According to reports, spurred by robust demand in artificial intelligence (AI) and high-performance computing (HPC), TSMC is facing a significant shortfall in production capacity for its 2nm and A16 processes, struggling to keep pace with the soaring market needs. NVIDIA, TSMC’s largest client, has been compelled to revise the design of its next-generation Feynman platform in response to these capacity constraints. Meanwhile, Meta has entered the fray, intensifying the competition for available production slots. The current waiting list for TSMC’s production capacity has stretched beyond 2028.
