The current balance of power in open models

Open-weight AI has become a serious enterprise option, but the center of gravity has shifted: Chinese labs now lead on widely used open models and are closing in on commercial agentic capabilities, while U.S. open models lag further behind frontier closed systems. For CIOs, this means open models can deliver lower-cost, more customizable AI deployments, but vendor strategy, model governance, and performance benchmarking must account for a fast-moving global market where capability leadership and release cadence are changing quickly. IT organizations should expect greater pressure to evaluate open models alongside closed APIs for coding, automation, and internal assistants, while tightening controls around licensing, security, and data residency.