our top customer ask about last week?” typically touches seven priced operations, billing roughly 35,000 input tokens and costing between $0.10 and $0.40. OpenAI's proposed program to give every Y Combinator startup $2 million in API credits reflects what it now costs to run an AI-native company through its first year of product.