TSMC can't keep up with AI chip demand, with shortages projected to last beyond 2027

TSMC's CapEx (Capital Expenditures) for 2026 is projected to hit $56 billion. This capital is being funneled into a three-pronged effort to address rising chip demand: new fabrication plants, increased capacity, and a massive build-out of supporting infrastructure. However, despite this record-breaking spending, TSMC CEO C.C. Wei admits that shortages will likely persist until 2027 and potentially beyond.

The primary driver of this insatiable demand is, as you guessed, AI hardware, including GPUs (Nvidia/AMD), CPUs, and HBM (High Bandwidth Memory). This demand from NVIDIA, AMD, Intel, and even automotive giants is drying up capacity for TSMC's mature 3nm process node family, which, despite already reaching production targets, is still not enough.

The next issue pertains to packaging constraints. Modern AI processors aren't monolithic chips. Instead, they use technologies such as CoWoS (Chip-on-Wafer-on-Substrate) to integrate logic with HBM (High Bandwidth Memory). As it stands, TSMC's advanced packaging capacity is significantly tighter than its front-end wafer fabrication capacity. Another constraint is lead times, as each new fab needs years to reach high-volume production.

It will take two to three years to build a new plant. There are no shortcuts. It will take another year or two to expand production

- C.C. Wei, TSMC CEO

TSMC's solution to this problem includes three new fabrication plants: a new facility in Tainan, Taiwan, slated for volume production in the first half of 2027, followed by a second fab in Arizona, USA, reaching volume production in the second half of 2027, and finally a 3nm plant in Kumamoto, Japan, by 2028.

2

Likewise, TSMC plans to convert existing 5nm production lines into 3nm capacity. This allows them to quickly and cheaply pivot manufacturing since the building and much of the machinery are already paid for. In terms of infrastructure, a large chunk of this CapEx will go toward the aforementioned backend facilities, including the construction of dedicated packaging plants to increase monthly wafer output.

Until now, we have never added that capacity when the node's production capacity reaches its target. But it is currently working on a plan to expand its production capacity globally to meet the powerful multi-year demand pipeline for 3nm technology

Semiconductor manufacturing is not about building physical plants but about establishing a complex, reliable pipeline that can take years to fully mature and stabilize. For consumers, the result is simple: as long as demand outpaces supply at TSMC, the cost of cutting-edge 2nm and 3nm chips, and the devices powered by them, will continue to climb.

This creates an opening for competitors like Samsung Foundry, which has already secured high-profile contracts with Tesla to diversify its supply chain and challenge TSMC's dominance. Intel Foundry (IFS) is also gaining traction, as its upcoming 14A test chips are currently being evaluated by major industry players. As TSMC remains booked, these competitors may be the only viable option for companies unable to wait until 2027 for new silicon.