Satya Nadella, CEO of Microsoft, made a keynote statement urging companies to rethink the structure of their assets in the era of artificial intelligence. According to him, businesses need to develop not only human capital but also "token capital"—a collection of AI systems capable of accumulating corporate knowledge and experience.
AI as Part of Intellectual Capital
Nadella divides organizational assets into two key categories. Human capital includes employee knowledge, professional connections, experience, and decision-making skills. Token capital, in turn, refers to AI systems owned by the company that are trained on internal data, business processes, and accumulated expertise. The task of business, according to Microsoft's head, is not to find the most powerful model, but to create its own continuous learning environment that allows preserving and scaling corporate knowledge regardless of the AI solutions used.
The Learning Cycle as a New Asset
Nadella emphasizes that companies must transform workflows, accumulated experience, and professional solutions into agentic AI systems that improve with each use. He calls such systems a new type of intellectual property. Unlike traditional assets, they are capable of constantly accumulating value through the interaction between employees and artificial intelligence. This creates a closed loop where human experience feeds AI, and AI, in turn, scales that experience.
Risk of Monopolization and the Need for an Ecosystem
Nadella paid special attention to the risk of value concentration among a limited number of foundational model developers. He warned that the economy cannot develop sustainably if a few AI platforms benefit from the knowledge of entire industries. Instead, Microsoft advocates for the formation of a broad AI ecosystem where each company retains control over its own data, knowledge, and learning systems. Long-term business success, according to Nadella, will be determined by the ability to simultaneously grow human and token capital, turning employee expertise into scalable AI systems.
Analytical Commentary: The concept of "token capital" is not just a buzzword, but a strategic shift in the perception of AI as a key productive asset. In an environment where foundational models become commodities, real competitive advantage will belong to those companies that can create unique, copy-protected knowledge bases and learning cycles. However, the main challenge here is to ensure the quality and relevance of this data; otherwise, the risk of "toxic" learning (garbage in, garbage out) could negate all investments.