Micron announced today their intent to build a greenfield fab in Clay, NY, north of Syracuse. Over the next eight years the company plans to invest $20B to build a leading-edge DRAM manufacturing. The site could eventually include four cleanrooms of 600,000 square feet each. These are planned numbers and don’t represent a commitment on the part of Micron, but make for a good press release. Vast plans for investment dollars and new jobs are what the politicians that hand out the grants that allow this type of project are paying for. Construction of the site will begin in 2024 with production output ramping in the latter half of the 2020s.
The primary reason Micron chose New York is the state provides massive grants to semiconductor manufacturers, in addition to the CHIPS act from the Federal Government. In this case, the state of NY is granting up to $5.5B in incentives. This is probably a total number, were Micron to build out the site fully with multiple massive cleanrooms. Secondary reasons to the massive grants are the semiconductor ecosystem already present in the area (Global Foundries in Malta) and the strong suite of universities in the region.
As an investor in Micron, this announcement doesn’t change my view of the value of the company. For one, the grants from the state and federal governments mostly just offset the cost of building in the U.S. rather than in another country with better semiconductor ecosystems and lower costs. Two, this is several years from having any effect on Micron’s financial performance. And third, adding this capacity is just for Micron to maintain market share. It doesn’t represent the company growing beyond the rate of the overall memory market. Investing in Micron is best done, in my view, as a cyclical. Buy during times like now when memory ASPs and the stock price are depressed. In a couple years, when the market has gone back into undersupply and ASPs and the stock price have recovered, sell. At that point an investor can either wait to buy back in at the bottom of the next cycle, or go short. It is not easy to time the cycles but the task has a higher probability of success the less concerned one is about selling close to the top.
-S Hughes (long MU)