Naval says open-source AI will not erase the profit moat of frontier labs

Silicon Valley angel investor Naval said open-source AI models will not threaten the profitability of frontier labs, arguing that the most economically valuable parts of the economy remain inherently adversarial and competitive. In his view, fields such as investing, product development, warfare, cybersecurity, and even scientific discovery reward those willing to spend in order to win, rather than those relying on open access alone. His comments came as Kimi K3’s decision to open its weights pushed debate over whether open-source models can approach the frontier into sharper focus. Supporters in the open-source community described the move as another major jump in model scale and capability, and said it highlights the leading position of Chinese labs in open-weight development. At the same time, the release has fueled fresh questions over whether increasingly powerful open-source models could directly challenge the business models and profit potential of closed-source leaders including OpenAI, Anthropic, and Google. Those concerns have centered on whether the traditional advantages of closed systems could narrow quickly.