The resources required to build frontier AI systems - data, compute, talent, and capital - are concentrated in a remarkably small number of organisations. A handful of US technology companies and a smaller number of Chinese firms account for the vast majority of large-scale AI development. This concentration raises legitimate questions. When a few companies control the most powerful AI systems, they effectively set the terms for how the technology is developed and deployed. Their choices about safety, access, and pricing affect entire industries. At the national level, AI capability is similarly concentrated: the US, China, and to a lesser extent the EU and UK dominate research output, investment, and deployment. Most countries are consumers of AI technology rather than producers, creating dependency relationships that echo earlier patterns in technology and resource extraction. For businesses, concentration means fewer choices and more dependency. If your AI strategy relies heavily on one or two providers, you're exposed to their pricing decisions, policy changes, and geopolitical entanglements. Diversification isn't always practical, but understanding the risks of concentration helps you make more informed decisions about vendor relationships and technology architecture.