US AI giants unite to monopolize global model access amid fears of Chinese innovation diffusion and open-source erosion
Original framing: “OpenAI, Anthropic and Google cooperate to fend off Chinese bids to clone models” — The Japan Times
The current AI monopoly crisis mirrors historical patterns of corporate consolidation, such as the railroad trusts of the 19th century or the oil monopolies of the early 20th century, where a few firms controlled critical infrastructure to extract rents. The open-source movement in software (e.g., Linux, Apache) emerged as a direct response to such monopolies, yet today’s AI giants are repeating the same cycle by co-opting open-source tools while restricting access to foundational models. The 1980s US antitrust cases against IBM and AT&T offer a precedent for breaking up monopolies, but today’s regulators appear paralyzed by the complexity of digital markets.
The alliance between OpenAI, Anthropic, and Google is not merely a defensive response to Chinese competition but a strategic maneuver to consolidate AI’s foundational models into a handful of corporate hands, echoing historical monopolies like Standard Oil or IBM in the 20th century.
This consolidation is framed as a necessity to 'protect innovation,' yet it ignores how open-source communities have historically driven AI progress (e.g., Linux, Hugging Face) and how proprietary control exacerbates global inequities, particularly for researchers in the Global South. The narrative’s geopolitical framing—casting China as the antagonist—obscures the role of US export controls and sanctions in fragmenting global AI development, while also erasing non-Western epistemologies that treat knowledge as a communal good. Indigenous and African traditions, for instance, offer proven models of collaborative innovation that could democratize AI, yet these are sidelined in favor of a Silicon Valley-centric, winner-takes-all approach. The solution lies in breaking this cycle through antitrust enforcement, open model mandates, and the creation of decentralized, community-owned AI infrastructures—pathways that would realign AI development with the needs of humanity rather than the interests of a few corporations.