This is an update to our previous article titled Thinking about DRAM, published on 10th September. DRAM, short for Dynamic random access memory, is a type of semiconductor memory that is typically used for data or program code needed by computer processors to function.
Since then, we have seen further evidence of demand weakness stemming from 2020’s WFH-related surge. This is also compounded by weakness from China’s public Cloud market, as Beijing directed domestic enterprises to align with government-approved Cloud computing providers.
Hence, recent 2Q21 data from the likes of Alibaba, Tencent showed a marked slowdown in CapEx spending. This is understandable since the Chinese titans will need time to lick their wounds.
However… just like there was a risk when the street extrapolated demand strength from WFH into 2021, ATEC thinks it will also prove illogical to extrapolate the current demand weakness into 2022, for the simple reason that the Cloud market in China is but a tenth (by some measure) versus that of the US.
True, ATEC does not know where the demand inflection point is. But, if it is indisputable that ‘data supremacy’ is the next battle frontier between these nations, then ATEC is willing to bet that it will not be long before China steps on the CapEx acceleration pedal.
Since our last update, the poster-child for DRAM, being Hynix (as shown), has declined another 13% or so. Based on ATEC’s framework, the stock/sector is in the last phase of its corrective decline, having re-traced some 76% of its peak price recorded earlier in March ‘21.
ATEC recalls how the street was always pitching that the DRAM sector had consolidated, shows CapEx discipline, registers higher trough operating margins and deserves higher multiples through the cycle. Well, look where we are now. Should we not be looking up instead? What’s missing?
Good luck out there.