This is a review of the financial results from Micron’s peer companies in DRAM and NAND manufacturing, released in March and April of 2022. In this review, I’m only looking for information on the markets for memory products; pricing, supply-demand balance, capacity investment, etc.
July 14, 2022
Bit shipments were down high single digits % and ASPs declined mid-single digits percent. This is the third consecutive quarter of drops in both bit shipments and ASPs. The company reduced their full year 2022 bit shipment forecast from flat to flat-to-down. This helps a little with the overall DRAM supply situation, but Nanya is small. They see weakness in consumer markets, which is affecting PC and mobile demand the most. Management said they believe this is a one-quarter correction. I don’t think this portends any special knowledge. Rather, it is hope winning out over history. The slowdown in memory demand is more from macroeconomic factors than anything else, and those don’t correct in one quarter.
July 28, 2022
Samsung saw bit demand for both DRAM and NAND below expectations and declining sequentially. Surprisingly, ASPs for both memory types were better than the previous quarterly trend. In DRAM, pricing was flat quarter-over-quarter and in NAND pricing increased slightly. These are both following sequential price declines from Q4 to Q1. The company said they will prioritize profitability in DRAM, which means they will adjust near term supply both by holding inventory and by lowering production. In NAND, they signaled they will use their competitive cost position to risk unbalancing the market by lowering prices to stimulate demand. This has been Samsung’s position for several quarters and NAND has held up well, indicating that there is adequate elasticity of demand. The company cited the same market conditions that Micron did in their call. Server demand is holding up, though customers are consuming inventory while consumer product demand (PC and mobile) has declined sharply. They were noncommittal about the second half of the year, a good position to have given their predictions in the call last quarter turned out to be wrong. The current macroeconomic environment is too uncertain to make predictions with much confidence.
July 27, 2022
This was a record quarter for company revenue. Gross and operating margins expanded. DRAM pricing declines have slowed in the last two quarters. Both DRAM and NAND cost declines were larger than ASP changes, leading to the expansion in margins. I wonder if the memory market isn’t a little better than many are giving it credit for. The company did a major down-revision in their forecasts for DRAM and NAND bit demand growth for all of 2022. They took DRAM down from high teens % to low teens %. NAND was reduced from 30% to 20%. Consumer demand in PCs and mobile phones is weakening but the server segment is holding up. They can’t do much about capex this year but signaled they will reduce capex next year. Management again said they will hold inventory, cut investment, etc. to protect DRAM profitability. Their statements about NAND profitability were not as strong, which has been their position for several quarters. They clearly said that the acquisition of Solidigm plus their technology roadmap in NAND will, they believe, enable them to take market share. This is the largest unsolved problem Micron faces today; the fact that they are the last sub-scale NAND maker left.
August 5, 2022
For memory company shareholders, the key takeaway from this report is the further weakening of demand. The PC segment was soft previously and is forecasted to decline 10% in 2022 in unit volume. WD said this quarter that mobile customers are pulling down their demand, following quickly that customers tell them this is a one-quarter inventory correction. I don’t know how customers would know that. Mobile has been weak most of the year so I think it is more likely there is continued softness in demand, not just inventory absorption. Even with the softness, like-for-like NAND ASPs went up for the company. NAND pricing has been flat from six quarter ago to today. While gross margins have increased some in NAND, they have not expanded as much as they should have given that NAND front end cost is targeted to decline 15% per year. I think one reason for this is they are starting up a new fab complex, which is inefficient at this phase of life. Also, they had an excursion earlier this year that lowered output and will hurt their costs during the time that material is coming out. Third, they are seeing higher labor and materials costs caused by the inflationary environment. Since Micron is not impacted by the first two factors, all else equal investors should expect Micron’s NAND business to be seeing increased profitability. The NAND market continues to be healthy, which is the opposite of what I have been predicting, based on statements from both Samsung and Hynix that they intend to grow their market share even if it risks oversupplying the market.
June 30, 2022
One quarter ago things were expected to get better as 2022 progressed. Now they are battening down the hatches. This has been the strangest cycle I’ve seen in more than two decades of owning and following Micron. The stock is down now below $60 per share in anticipation of demand, and thus pricing, weakness. Nobody knows how long the weakness will last or how bad it will be. This quarter was a record for a company and pricing only declined slightly. The low share price is because the overall tech market is well into bear territory, and because when demand declines in memory, nobody knows where it will stop. I won’t even venture a guess as to what will happen. I’ve been wrong too many times since COVID hit and have no reason to think I’ll be more right in the future. If you want to play the memory cycle, this would typically be too early to start buying. There are usually at least four quarters of pain before things recover. But since the Pandemic hit, history has not applied. Micron is taking action, lowering bit growth and contemplating lower fab utilization to reduce supply growth. They will also grow inventory rather than let pricing collapse. Sadana said something that gave me optimism that cycles will dampen in the future, the fact that most big customers have LTAs with the major DRAM makers. This makes it less likely the overall DRAM industry will overbuild and overshoot demand. This won’t prevent cycles, but it should help to dampen them. As for this cycle, I hope it is short and I hope the recovery matches up with a rebound in equity valuations. I’d still like to be out of my Micron position by the end of calendar 2022, though the odds of that are much lower than they were three months ago. I’ll be closely watching what Samsung and Hynix say in their earnings reports in the next couple weeks.
On August 9th, Micron pre-announced lower earnings for their fourth fiscal quarter. During that announcement, they also lowered calendar 2022 bit demand growth for both DRAM and NAND. DRAM will be mid-to-high single digits percent growth. NAND will be low to mid-teens percent growth. Long term growth, they believe, is mid-to-high teens percent for DRAM and 28% for NAND.
The most recent information on the state of the memory markets came out of Micron, in early August. They said that 2022 bit demand will be well below the long-term trend for both DRAM and NAND. For the quarter, pricing was surprisingly good when compared to what is believed to be happening today. Expect ASPs to be significantly down when these companies announce in September and October. Softness in consumer segments, particularly PCs and mobile, is the common theme across all manufacturers. In response, the big three DRAM makers all say they plan to hold inventory rather than sell at prices they believe are below the value of their products. This sets up a game of chicken with memory customers, who have switched to taking down their inventory in hopes of buying memory later at lower prices. Expect a round of capital expenditures cuts to come out in the next couple months with the next set of financial results. It will take at least a couple of quarters for these decisions to flow through to lower supply. Demand returning could end this down-cycle, but given the persistence of inflation and continued COVID-19 issues in China, I don’t believe that will happen until at least sometime in 2023. Supply reductions are the nearest hope to bringing the DRAM market back into balance. The coordinated action of DRAM makers to hold inventory is and will soften the drop in pricing, but they can’t stop it completely. Micron’s stock in the mid-$50s seems to have fully priced the down cycle in at this point, making it historically a good time to buy. If I wasn’t still holding a large long position that is what I’d be doing.
-S. Hughes (long MU)