The US AI camp is clearly diverging over the issue of control over intelligence. One side comprises cutting-edge, closed labs like OpenAI and Anthropic: possessing the most advanced models currently available, emphasizing security reviews and strong regulation, and aiming to control the scope of release of cutting-edge capabilities and the pace of intelligent advancement. Critics argue that this security narrative could raise industry barriers and strengthen the control of leading labs.
The other side consists of open-model or sovereign AI-first groups like Meta, Palantir, Microsoft, and Nvidia, advocating for companies to independently control models, data, and intelligent infrastructure, competing in the enterprise market with near-cutting-edge performance, lower costs, and privately deployable technology stacks. Furthermore, whether Google will further shift towards an open ecosystem remains a key variable for the expansion of this camp.
