3 ms·

2020 has seen record investment in semiconductor manufacturing equipment (+20% to 70 B USD). https://www.scmp.com/tech/tech-trends/article/3129611/us-china-tec

• scotty79 6y ago
  2 years should be enough from what I've read.
  • Bombthecat 6y ago
    Yeah, most of the time i read it is something between 2 or 3 years before a new fab opens.
    • throwaway4good 6y ago
      To me it looks that following this chip shortage there will come a massive supply glut as all this manufacturing capacity comes online. Maybe China and South Korea will be alright, as they can focus on the Chinese market in a world bifurcated by the US-China Tech war, but for the rest of the world, it seems that a supply glut is sure to follow?
      • 11thEarlOfMar 6y ago
        Yes. This pattern has been repeating since the advent of mass-producing semiconductors. Most recently, 2016-2017, a new NAND technology emerged. Referred to as 3D NAND, it greatly increased bit-density and reduced cost of NAND flash. Samsung, Toshiba and to some extent Micron all decided they were going to lead that massive market. Moreover, the 3D NAND required many more layers and therefore more 'equipment per chip'. These three bought new equipment it required in record amounts. You'll see companies like Lam Research with revenue spikes in 2017. 2019, though, not so much. All the capacity came online at the same time, there was oversupply, the price of 3D NAND plummeted, and the chip companies pulled back buying dramatically, sending the equipment suppliers from feast to famine, again. I recall a 2-year period in the late '90s where the San Jose Mercury News headline in August read '1,000 Jobs Open at Lam and Applied' to one year later, 'Lam to Lay Off 1,500'.
        • iancmceachern 6y ago
          I've worked in, and on the periphery of the semi world for much of my career and can second this. Not oy for direct employees but also the huge world of contractors and sub contractors that supply engineering services and equipment to the semi industry. It's very much a boom and bust industry, with huge swings from everyone is busy as could be to everyone is looking for work. It's difficult because many of the skillets and manufacturing processes for these systems are mostly specialized to this industry. Edited- typo.
          • wbc 6y ago
            Thoughts on reducing the cyclical nature of this industry?
            • iancmceachern 6y ago
              The challenge is that to design a new line, or production cell, or process they need a broad range of expertise that is very specialized. For example, they may need folks that are phds/experts in plasma deposition/sputtering/etc. to do a bunch of multiphysics computer simulations to get the operation in the chamber right. Then they need folks with expertise in building the chambers, and associated cryo pumps, then they need robotics experts to architect, design and implement all the automation, chemical engineers for all the crazy chemistry going on, etc. It would be very expensive to keep all those folks on staff, and even harder to keep them all busy. I view it like the construction industry. I worked as a concrete/asphalt inspector for my local municipality in college and the cycles of the road repair industry are similar in my mind. In this world there are lots of "crews" that travel around the country following the work. They start in the south in the winter, and gradually work their way up to the north as the ground thaws and road construction can be done. They usually have a 2 week - 2 month contract in each town and then move on. No one area has enough work to keep them busy for a season. The industry is built on that ability to maximize work/profit and minimize downtime by moving from customer to customer as the work does also. I view it as similar for many of us who do/have worked in semi. When there is a glut of work, it's all good. When there isn't there are hundreds or thousands of highly skilled, highly specialized folks needing work. Those folks/consultancies typically are able to pivot to other industries that need the same skillets readily and easily (like medtech, aerospace, automotive, etc.) and keep working during the slow period for semi. The alternative would be for semi equipment designers/manufacturers (like ASML, etc) to rely more on in house permanent workforces and less on contractors and subcontractors but I fear a fundamental change like that may not be sustainable from a cost standpoint. These companies already have huge swings in their workforce numbers every few years, often hiring or letting go hundreds or thousands at a time. Edited - typo
• whazor 6y ago
  It depends, if you are upgrading machines to produce smaller (5nm instead of 7nm) then you lose some productivity as you get lower yields. But many of the companies are adding totally new semiconductor equipment and not removing the older once. That being said, the massive chip shortage is across the entire spectrum, including the bigger older chips.
  • Kliment 6y ago
    Nobody is scrapping old equipment when they move to a new node. The equipment just gets sold to whoever is one rung down on the pecking order, to keep producing parts at a lower profit margin. Early 2000s processes are still in extensive use.
• emodendroket 6y ago
  Both those countries are trying to compensate for sanctions/export controls though, right?
  • throwaway4good 6y ago
    China certainly is. South Korea, I don't know. The spike in equipment purchase certainly sticks out, maybe they (SK) expect to be able/allowed to supply the Chinese companies, that Taiwan can't/won't?
    • emodendroket 6y ago
      Japan has put pretty serious sanctions on them though, on a (frankly rather dubious) theory that the North Koreans are getting materials from them.