Workflow
英伟达的泡沫,或许能再吹5万亿美元
虎嗅APP·2025-10-30 13:13

Core Viewpoint - Nvidia's stock price has surpassed $210, making it the first company in history to reach a market capitalization of $5 trillion, raising questions about the sustainability of this valuation and whether the bubble will burst soon [2][3]. Group 1: Market Comparison - Nvidia's market growth is compared to Intel's historical growth from $120 billion to $509 billion between 1996 and 2000, driven by the PC market, while Nvidia's growth is supported by diverse markets including AI, data centers, consumer graphics, and autonomous driving [5]. - Nvidia's revenue compound annual growth rate (CAGR) is projected to exceed 100% from fiscal year 2022 to 2025, contrasting with Intel's 12.6% CAGR from 1996 to 2020 [5]. - Nvidia operates as a fabless company, avoiding the heavy asset burdens of traditional chip manufacturers, which allows for more flexible capital allocation [6]. Group 2: AI Industry Context - The article discusses the potential for Nvidia's business model to be likened to an "energy company" in the AI industry, as it provides essential computational power rather than just infrastructure [10][13]. - Nvidia's recent $100 billion investment proposal to OpenAI for building data centers illustrates its role in the AI ecosystem, where it acts as a provider of computational resources [11][12]. - The AI industry is experiencing a bubble, potentially larger than previous internet bubbles, with companies like OKLO achieving high valuations despite minimal revenue [17]. Group 3: Future Outlook - Nvidia's growth is expected to continue for the next one to two years, driven by technology-driven industries that have yet to fully utilize computational power [18][20]. - The company has positioned itself to support various AI research directions, ensuring a steady demand for its computational resources, even if the commercial viability of these applications remains uncertain [21]. - Concerns about potential computational resource overcapacity are not immediate, as the demand for AI-related computational power is still growing [21].