Twenty-five companies, including Nvidia, Meta, and Microsoft, signed an open letter on July 24 advocating for open-source AI models amid Washington pressures to restrict American AI development in the name of national security. Nvidia CEO Jensen Huang supported the letter, stating that open models strengthen safety, accelerate innovation, and enable sovereignty. The signatories addressed concerns about China, where labs like Moonshot AI and DeepSeek are publishing sophisticated open-source models: according to them, restricting American models would simply make Chinese versions the default choice worldwide. Closed companies like OpenAI and Anthropic oppose this approach, arguing that the most powerful systems must remain protected to prevent abuse. Meta, with its Llama model family, and Microsoft, present in both open source via Azure and closed models via OpenAI, illustrate the sector’s divergent strategies.
Source: Read the original article
