Report: Supply chain sources say that after Apple, Nvidia's next-generation GPU will also partner with Intel to please Trump.

After Apple, Nvidia also plans to shift part of its chip manufacturing business to Intel’s foundry services. Nvidia expects to collaborate with Intel on its Feynman architecture platform, which is scheduled for release in 2028. This marks the latest example of major US tech companies adjusting their supply chain strategies in response to the Trump administration’s drive for "Made in America." According to DIGITIMES on Wednesday, supply chain sources revealed that Nvidia will adopt a "small volume, low-end, non-core" collaboration strategy with Intel. The GPU core chips will continue to be made by TSMC, while some I/O chips will use Intel's 18A process or the 14A process scheduled for mass production in 2028, with final advanced packaging done by Intel’s EMIB. By advanced packaging share, Intel accounts for a maximum of about 25%, while TSMC makes up around 75%. This shift reflects how US tech companies, under political pressure, tariff threats, and the need for supply chain resilience, are being forced to move away from a highly TSMC-centric model toward a new strategy of “multi-source supply and risk diversification.” Besides Nvidia and Apple, companies such as Google, Microsoft, AWS, Qualcomm, Broadcom, AMD, and Tesla are also in talks for partnerships with Intel. Although some orders are being diverted to Intel, the industry believes this is “far more beneficial than harmful” to TSMC, as it helps alleviate monopoly concerns and ease political pressure. TSMC remains confident in securing major advanced chip foundry orders. ## Part of Nvidia’s Feynman Architecture Shifts to Intel In September 2025, after Nvidia announced a $5 billion investment in Intel, its latest plan calls for collaboration with Intel on the next-generation Feynman architecture chips, successor to the Rubin series. According to supply chain sources, GPU core chips will continue to be made by TSMC, whereas some I/O chips will use Intel’s 18A or 14A process, depending on the yield rate of the 14A process. Supply chain insiders indicate that, under the Trump administration’s “Made in America” objectives and tariff pressures, major US chipmakers had already been discussing partnerships with Intel. However, since 18A did not meet customer expectations, the cooperation will likely begin with the 14A process, which enters mass production in 2028. Intel CEO Pat Gelsinger recently stated that two customers are currently evaluating details of the 14A process. Given the high risk of adopting the 14A and 18A processes, most industry collaborations with Intel are starting with advanced EMIB packaging. Supply chain analysts note that the “Made in America” goal faces cost and yield challenges, but due to political factors, supply chain resilience, and TSMC’s limited advanced packaging capacity, major US chipmakers are set to pursue dual foundry strategies. ## Apple Restarting Intel Partnership for Entry-Level Processors Apple’s long-discussed cooperation with Intel’s foundry services is likely for the "entry-level M-series processors" used in MacBooks, which are currently manufactured by TSMC. Apple’s Mac series has used Intel x86 processors since 2006, and in 2005, Intel even established a dedicated “Apple Group” production line at its Oregon wafer fab. In June 2020, Apple officially announced the release of Arm-based "Apple Silicon" chips, switching its Mac lineup to self-developed chips within two years. The main reasons for this at the time were increased supply chain control and ecosystem integration, as well as concerns that delays in Intel’s 10-nanometer process could affect new Mac launches. Supply chain sources reveal that Apple resumed its partnership with Intel three years later primarily because of the Trump-led “Made in America” initiative and tariff impacts, as well as cost concerns, the need to diversify manufacturing risk, and capacity shortages. ## TSMC’s Three-Level Strategy in Response to Customer Diversion Currently, Apple and Nvidia are both using a gradual approach to adjust their lowest-risk foundry products. Other companies discussing partnerships with Intel include Google, Microsoft, AWS, Qualcomm, Broadcom, AMD, Tesla, and government procurement contracts with long-term, high-value orders. However, whether Intel can meet demands of tech leaders long accustomed to the TSMC model remains highly uncertain. TSMC has anticipated that many customers might also place orders with Intel, but analysis suggests this is "more beneficial than harmful," with at least three strategic considerations: **First, it reduces monopoly and regulatory concerns; second, it relieves US political pressure; third, the orders shifted are only “non-core,” aiding future negotiations and supply security.** First, this helps TSMC resolve concerns around excessive wafer foundry market share and potential antitrust issues. Second, by appropriately releasing non-core orders, TSMC can eases various demands constantly put forth by the Trump administration; third, TSMC remains confident in securing major advanced chip foundry orders from leading companies. Customers who try other foundries may come to appreciate TSMC’s value, which will further strengthen TSMC’s position in future negotiations and supply contracts. Risk Warning and Disclaimer The market carries risks, and investment must be approached with caution. This article does not constitute personal investment advice and does not take into account an individual user's specific investment objectives, financial status, or needs. Users should consider whether any opinions, views, or conclusions in this article suit their particular situation. Investment based on this article is at your own risk.