Nvidia is investing $26 billion to develop open source AI models aligned with its GPUs, aiming to expand beyond chips and build a powerful developer-driven AI ecosystem.
Nvidia plans to invest $26 billion over the next five years to develop open-source artificial intelligence models, marking a major expansion beyond its traditional chipmaking business into AI software and model development. The development was first reported by Wired and later confirmed by Nvidia executives.
The move is aimed at enabling Nvidia to build, test, and scale next-generation AI models that are closely aligned with its industry-leading GPU hardware. By developing open AI models optimised for its chips, Nvidia could deepen integration between software and hardware while strengthening its developer ecosystem.
Open source AI models are free and accessible to anyone, providing researchers, data scientists, and start-ups with a foundation to modify and build customised AI systems. Many developers already rely on open-source models rather than proprietary systems.
Currently, several leading AI models in the United States remain proprietary and typically require paid subscriptions or cloud access. Examples include GPT-5.4 from OpenAI, Gemini 3.1 Pro from Alphabet, and Claude Opus 4.6 from Anthropic.
Nvidia’s entry into open AI development could also provide a Western alternative to many freely available models that currently originate from China.
The company already maintains a strong developer ecosystem through its CUDA programming platform, which offers more than 400 libraries that help developers build, optimise, deploy, and scale applications across PCs, cloud systems, workstations, and supercomputers.
If Nvidia releases open AI models tuned specifically for its GPUs, it could further strengthen its ecosystem and increase switching costs for developers. The strategy could reinforce Nvidia’s dominance as the world’s most valuable publicly traded company, with a market capitalisation of $4.5 trillion.
