token consumption

A metric measuring the volume of data processed by generative artificial intelligence models, where tokens represent pieces of words.

Token consumption is a key operational metric used to measure the utilization and scale of generative artificial intelligence (AI) applications. In natural language processing, a token is a basic unit of text (roughly equivalent to four characters or 0.75 words in English) processed by an AI model during training or inference. For developers and enterprises, tracking this volume is essential for managing API costs and system performance. For investors evaluating AI and software-as-a-service (SaaS) companies, token consumption serves as a proxy for customer engagement, workload volume, and infrastructure demand. High token consumption indicates strong product adoption, though it also directly correlates with the company's underlying compute and cloud infrastructure costs. When analyzing companies with consumption-based pricing models, investors look for rising token consumption as a leading indicator of future revenue. However, because processing tokens requires significant hardware resources, rapid growth in consumption without a corresponding scale in pricing can compress gross margins due to rising compute costs.

An enterprise AI platform charges clients $0.002 per 1,000 tokens processed. If a corporate client processes 500 million tokens in a month, the total token consumption billing is calculated as: (500,000,000 / 1,000) $0.002 = $1,000.