TSMC’s dominance in advanced process and packaging has made it a prime target amid US manufacturing mandates. Chip customers now face mounting pressure to diversify supply chains due to cost and capacity constraints, accelerating the shift toward multi-sourcing strategies.

Recent supply chain reports reveal that Nvidia, alongside Apple, plans to collaborate with Intel on its 2028 Feynman architecture platform. Both companies are targeting "low volume, low-tier, non-core" production runs to align with Trump administration directives while preserving their core TSMC relationships. This dual-foundry approach is designed to minimize mass production risks while satisfying political pressures.

Nvidia’s Feynman generation shifts partly to Intel

Following Nvidia’s US$5 billion investment in Intel announced in September 2025, the company intends to partner with Intel on the successor to its Rubin series—the Feynman GPU chip.

The GPU die will remain with TSMC, but portions of the I/O die are expected to leverage Intel’s 18A or the planned 14A process slated for 2028, contingent on yield improvements. Advanced packaging will utilize Intel’s EMIB technology, with Intel handling up to 25% of final packaging and TSMC managing the remaining 75%.

Supply chain insiders note that major American chipmakers have long explored collaboration with Intel under US manufacturing mandates and tariff pressures. However, given current limitations in 18A technology, meaningful cooperation will likely begin with 14A production in 2028.

Given the inherent risks of adopting 14A and 18A processes, most Intel partnerships are starting with EMIB advanced packaging. Intel CEO Lip-Bu Tan recently confirmed that two customers are evaluating 14A specifications, while capital expenditure trends signal when Intel secures firm orders.

Analysts observe that while US manufacturing targets face cost and yield challenges, political considerations, supply chain resilience needs, and TSMC’s constrained advanced packaging capacity are driving American chip firms toward dual-foundry strategies.

Apple taps Intel for entry-level M-series chips

Apple’s discussions with Intel center on an entry-level M-series processor for MacBook models currently produced by TSMC. This marks a notable shift in their relationship, which has evolved considerably over two decades.

Apple relied on Intel x86 processors in Macs starting in 2006, with Intel establishing a dedicated "Apple Group" production line at its Oregon fab in 2005. That partnership lasted until June 2020, when Apple launched its Arm-based Apple Silicon chips, completing a full transition to in-house designs within two years. The move enhanced Apple’s supply chain control and ecosystem integration while sidestepping delays in Intel’s troubled 10nm process development.

The rekindled partnership stems primarily from US manufacturing policies and tariffs under the Trump administration, compounded by cost concerns, risk diversification imperatives, and capacity constraints at existing suppliers.

TSMC’s strategic response

Apple and Nvidia are proceeding cautiously, shifting only low-risk products to Intel initially. Other potential collaborators include Google, Microsoft, AWS, Qualcomm, Broadcom, AMD, Tesla, and entities tied to large US government contracts.

Whether Intel can satisfy tech leaders accustomed to TSMC’s execution standards remains an open question.

For TSMC, the anticipated order diversions to Intel may represent more opportunity than threat. Industry experts identify three strategic advantages: first, reducing monopoly concerns and regulatory scrutiny; second, alleviating US political pressures; third, offloading "non-core" orders to strengthen future pricing and supply negotiations.

This dynamic addresses antitrust concerns surrounding TSMC’s market share while easing demands from the Trump administration. TSMC remains confident in retaining core, high-end chip manufacturing contracts—the most profitable and technologically demanding work.

As customers experiment with alternative foundries, they may ultimately develop a deeper appreciation for TSMC’s capabilities, reinforcing its bargaining power and supply reliability in the long term.

Nvidia, Apple, and others typically refuse to comment on supply chain rumors.

Article edited by Jerry Chen