Memory Market Analysis:-

1.DRAM and NAND cycles are different.

DRAM and NAND are two distinct types of memory and the cycles in the two memory types are far from perfectly correlated.Memory manufacturers had record earnings recently primarily because of the high DRAM prices. Over the last year, DRAM prices peaked due to supply shortage. For example, Micron to had 71% adjusted gross margin in its fiscal Q4 2018 in the DRAM segment. The economics of NAND remained more stable.

The fabs are optimized for either DRAM or NAND and memory manufacturers can not easily reallocate capacity to one or another. Therefore, the change in economics of DRAM manufacturing would not directly affect the economics of NAND.

2. Secular growth in demand for memory reduces cyclicality of the market.

The memory market became less cyclical. (1) Big-data, artificial intelligence, and cloud services proliferated. (2) These technologies require computing power and data storage, thus creating more demand for memory from data centers. (3) In 2016-2018 3D NAND memory became cheap enough per terabyte to be installed in data centers. (4) As a result, the demand on memory now depends less on consumer devices and PC cycle than it did in previous cycles.

3. Memory market is consolidated, so oversupply is less likely.

The memory market is more consolidated now than during any previous troughs of the cycle. There are only 3 largest manufacturers of DRAM (Samsung, SK Hynix, and Micron) and 5 major manufacturers of NAND (Samsung, SK Hynix, Micron, Toshiba, and Intel) compared to 8 major DRAM manufacturers at the time of previous trough in 2012.

The manufacturers would not allow a serious oversupply this time. Let’s look at the market from a game theory perspective. All key market participants have enough resources to live through a trough and seems to be long-term players in the market. Imagine if one that the manufacturers decides to increase volume and decrease prices to achieve higher short-term profits. This would trigger a similar response from other manufacturers resulting in a potentially prolonged period of oversupply and low prices. During previous troughs it could make sense if one believes that some of the market participants can go out of business, fail to catch up with latest technology due to limited resources, or be acquired by another memory manufacturer at a lower valuation.

Now, all the players have enough resources, positive net cash balances, or diversified operations. Regulators are unlikely to allow any further consolidation. Therefore, equilibrium strategies for memory manufacturers would be to prevent oversupply by increasing the capacity inline with the growth of memory demand.

The memory manufacturers can still have wrong demand projections. But I believe that this time oversupply is less likely and would be less pronounces if it occurs.

4. Mobile device memory demand is likely to increase in 2019.

The demand for memory from mobile device segment is driven by (1) the number of new phones sold and (2) the average volume of memory in a phone.

The base model iPhone XS has 64 GB, the same amount that iPhone X had a year ago. The base model of Samsung Galaxy S9 has 64 GB of memory too, the same as S8 had. It means that the industry standard amount of memory in a phone has not changed in 2018. Additionally, most of customers choose the base models.

It is likely that the next generation would have more memory in the base models, say 128 or 256 GB.

This change would boost the demand for memory from the smartphone segment that represents ~37% of the memory demand in 2018.

.

.

.