OpenAI:我们需要100亿个GPU

Core Viewpoint - The collaboration between OpenAI and Nvidia is seen as a significant step towards a future where artificial intelligence (AI) operates continuously, with the need for approximately 10 billion GPUs to support this vision, surpassing the current global population of 8.2 billion [1][2][3] Group 1 - OpenAI President Greg Brockman envisions a future where AI acts as a dedicated agent working for individuals while they sleep, requiring each person to have their own dedicated GPU [2] - Sam Altman, CEO of OpenAI, compares the scale of their collaboration with Nvidia to the Apollo program, suggesting it is even more ambitious [1] - Brockman emphasizes that the AI industry is currently "three orders of magnitude" behind in computational power, indicating a significant gap that needs to be addressed to achieve the goal of always-on AI [2] Group 2 - The world is moving towards a "computational scarcity" state, where AI data center services may become a form of currency, reflecting a shift in economic dynamics [3] - The geopolitical implications of Nvidia's GPUs are highlighted, as they have become a focal point in US-China relations, suggesting that this trend will continue [3]