Companies are struggling to choose which AI models to use, according to Arena CEO Anastasios Angelopoulos, who said businesses face a “really tricky situation” weighing proprietary models from companies such as OpenAI and Anthropic against open-weight models from Chinese companies including DeepSeek and Moonshot. Frontier AI companies provide cutting-edge models, but access could change, or safeguards may be implemented, whereas Chinese open models can be run locally at a lower cost, but with security concerns and potential government restrictions. “It’s not only true that they’re terrified of working with the frontier labs, but they’re also terrified of working with the Chinese open source,” Angelopoulos said during a recent episode of the “20VC” podcast. He said enterprises are trying to determine where to build their long-term AI infrastructure while dealing with concerns around reliability, cost, data and access. “Anthropic or OpenAI could decide to turn off their AI at any moment if they feel that it is, quote unquote, ‘unsafe,’” Angelopoulos told Business Insider. The Trump administration has also considered restrictions on Chinese AI models, according to an Axios report. Angelopoulos also said some providers had used Arena’s internal data “to serve synthetic battles and arenas to our employees in an effort to essentially get them to label data,” though he did not identify the companies involved.