TSMC Looks to Build Multibillion-Dollar Chip Plant in Singapore – WSJ
For the Singapore project, TSMC is studying the feasibility of production lines that would make seven- to 28-nanometer chips, another person familiar with the plans said. These chips are based on older production technologies and are widely used in cars, smartphones and other devices. The company is building a $12 billion chip factory in Arizona and expects to receive support from a $50 billion chip-manufacturing incentive plan that Congress is weighing. TSMC is also building a new plant in Japan with financial help from the Japanese government and investment by Sony Group Corp.
The new plant will use the older nodes to manufacture chips, while keeping the “state-of-the-art”) processes in Taiwan itself. I am not surprised at the choice for Singapore. TSMC is betting on the established geopolitical norms and established infrastructure to base its decision for investments. It’s expanding on production in different hubs to derisk itself politically and economically. Interesting times.
Here’s a Linkedin post I came across:
A FABulous reality! | LinkedIn
Fact #1: A semiconductor Fab is a highly capital intensive enterprise, the latest node of which will require approx. $10-20 Billion to setup and more to keep running.
Fact #2: Semiconductor fabrication employs highly complicated technology and processes whose viability and profitability depends on being able to achieve a highly precise combination of feats in physics, chemistry and engineering.
Fact #3: A semiconductor Fab functions along with an ecosystem of raw materials, manufacturing equipment, design, tools, soft & hard IPs, testing and packaging that also need to move in step with technology and process updates.
Building a fab requires an ecosystem to thrive-electronics, automobiles, and now 5G (increasingly relying on the Open-RAN) and software through cloud. If you account for the enterprise 5G deployments, network splicing will add more value to the available spectrum. I expect a surfeit of investments in downstream “ecosystem” around “telemedicine” and services. I am not including the “potential-value” of the “market size”, but these have huge spin-off benefits that are difficult to capture in the final value of the product. Therefore, public investments are warranted.