Fierce Competition Among AI Customers as TSMC’s 2nm Chip Powers Next-Gen AI Chips

Reports indicate that TSMC’s 2nm chip manufacturing process is witnessing a surge in demand from major industry players, including NVIDIA, AMD, and various ASIC designers, all vying for available production capacity.

Intense Competition for TSMC’s 2nm Process Capacity

In the current climate fueled by advancements in artificial intelligence (AI), high-performance computing (HPC) clients have accounted for a considerable portion of TSMC’s revenue. As the industry shifts towards next-generation N2-class nodes, the influence of chip manufacturers is anticipated to amplify. Recent insights from a Ctee report reveal that demand from mobile sector clients such as Apple and Qualcomm will spearhead the initial phase of 2nm technology adoption, with AI giants quickly following suit. This transition signifies a critical challenge for TSMC in managing production yields, especially given the growing complexity of chip packages.

Looking ahead, it’s expected that AMD will integrate the 2nm process into its upcoming Instinct MI400 series, targeting a launch in the second half of the year. Meanwhile, NVIDIA plans to unveil its next major upgrade, codenamed Feynman, which will feature the A16 chip at a 1.6nm node. Additionally, prominent ASIC developers, including Amazon and Google, are set to utilize a substantial portion of TSMC’s N2 production capacity for their next-generation architectures, underscoring a significant escalation in the race for premium fabrication processes.

Image Credits: Wccftech

It’s important to note that with each advancement in process technology—from 5nm to 4nm and now to 3nm—TSMC has broken previous records in customer adoption. This trend is closely tied to the pressing need to leverage Moore’s Law, reflecting a paradigm shift where computational power has become paramount. As the leading player in the semiconductor foundry landscape, TSMC stands as the go-to option for most chip manufacturers, given the limited alternatives available.

NVIDIA’s CEO recently emphasized the urgency for TSMC to expand its production capabilities in response to the AI-driven demand surge. He projected that TSMC could potentially “double production”over the next decade. The global buildup of infrastructure presents an overwhelming need that TSMC must address, especially when factoring in the complexities of OSAT packaging services.

Source & Images