Intel May Produce Chips for NVIDIA and Apple by 2028

According to supply chain sources, Intel is steadily emerging as a potential foundry partner for tech giants such as NVIDIA and Apple, with related projects expected to move into mass production around 2028.

For years, TSMC has dominated advanced chip manufacturing thanks to its leading-edge process and packaging technologies. That dominance, however, has also put the company under close scrutiny from U.S. regulators. To ease policy pressure and address capacity constraints, more chip designers are now shifting away from relying on a single foundry toward a "multi-sourcing, risk diversification" strategy.

NVIDIA is reportedly planning to use Intel's 18A or possibly 14A process for parts of its Feynman-architecture chips scheduled for release in 2028, depending on how 14A yields perform. Under the current plan, the GPU compute cores would still be made by TSMC, while the I/O dies would be handled by Intel and integrated using Intel's EMIB advanced packaging technology. In this setup, Intel's share of advanced packaging work is expected to reach about 25%, with TSMC covering the remaining 75%.

Apple is also said to be in talks with Intel about outsourcing production of some entry-level M-series processors. The discussion reflects the long history between the two companies during the x86 era. Although Apple transitioned to its own ARM-based silicon in 2020, factors such as U.S. manufacturing goals and potential tariff pressures are pushing Apple to reconsider selective cooperation with Intel.

Industry observers note that TSMC anticipated this shift and has already adjusted its strategy. By selectively releasing non-core orders, TSMC can reduce regulatory and monopoly risks while easing political pressure—without losing its leadership in high-end chip manufacturing. From TSMC's perspective, the trade-off appears more beneficial than harmful.

Beyond NVIDIA and Apple, companies including Google, Microsoft, AWS, Qualcomm, Broadcom, AMD, and Tesla are also exploring possible collaborations with Intel. Still, whether Intel can consistently meet the strict technical and yield requirements of these customers remains an open question.

NVIDIA, Apple, and other companies mentioned have declined to comment on supply chain–related reports.