Intel’s recent decision to outsource its flagship 2nm production to TSMC raises significant concerns for the U.S. chipmaker. Back in October 2021, then-CEO Pat Gelsinger touted an ambitious journey for Intel, claiming that the company’s transistor sizes would shrink at an unprecedented rate, surpassing Moore’s Law. This principle, articulated by Intel co-founder Gordon Moore, states that transistor counts in chips double approximately every two years. Gelsinger’s statement was intended to demonstrate Intel’s intent to reclaim its leadership in process technology from competitors like TSMC and Samsung Foundry.
However, recent reports indicate a troubling shift for Intel. According to Taiwan Economic Daily, Intel is having TSMC manufacture its Nova Lake CPU using the latter’s 2nm process node. Initially, Intel had projected its 18A process node as a superior alternative to TSMC’s 2nm technology. Yet, by opting to utilize TSMC for its flagship chip production, Intel’s actions contradict its previous claims about the competitiveness of its own processes.
Moreover, this dual-sourcing strategy—using TSMC for flagship chips while relying on its own Intel Foundry Services for non-flagship products—signals a pragmatic but troubling acknowledgement of its position in the market. This comes in the wake of AMD’s decision to become the first TSMC customer for 2nm production for its EPYC processors. TSMC is also set to manufacture Apple’s A20 processors for the upcoming iPhone lineup. Despite continuing to use its 18A node for certain products, Intel’s overall trajectory seems misaligned with Gelsinger’s original vision of aggressively competing with TSMC and Samsung by 2025.
Following a challenging year that saw Intel incur losses of $18.8 billion, the market remains skeptical about the company’s ability to compete in contract chip manufacturing. Intel now faces a critical juncture and may need to reevaluate its strategy for the coming years.