74% of developers now use AI tools, but the industry is increasingly dependent on proprietary, usage-billed platforms from Anthropic, Google, Microsoft, and OpenAI. Gartner predicts AI coding token spend could exceed developer salaries by 2028. Industry voices argue that open infrastructure — open-source models, open protocols like MCP, open orchestration layers, and Kubernetes-based deployments — is essential to avoid vendor lock-in, fragility, and economic concentration. Parallels are drawn to Linux, Docker, and the open internet as evidence that open foundations tend to win long-term. Initiatives like the Agentic AI Foundation, MCP, and llm-d donation to CNCF signal growing momentum toward openness, though gaps remain in governance and API standardization.
