3 6 Ke·2025-10-15 00:27Core Viewpoint - From September 2025, major U.S. tech giants are abandoning their rivalries to collaborate on computing power, forming a strategic alliance centered around AI capabilities, significantly boosting their market valuations [1][4]. Group 1: Strategic Collaborations - OpenAI has placed a $300 billion order for computing power from Oracle, which has a close partnership with Nvidia, indicating a surge in demand for computing chips [1]. - Nvidia announced a $100 billion investment in OpenAI to jointly build large-scale AI data centers, creating a powerful "AI triangle" among OpenAI, Oracle, and Nvidia [1][4]. - AMD has also entered into a strategic agreement with OpenAI to deploy a total of 6 gigawatts of AMD GPU computing power, potentially granting OpenAI nearly 10% equity in AMD [4]. Group 2: Market Dynamics - The AI industry is experiencing a "computing power anxiety," which is a critical bottleneck for development and technological advancement [9][12]. - AI computing power expenditure has increased from 9% of total computing power spending in 2016 to an estimated 25% by 2025, indicating a rapid shift towards AI capabilities [10]. - The high costs associated with computing power are a significant barrier for AI companies, with OpenAI's computing costs reaching $16 billion annually, surpassing its human resources and R&D expenses combined [13][14]. Group 3: Political and Economic Context - The Trump administration's relaxed regulations on large tech companies have facilitated this collaboration among Silicon Valley firms, contrasting with the more interventionist approach of the Biden administration [15][16]. - The Biden administration's focus on market intervention has limited large-scale cooperative projects among tech giants, highlighting the importance of a favorable business environment for industry growth [17][22]. Group 4: Comparison with China - China's AI industry, while rapidly developing, lacks the collaborative spirit seen in the U.S. due to regulatory constraints and a less mature computing power infrastructure [23][26]. - The disparity in computing power supply in China, characterized by both surplus and scarcity, complicates the potential for large-scale collaboration among tech firms [26][28]. - Despite these challenges, China's AI market holds significant potential, and with the right market conditions, it could narrow the gap with U.S. counterparts [29].