| More insight from me. First, the structure. TSMC will have tool owners who manage the tool on their side. The tool suppliers will have engineers who maintain the tools from their side. A tool is a piece of equipment involved in the manufacturing process. This relationship is interesting. The tool owner's performance is heavily dependent upon their ability to maximize availability. This mainly involves managing relationships with the suppliers and fighting for more resources from the supplier's engineers to fix the tools. Supplier sign availability contracts to provide maintenance for the tools. There is a lot of pressure and politics involved. Second is hiring. Onboarding is a big problem. Training a college grad from scratch requires 1.5 years, a lot of domain knowledge. It is a huge investment to fly them around the world to different training sites and to give them opportunities to practice on the tools. Lastly is culture. TSMC pushes their guys and their tools to their limits. They work in high pressure environments and are expected to proactively step up. In Taiwan, the managers often yell at their workers and the workers have little room to negotiate. TSMC tool owners know a lot more about how the tools work than their competitors, sometimes even more than their suppliers. TSMC tool owners often step in and do things themselves. TSMC will also not be able to hire the same quality of talent as in Taiwan. TSMC in Taiwan pays more and has higher prestige. This is not the same in Arizona. They will also have trouble convincing "target school" talent to relocate to Arizona. Now what does this amount to? First, the supplier's engineers will have options to work for Intel and Samsung and will probability prefer to work at those sites. Intel has the best work culture. Quality, safety, etc. This will force TSMC to make changes. Second, TSMC will have trouble keeping talent if they push phds to work as tool owners, as the job market provides better opportunities in the US. This will result in the new fab to staffed with more inexperienced workers. This is very problematic because domain knowledge from experience in this field is key. The inexperience will result in availability issues (5-10 percent) affect production. So what should TSMC do? What TSMC must adapt to the local culture and lower their expectations. They must be able to convince experienced engineers to work at their site. They can also push this to their supplier's side in terms of availability contracts. They need to relocated some experienced process engineers and tool owners from Taiwan or the throughput and yield will drop. If they do hire new talent, they have a great opportunity to break industry norms. This industry is filled with bad practices from older engineers. These habits are incredibly difficult to break. If they train new engineers correctly, they can outperform the older engineers. End of the day, the fab will come up, it will run. The question is how many chips will they be able to produce in the first few years. |