AI companies are moving away from traditional per-user licensing fees. They are now charging for the actual work their tools perform, a model described as selling 'units of labor' or 'units of productivity.' This fundamental shift in software economics aims to unlock larger corporate spending pools.
According to a recent Goldman Sachs analysis, this strategy allows firms to tap into new budget allowances. The investment bank's note, based on meetings with about 40 software and internet companies, suggests the change is driven by the need to justify the high costs of building and running AI. Pricing based on delivered value, rather than simple access, is seen as a key to maintaining strong profit margins.
The new model directly ties a company's expense to the productivity or labor output its AI tools generate. This separation of pricing from the underlying computational costs is a deliberate move. It represents a significant departure from the seat-based licensing that has dominated enterprise software for decades.
The shift could reshape how businesses allocate their technology budgets, moving funds from traditional software lines to new AI-specific categories. For software vendors, success will depend on clearly demonstrating the tangible value of each 'unit' of AI-powered work. This transition may accelerate as more companies seek a return on their substantial AI investments.
If widely adopted, the model could create more predictable revenue streams for AI providers. However, it also places greater pressure on them to continuously prove their tools' efficiency and cost-effectiveness to enterprise clients.