Intel’s CEO, Lip-Bu Tan, is reportedly considering a major change in the company’s contract manufacturing strategy.

The potential shift may see Intel move away from marketing its 18A chipmaking technology to external clients.

Tan, who became CEO in March 2025, is considering focusing resources on 14A, Intel’s next-generation chip process, amid delays in 18A.

Tan aims to position the 14A technology as a competitive alternative to TSMC’s N2 technology.

Shifting 18A to internal use only could result in a major financial loss, with analysts estimating a hit of hundreds of millions to billions.

However, Intel has not confirmed these considerations and maintains its commitment to fulfilling customer obligations.

Intel still plans to meet 18A orders for Amazon and Microsoft, while preparing to scale internal Panther Lake chips by late 2025.

Intel’s board will soon review Tan’s proposal as the company seeks to regain competitiveness against TSMC in the global chip race.

🔗 Source: Reuters

🧠 Food for thought

1️⃣ Intel’s second attempt at foundry business faces steeper competition

Intel’s current foundry pivot represents its second major attempt to compete in the contract manufacturing space, with history not on its side.

The company previously shuttered its Custom Foundry business after struggling to attract customers, with industry analysts describing the earlier effort as “ill-conceived” and a distraction from Intel’s core microprocessor manufacturing 1.

Today’s competitive landscape is even more challenging, with TSMC commanding a dominant 64.9% market share in Q3 2024 while Intel’s foundry revenue ($291 million in Q4 2023) remains a fraction of TSMC’s quarterly revenue ($23.53 billion) 2.

Intel’s difficulty scaling its foundry operations stems partly from its historical focus on internal manufacturing processes optimized specifically for its own chips, contrasting with TSMC’s decades of experience serving diverse external customers with differing requirements.

The technical comparison between Intel’s 18A and TSMC’s competing processes further explains Intel’s strategic recalculation, as TSMC’s N2 process achieves superior SRAM density (38 Mb/mm² versus Intel’s 31.8 Mb/mm²) – a critical metric for high-performance computing applications 3.

Recent Intel developments