Nvidia Commits $12.9 Billion to Open AI Models

The New Stack reported that Nvidia is paying $12.9 billion to keep open AI models running on its chips. The article connects the move to a reported Hugging Face acquisition intended to sustain the open-model ecosystem on CUDA. It also notes that Ollama has placed Qwen, DeepSeek and Kimi inside Claude Desktop. The post presents the spending as a step to maintain compatibility between open models and Nvidia's platform.