Government restrictions on access to some of the most advanced AI models are accelerating a shift that was already underway: developers are increasingly embracing open-source and open-weight alternatives. The article argues that reliance on proprietary AI services carries real business risk, as policy changes or provider decisions can suddenly limit access.
What’s especially notable is that competitive open models from companies such as China’s DeepSeek and Zhipu AI are closing the performance gap while offering greater flexibility and lower costs. My view is that this isn’t just an open-versus-closed debate anymore. It’s becoming a question of resilience. Businesses building AI products are likely to diversify across multiple models rather than depend on a single provider, making open-source AI an increasingly important part of the industry’s future, even as governments may eventually seek to regulate those models as well.