Reply to thread

Message

[QUOTE="Munyaradzi Mafaro, post: 58466, member: 636"]
NVIDIA breaks from its traditional manufacturing strategy by securing the position as TSMC's inaugural A16 node customer. The graphics chip maker previously relied on mature semiconductor processes rather than pursuing the latest fabrication technologies. Apple, MediaTek, and Qualcomm typically claimed first access to TSMC's newest production nodes. This shift represents a departure from NVIDIA's approach since the 110nm generation. The company historically focused on architectural improvements rather than cutting-edge manufacturing processes.

The A16 process represents a significant technological advancement for TSMC with Gate-All-Around FETs and Super Power Rail backside power delivery systems. TSMC expects to begin high-volume manufacturing by late 2026. NVIDIA may implement the A16 technology between late 2027 and early 2028. The company's Feynman GPU lineup appears more likely than Rubin Ultra to debut this process node. This partnership is expected to generate substantial revenue for TSMC from AI hardware manufacturers seeking advanced semiconductor capabilities.
[/QUOTE]

Name