Over the last decade, the production of artificial intelligence (AI) knowledge and technologies has stood out from all other sectors of the economy. In the pursuit of AI evolution, some data and algorithms (i.e., models, the software side of AI) are typically freely available as open source by its authors and sponsors, who value academic impact and scientific reputation. Lately, the open science dynamics of the AI industry has produced a fierce competition for breakthroughs, which depend on firm-specific knowledge—obtained through investments on research and development (R&D)—and expensive technology (respectively the human and hardware sides of AI) to expand the field. While studies on investments in AI have favored the role of nations and governments, the aggregate role of large companies has been anecdotal. This paper tracks global trends in R&D expenditure by key players in the AI knowledge field for 10 years (2012-2021), a decade that has produced AI to transform society. The findings show a highly concentrated industry, suggesting the formation of knowledge and technological clusters (that we relate to a broad concept of oligopoly). Future research in the “periphery” (competing companies and independent or publicly funded research institutions) would be at the service of this oligopoly or would be limited to the resources offered by the oligopoly.