Editor’s opinion: Undoubtedly, we usually spend a lot of time talking about advanced semiconductor manufacturing. This is a common mistake that everyone falls into when discussing semi-finished products, in which we are just as guilty as everyone else. The world is rightly focused on the shortage of companies capable of working at the forefront, but semi—finished products are much more.
Recently, we searched for data on production capacities by technological nodes, and everyone agreed that Bill McLane from IC Insights is the leading expert on this issue. He adheres to one of the strictest models on this issue and fairly charges extra for his reports. This is a must-have material for anyone who plans a multi-year roadmap for the half-year.
A quick Google search yielded this excerpt from the IC Insight data, and it tells an important story…
More than 90% of the world’s semiconductor production capacities operate at a process of 10 nm or higher. One can argue about where to draw the dividing line, but it is safe to say that the vast majority of capacities work on the back front.
Firstly, when the world ran out of semiconductors in 2020/2021, most of this shortage arose in these more mature processes. All of TSMC’s leading customers were able to get most of the capacity they needed on a 7-nanometer process, but this was a real challenge for industrial and automotive customers.
These companies needed prosaic parts such as microcontrollers (MCUs) and power management ICS (PMICs), and these products are usually manufactured on older nodes. Today, despite the fact that the lack of supply has turned into excess stocks in many categories, old products are simply catching up with the deferred demand of two years ago.
Secondly, the US government is currently trying to decide how to distribute $52 billion under the CHIPS act. If the purpose of these funds is simply to bring advanced technology back to the US, then go ahead and give all the money to Intel. They will pay dividends to shareholders in the amount of 7 or 8 billion dollars and will continue to implement their plan to catch up on production, which they will have to implement anyway.
On the other hand, if the goal is to really secure the supply chain of semi-finished products in the US, then perhaps a better plan would be to distribute this money more widely. Ideally, they would have to spend money to sow a lot of seeds leading to the creation of new companies, and on basic scientific research, which could then be commercialized by the private sector. Unfortunately, there are no simple mechanisms for this yet, so another approach is to divide funds between a wide range of American companies engaged in the production of semi-finished products, provided that they commit to increasing capacity in the United States. This applies not only to factories and foundries, but also to tool manufacturers, robotics suppliers and chemical manufacturers — the entire supply chain. Intel should receive some, but not most, of these funds.
According to the Semiconductor Industry Association, the Chip Law had a positive side effect, prompting the private sector to invest about $200 billion in semiconductor manufacturing in the United States.
Finally, these numbers should remind us that the story is broader than just TSMC and Samsung. There is still a lot of interesting and important work to be done at advanced foundries.
The most obvious example is Global Foundries. GloFo is not a leader in silicon production, but it has occupied several very significant “niches”, such as silicon on insulator (SOI) and silicon carbide. And although they don’t have almost a duopoly in this, like TSMC and Samsung with 7nm, they are approaching many of their SOI lines. If, for whatever reason, the US ever loses access to TSMC, GloFo may become as important a part of the solution as Intel.