On August 9, according to a report by The Wall Street Journal, Intel's chairman Frank Yeary attempted to spin off Intel's manufacturing business, even selling it to TSMC earlier this year, but faced strong opposition from CEO Lip-Bu Tan, who was appointed in March.

The report said that due to some directors also supporting Yeary, conflicts arose within the board, ultimately leading to the failure of some of Tan's strategic initiatives. While the Intel board officially supports Tan, his leadership is under increasing pressure both internally and externally.

The Chairman's Attempt to Sell Intel's Manufacturing Business

Earlier this year, there were rumors that Intel planned to split its Intel Foundry Services into an independent company and sell it, with Intel, TSMC, and large fabless chip designers such as Broadcom or NVIDIA jointly holding shares.

Another rumor suggested that Intel would sell all or part of its foundry to TSMC. TSMC would gain operational and strategic control over the production entity and must "fix" Intel's process technology.

However, Intel never confirmed these two plans publicly.

Nevertheless, according to The Wall Street Journal's report, both plans were driven by Intel's chairman, former investment banker Frank Yeary.

Therefore, since becoming CEO of Intel in March 2025, Tan has been in conflict with the chairman and some directors over whether Intel should maintain its semiconductor manufacturing business or completely exit the business.

According to the report, unlike Yeary and other directors, Tan believes that internal chip manufacturing is crucial for Intel's competitiveness and protecting the U.S. supply chain from excessive reliance on foreign companies like TSMC and Samsung.

But are these deals feasible?

TSMC has never considered acquiring Intel's wafer fabrication plants or manufacturing assets for technical and business reasons. This stance has also been publicly stated by TSMC executives.

Technically, TSMC's employees lack experience with Intel's EUV-based process flow, making it difficult for them to help Intel improve its Intel 3 or Intel 18A manufacturing technology.

Moreover, both nodes have already been used by Intel and third-party customers. Transferring TSMC's process technology to Intel's U.S. wafer fabs is complex because, although both companies use ASML's EUV lithography tools and many other tools from manufacturers such as Applied Materials, KLA, and Lam Research, each company's tools, calibration, and vendor-specific settings differ. ASML lithography systems often have custom adjustments or added modules based on each company's process recipes.

Intel and TSMC also use different etching and deposition methods. Even minor changes in materials, such as photoresists or etching chemicals, can lead to yield loss or performance changes. The supplier lists are also different, so materials approved by TSMC may not reliably work on Intel's product lines. Most importantly, even small changes in airflow, temperature stability, or mask handling can alter dimensions and electrical characteristics, making the transfer costly and risky.

As for the business reasons, TSMC has little incentive to help a direct competitor perfect its chip production. It also has little interest in investing billions of dollars in a wafer fab without knowing if it will be fully utilized and profitable. Additionally, Intel's EUV capacity in the U.S. will not be sufficient to meet the needs of both its own and TSMC's customers.

Editor: Xinxizhun - Langkejian

Original article: https://www.toutiao.com/article/7536399502777795114/

Statement: This article represents the views of the author. Please express your opinion below using the [Up/Down] buttons.