AI customers will eventually move their orders towards TSMC’s 2nm process in droves to target higher inference performance while utilizing less power to reduce the datacenter footprint. This demand surge has already forced the biggest semiconductor manufacturer in the world to make production adjustments for its lucrative customers, with a new report stating that NVIDIA, AMD, and others will push TSMC towards achieving that 100,000 monthly wafer milestone by the end of 2026.

A report from Economic Daily News claims that NVIDIA, AMD, Broadcom, and other customers have enabled TSMC to achieve its goals before its Q4 2026 schedule. Previously, it was mentioned that the semiconductor behemoth’s monthly production would reach up to 175,000 wafers . Now that the target is said to touch 180,000 units, forcing TSMC to make expansion plans rapidly.

However, the company isn’t witnessing any slowdown on the 2nm front, because despite no silicon in sight that leverages this lithography, TSMC has said that its most cutting-edge process has not just contributed to 3 percent of its Q3 2026 revenue , but the company has seen four times as many tapeouts compared to the 3nm process.

The previous 2nm wafer production figure from Fab 20 was 20,000 units , and in just four months, that figure is estimated to touch 100,000 units. While we recommend treating these numbers with a pinch of salt, TSMC can still achieve this milestone through an ultra-aggressive expansion plan.

After all, Apple will largely contribute to that tally from the mobile side, as its A20 Pro , said to be found in the iPhone 18 Pro and iPhone 18 Pro Max , will launch in September, followed by Qualcomm and MediaTek, with their SoCs powering flagships in 2027. As for AMD, its MI455X GPUs will challenge NVIDIA’s Rubin Ultra offering. Seeing the scale at which these datacenter racks will be deployed, TSMC’s 2nm supply could eventually choke.

This reason might explain why Apple reportedly won’t stick with the 2nm process for more than two generations and wants to transition to the 1.4nm node to avoid this supply disparity problem.

News Source: Economic Daily News

Follow Wccftech on Google to get more of our news coverage in your feeds.