Naval said open-source AI models are unlikely to threaten the profitability of frontier AI labs, arguing that the economy's most valuable activities, including investing, product development, warfare, cybersecurity and even scientific discovery, are fundamentally competitive. In that view, participants either spend to win or risk being outperformed. The remarks come as Kimi K3's open-weight release fuels a fresh debate over whether open-source models can approach the frontier. Supporters in the open-source community see the launch as another major step up in model scale and capability and as further evidence that Chinese labs lead in open-weight frontier AI. At the same time, it has raised questions over whether powerful open models could directly pressure the business models and earnings power of closed-source developers such as OpenAI, Anthropic and Google.