Semiconductor Memory Industry: Key Insights

Summary
- Structure of memory industry has changed.
- China is the big unknown to the future of the industry.
- The supply would slow owing to technological constraints.
In this article, I would try to develop key insights pertaining to the memory and storage industry of the world. Along with a focus towards the future emerging trends.
According to the MicKinsey, during 1996 and 2012, the memory players were unable to create economic value, despite contributing significantly to the overall semiconductor industry through technological innovations. This was due to following reasons:
- Intense competition among the rivals.
- Commoditization of DRAM and NAND flash.
However, recent structural changes such as:
i) High barriers to entry,
ii) Consolidation in DRAM segment owing to a diversified demand for products, iii) lower capacity increments, and
iv) Slowdown in Moore’s law have swelled the margins of the sector. It should be noted that the capacity of memory bit is determined by the capacity of memory wafer and the number of bits per wafer.
Although the capacity of memory wafer has increased by ~86 percent in the last seven years, I anticipate slower technological progress in process node shrinkage due to the factors mentioned below. Notwithstanding Moore’s law, the International Technology Roadmap for Semiconductors has also stated that the process miniaturization has slowed.
Presently, the industry expects its future growth to come from IoT (Internet of Things). Although its inflection point has not happened, the main challenge to the industry originates from the complexity of products and processes. This limits the industry's ability to respond to the current pace of innovation.
Demand Drivers
Demand for memory and storage products would come from the following factors.
- Increasing graphics memory in games and the prominence of gaming as a mass sport, particularly in emerging countries. Such trends would inevitably create a demand for a higher bandwidth and density.
- Average memory content in server units is raising the demand for server DRAM and NAND flash. Google, Amazon, and Microsoft are setting up more data centers for their services. Moreover, automobiles (ADAS and autonomous driving) together with smartphones have also witnessed an increase in average memory and storage content per unit.
- Rising SSD (Solid State Drive) demand together with higher content explosion have given rise to the insatiable demand for memory and storage products. As per SEMI, the demand for non-volatile memory would grow from ~$35 billion in 2016 to ~$80.3 billion till 2025.
IoT the future driver of growth
As per McKinsey Global Institute, the Internet of Things ((IoT)) could generate from $4 trillion to $11 trillion in value globally by 2025. However, it is to be noted that only 10 percent value is in “things” or devices, while 90 percent of the value resides in other elements of the value chain. For instance, retailers might get approximately over a $1 billion in value from IoT, but the device makers can only get approximately $100 million in value. Therefore, semiconductor companies need to offer a comprehensive set of solutions rather than concentrating solely on hardware. For example, the company can offer the suite of cloud computing solutions in addition to the server hardware.
Increase in complexity and processes
3D die stacking would involve more complexity and process steps per square inch of wafers. Such sophisticated manufacturing involves more purity chemicals coupled with more filtration steps. The increase in filtration steps is due to higher points of filtration.
Further, the production process of 3D NAND is more disposition and etch-intensive. In addition to that, such processes also lead to more etch and disposition steps, which in turn require more gas filters. Therefore, more services of Micro-contamination and material handling are needed.
Higher investment per unit of wafer capacity
Materials account for almost 30 percent of the cost of the die, having a greater impact on the semiconductor performance. Transition into new process node would require more advanced and process materials. For instance, vertical NAND production involves new dielectrics metals and polishes. 3D NAND manufacturing needs 3x to 4x more tools for conversion per unit of wafer capacity. Further, Atomic-layer deposition ((ALD)) and atomic-layer etching ((ALE)) require new equipment features.
The development cost of extreme ultraviolet ((EV)) technology amounts to more than $4-$5 billion. It is worth noting that the cost of leading edge node rose from $1.7 billion in 2001 to $10 billion in 2010.
HDD’s future and the emergence of SSD
The future of the HDD in the storage industry lay in the ability of HDD manufacturers to keep their cost advantage with respect to SSDs. The investors must be wary of the continued positive development in NAND technology in terms of cost and performance. The customers of both notebook and enterprise applications are increasingly switching to SSD.
The demand for all-flash array has surged due to the complex caching and tiering system associated with hybrid arrays. Further, AFA-enabled data reduction (including some vendor-guaranteed reductions of 3:1 or more) is significantly reducing acquisition costs.
The introduction of newly developed 32TB SSDs has offset the capacity advantage of HDDs
China’s ambitions for a semiconductor industry
China’s plans to build a plant that is expected to manufacture 3D-NAND and DRAM memory chips by investing more than $30 billion. This plant would have the initial monthly capacity of 100,000 wafers. It usually takes on average three years to build a functional fab facility, thus, after 2020, the industry could witness substantial supply coming from China.
Moreover, Chinese acquisition of Toshiba’s semiconductor business, the second largest flash memory manufacturer, could initiate a capex war with Samsung (the largest player in the semiconductor memory business). If this happens, then the other small players such as SK Hynix, INTC, WDC, and MU could witness a steady erosion in their market share coupled with shrinking margins.
China accounts for ~45 percent of the worldwide demand for chips. Earlier, bureaucratic-led initiatives were unable to produce desired results. However, Chinese endeavors are different due to the following reasons:
- The investment is 40 times more than the previous time.
- Greater focus on creating national champions, particularly through mergers and acquisitions.
- Market-based approach rather than bureaucratic-led effort.
Conclusion
Key insights drawn from the above content in the article are:
- China’s ambitions can be disruptive to supply and demand dynamics of the memory-chip industry.
- The progression towards lower process node would take more than two years
- The industry would develop a more cooperative model, where cross-licensing and technology exchange agreements would become the characteristics of the market. However, the emergence of Chinese players could unleash fierce competition in the industry, but that is a far cry from today as industry’s developmental cycle are very long.
- The future of HDD is very bleak, given advances in SSD technology.
- Rising capital and technological investments for memory industry would ensure fewer players dominating the market, particularly in DRAM segment.
This article was written by
Analyst’s Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Seeking Alpha's Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.
Recommended For You
Comments (7)

