Software companies are fundamentally shifting their AI pricing strategies, moving away from per-user subscriptions toward models based on the actual work performed. This transition, highlighted in a recent Goldman Sachs note, aims to help technology firms access larger corporate budgets by positioning AI tools as vendors of labor units or productivity gains rather than simple employee licenses. After consulting with approximately 40 companies across the software and internet sectors, Goldman Sachs analysts observed that businesses are increasingly selling value in terms of output. This approach allows vendors to charge based on the utility delivered, effectively separating their profits from the operational costs of running complex AI models. By doing so, companies can maintain strong profit margins while tapping into new budget categories that were previously inaccessible under traditional licensing models. Concrete examples of this trend are already emerging in the market. Salesforce has introduced "agentic work units," while Workday sells credits specifically tied to units of work. These structural changes reflect a broader industry movement away from monthly per-seat licenses toward usage-based, pay-as-you-go pricing, a trend Business Insider reported on in early 2023. The drive for this shift is largely due to the substantial costs associated with building and deploying AI infrastructure. As these expenses rise, vendors must rethink how they monetize their services to ensure sustainability. Consequently, OpenAI CEO Sam Altman previously noted that the industry may eventually evolve to treat AI more like a utility, similar to electricity or water. Altman suggested that business models would resemble selling tokens, the units used by AI systems to process and price data. In this vision, intelligence becomes a metered resource that customers purchase on demand for whatever tasks they need to accomplish. This transition from volume-based to value-based pricing could significantly reshape corporate spending habits. While it offers vendors a path to higher deal sizes, it introduces an element of unpredictability for customers. Expenses will now fluctuate based on usage intensity and workflow complexity rather than remaining fixed at a set monthly rate per employee. As the industry matures, the alignment of pricing with tangible business outcomes is expected to become the standard, driving a more dynamic relationship between AI providers and enterprise clients.