NVIDIA, the global leader in computing power, is undergoing a significant strategic shift in its history. According to the company's latest financial documents, NVIDIA plans to invest approximately $26 billion over the next five years, specifically for the research and training of "open-weight" artificial intelligence models. This investment, one of the largest in the history of the open-source community, marks NVIDIA's deep entry into the core model development sector of the AI industry, moving beyond being just a provider of hardware infrastructure.

The core of this strategy lies in promoting the "open-weight" model, which involves publicly releasing model parameters, allowing global developers to download, modify, and deploy them according to their needs. Through this approach, NVIDIA aims to leverage the diffusion effect of the open-source ecosystem, enabling more AI applications to be built natively on its computing platform, thereby further strengthening its GPU hardware ecosystem moat. Currently, the strategic flagship product Nemotron 3 Super has demonstrated strong performance, with its 128 billion parameter scale outperforming similar open-source models in multiple benchmark tests, showcasing NVIDIA's ambitions in cutting-edge fields such as embodied intelligence, climate simulation, and enterprise-level AI agents.

This move undoubtedly changes the competitive dynamics in the AI industry. For a long time, NVIDIA has supported the expansion of top laboratories like OpenAI and Anthropic by selling high-performance chips, but these major clients now face direct technical challenges from their hardware suppliers. At the same time, NVIDIA's large-scale investment is seen as a strong response to the rise of global open-source model forces, aiming to ensure that the global developer community remains closely tied to its technology stack by providing high-performance open models deeply optimized for its underlying hardware.

As the boundaries between computing power, algorithms, and the developer ecosystem become increasingly blurred, the AI industry has officially entered a new phase of "full-stack platform warfare." NVIDIA's massive bet indicates that the key to success among top AI players in the future will no longer be limited to single model performance, but rather the ability to provide a complete closed-loop capability from the underlying transistors to the top application architecture. In this technological dominance game, NVIDIA is reshaping the global AI model competition landscape by unleashing the "computing power dividend."