硬AI·2026-01-28 08:24Core Viewpoint - Nvidia plans to collaborate with Intel on the Feynman architecture platform expected to launch in 2028, adopting a "low-volume, low-tier, non-core" strategy in this partnership, reflecting a shift in supply chain strategy among US tech giants due to political and supply chain pressures [2][3][6]. Group 1: Nvidia and Intel Collaboration - Nvidia's Feynman architecture will involve collaboration with Intel, with core GPU chips still being manufactured by TSMC, while I/O chips will utilize Intel's 18A or the anticipated 14A process, depending on yield conditions [3][6]. - The collaboration is part of a broader trend among US tech companies to diversify their supply chains and reduce reliance on TSMC, driven by political pressures and supply chain resilience considerations [3][4][10]. Group 2: Impact on TSMC - Despite some orders being diverted to Intel, industry analysts believe this shift will benefit TSMC by alleviating monopoly concerns and political pressures, while TSMC remains confident in securing high-end chip orders [4][10]. - TSMC is expected to maintain a dominant position in high-end chip manufacturing, as the orders moving to Intel are primarily non-core, allowing TSMC to strengthen its bargaining power and supply capabilities in the future [10]. Group 3: Other Companies Involved - Other major companies such as Apple, Google, Microsoft, AWS, Qualcomm, Broadcom, AMD, and Tesla are also in discussions with Intel for potential collaborations, indicating a significant shift in the semiconductor landscape [2][3][8]. - Apple's renewed partnership with Intel for entry-level M-series processors is driven by the need to mitigate manufacturing risks and respond to US manufacturing goals and tariff impacts [8].