Loding Loading ...
X
Century is regulated by the Capital Market Authority. CFDs are leveraged products that incur a high level of risk. Know more

Tuesday, March 10, 2026

The Next Wave of the AI Boom: Memory Chips

By Century Financial in 'Investment Insights'

The Next Wave of the AI Boom: Memory Chips
Memory Flyer

What's Actually Happening to Memory Stocks Right Now?

Think of HBM (High Bandwidth Memory) like the working desk space of an AI model. The bigger and more complex the task, the more desk space you need. Google's TurboQuant claims it can shrink the desk — and the market panicked, selling Samsung, SK Hynix, Micron, etc., fearing AI will need less memory going forward.

The sell-off is a classic "sell first, read later" reaction. Here's why the market is likely wrong:

The benchmarks are inflated. TurboQuant's 6x memory reduction is measured against old, inefficient baselines — like claiming a car saves 50% fuel versus a 1970s engine. Real AI labs already use far leaner systems (FP8/FP4 precision), so the actual saving over what's deployed today is much smaller.

Efficiency has always grown the market, not shrunk it. Every time AI got more efficient — cheaper tokens, longer context windows, better models — users consumed more of it, not less. DeepSeek proved this exact point in January 2025.

Memory is already the bottleneck. Memory has become "THE bottleneck" for AI and next-generation CPU builds, with shortages intensifying and customers now prepaying for large volume deals, investing — not the behaviour of an industry expecting demand to fall.

The Bull Case for Memory Stocks

1. Structural AI bottleneck, not cyclical. With DRAM slack now gone, indications that it is a true bottleneck investing — meaning supply is already stretched before any new wave of AI model deployment.

2. Efficiency = more usage, not less (Jevons Paradox). Every major efficiency improvement in AI history — faster chips, better compression, cheaper inference — has resulted in greater aggregate demand. TurboQuant, if it works at all at the margins it claims, will enable longer context windows that users and enterprises will immediately fill. This is the single strongest structural argument.

3. TurboQuant is old news overhyped. The paper was on arXiv a year ago. Every frontier lab has already been compressing KV cache as aggressively as possible. This is not a new risk — the market is discovering something technologists already knew and priced in at the model level.

4. Customers are prepaying — a leading indicator. When hyperscalers lock in large volume deals with prepayment, that signals multi-year confidence in the supply scarcity. You don't prepay for a commodity you expect to get cheaper or more abundant.

5. Geopolitical supply risk is a floor under prices. Helium supply concerns for semiconductors alongside Iran war disruptions — adding a supply-side risk premium that further tightens the market for Asian memory producers.

Bottom line: The market is confusing "efficiency improvement" with "demand destruction." In AI infrastructure, those have historically been opposites. The selloff in Samsung, SK Hynix, and Micron looks like a knee jerk reaction to a year-old paper.

Major Beneficiaries of the Memory Trend

Investors can get exposure to the memory theme by investing in the individual stock picks or via the iShares MSCI South Korea ETF (EWY), which gives a total exposure of 46% to Samsung and SK Hynix. These two companies are the top memory chip producers in the world.

Stock Snapshot

Name Sector 52 Week Low *Last Price 52 Week High Market Capitalization (Billions) Discount from 52 Week High Analyst Target Price
Micron Technology Inc Semiconductors $61.54 $370.30 $455.50 $416.78 -19% $408.81
Sandisk Corp Computer Hardware $27.89 $527.33 $725.00 $80.89 -27% $747.73
Western Digital Corp Computer Hardware $28.83 $245.25 $309.90 $83.15 -21% $326.56
ASML Holding NV Production Technology Equipment $578.51 $1,292.80 $1,547.22 $501.80 -16% $1,712.66
Lam Research Corp Production Technology Equipment $56.32 $199.33 $256.68 $248.92 -22% $276.42
KLA Corp Production Technology Equipment $551.33 $1,344.55 $1,693.35 $176.24 -21% $1,707.69
Applied Materials Inc Production Technology Equipment $123.74 $324.74 $395.95 $257.72 -18% $413.36
Name Price Appreciation Total Analyst Recommendations Recommendation
Consensus
(Out of 5)
Beta 30 D volume ISIN Number
Buy Hold Sell
Micron Technology Inc 10% 47 5 1 4.68 2.58 35,834,336 US5951121038
Sandisk Corp 42% 16 7 1 4.21 2.53 21,373,444 US80004C2008
Western Digital Corp 33% 23 6 0 4.59 2.25 10,726,246 US9581021055
ASML Holding NV 32% 19 2 1 4.55 1.57 1,824,132 USN070592100
Lam Research Corp 39% 28 10 0 4.42 2.11 11,416,894 US5128073062
KLA Corp 27% 19 11 2 4.03 1.96 1,162,621 US4824801009
Applied Materials Inc 27% 35 7 0 4.62 1.83 8,207,308 US0382221051

Source: Bloomberg
Last Price: 31st March 2026

ETF Snapshot

Ticker Issuer Name Inception Date 52 W
Low
Last
Price
52 W
High
NAV % Premium/
Discount
Total Assets
(Billions)
Expense
Ratio (%)
30 Days
Average Volume
Beta ISIN
EWY US iShares MSCI
South Korea ETF
12/05/2000 $48.5 $123.01 $154.2 $120.31 2.25 $15.60 0.6 16,956,310 0.8 US4642867729

Source: Bloomberg
Last Price: 31st March 2026

iShares MSCI South Korea ETF (EWY)

Over the past year, the South Korean index KOSPI has rallied 105.41%. The iShares MSCI South Korea ETF is a proxy for investing in the South Korean market. The rally has been driven mainly by the Memory Supercycle, which is expected to continue into 2026, supporting further growth in the South Korean market. The major constituents of the ETF include Samsung Electronics, which accounts for 27.15%, followed by SK Hynix at 18.74%, Hyundai at 2.91%, and others.

South Korea is well positioned to lead in 2026, supported by strong earnings growth and a favorable momentum profile. The main driver behind the rally is the expectation that South Korea will deliver close to 50% EPS growth, which stands out compared to other emerging markets. This growth is largely driven by the AI supercycle and the recovery in the semiconductor industry. The AI industry is currently facing shortages in NAND, DRAM and high-bandwidth memory (HBM) chips due to strong demand. As AI model sizes continue to expand, rising demand for higher memory content and bandwidth could drive a 24% annual increase in HBM content per AI chip. Semiconductor export prices jumped by more than 40% by late 2025, and this trend is expected to continue into 2026. Major companies such as Samsung and SK Hynix have seen strong upward revisions to their 2026 earnings. The ongoing shortage of high-bandwidth memory has further lifted expectations for both companies, placing South Korea ahead of other emerging markets.

Despite the strong rally, South Korean equities still appear inexpensive. Forward earnings multiples are around 9.3x to 9.5x, which is the lowest among major emerging markets, even after the strong performance in 2025. In addition, the MSCI Korea Index trades below its five-year historical average. This valuation gap, which is often referred to as the “Korea Discount,” presents a huge opportunity for investors, especially as the government focuses on reforms aimed at unlocking greater shareholder value. Despite concerns about a potential AI bubble, earnings trends for regional chipmakers remain strong, and expanding partnerships between Korean chipmakers and Nvidia continue to support a positive outlook.

Lastly, Korean conglomerates have accelerated share buybacks, totaling $10.9 billion in the third quarter of 2025. Samsung Electronics announced a 10 trillion won buyback plan, alongside other companies, highlighting a clear commitment to narrowing the valuation gap.

South Korea’s central role in the global AI supply chain continues to attract strong foreign investor interest and presents an excellent opportunity for investors.

Stock Descriptions:

Micron Technology Inc:
Micron Technology is an industry leader in innovative memory and storage solutions. Micron has a 20-25% share in the highly consolidated and commodity-like DRAM market that plays a crucial role in processor and system performance. This, combined with increasing unit growth, should drive double-digit bit gains through at least the next five years. Given the demand-supply dynamics in the HBM market, Micron's HBM revenue could more than double in fiscal 2026 to $12.4 billion, up from $5.4 billion a year ago. Micron's cyclical upturn is extending beyond historical precedent as AI and HBM demand and muted industry capacity increases in recent years are driving a structural undersupply that can sustain memory-pricing growth and strong margins in 2026.

Western Digital Corp:
Western Digital is well positioned to capitalize on surging hard-disk drive demand as cloud-capacity investments and AI-driven data growth accelerate. The shift toward larger-capacity drives allows Western Digital to capture more exabytes per unit shipped, strengthening its positioning with hyperscale customers. The company's build-to-order strategy has led to unprecedented sales visibility, supply constraints and higher pricing, resulting in margin expansion and stronger profitability.

Sandisk Corp:
Tight supply and high demand are driving an elongated NAND upcycle for Sandisk that we expect can last through 2026, given limited industry appetite for capacity expansion and strong AI inferencing demand. Pricing tailwinds should drive meaningful margin growth that can exceed previous cyclical peaks, assuming that NAND suppliers remain disciplined about technology upgrades.

ASML Holding NV:
ASML, a leading lithography-tool maker, is expected to report strong sales growth until around 2030. ASML has a 95% market share in lithography tools, a fundamental process in chipmaking. It is also the only commercial vendor of lithography machines that use EUV, an advanced technology which is essential to making 3- and 2-nm logic chips. With the evolution of 1-nm technology, the EUV market will grow significantly as the next-generation system, high-numerical aperture EUV, is introduced.

KLA Corp:
Advanced chips' increasing structural complexity is powering demand for patterned wafer-inspection tools, and that helps sales and profit at KLA, which is the global leader in this field. KLA's process-control tools are set to play important roles in the evolution of chip-shrinking technology and 3D structuring, boosting its sales toward 2033. As semiconductor wiring scales down to sub-1-nm technology from 3-nm, the need for precise measurement of circuit patterns at all stages could grow strongly. KLA could achieve or even exceed 40% operating profit margin in the long term due to its overwhelming market share in chip-defect inspection and metrology machines.

Lam Research Corp:
Lam Research's sales might expand faster than peers as its chipmaking tools could be more used by increasing numbers of both memory-chip and logic-chip customers. The company, whose etching and deposition tools are indispensable for making advanced NAND memory chips with 300 or more layers, could keep a dominant market share in 3D NAND tools. Lam Research could add to sales as the AI server market expands, since AI servers tend to use 8 times as many DRAMs, including HBM chips, as general-use servers. Its dominant share in electroplating tools used in the TSV process for HBMs might be another growth driver.

Applied Materials Inc:
Applied Materials could have a revenue share of 20% in the global chipmaking-tool market in 2033, rising from 19% in 2024. The company has deep know-how in equipment for multiple processes that will be used in next-generation chip production. These include 2-nm technology and beyond, which could differentiate it from competitors. The company could achieve a high return on equity of 40-50% due to its large market share. Applied Materials’ sales could also grow faster than most peers due to solid demand for tools to make mature-node chips for internet-of-things, communications, automotive, power and sensor applications. Its service business can achieve solid secular sales and operating-profit growth due to its substantial base of installed equipment.

Risks and Assumptions related to Back-tested trading strategies
The risks and assumptions listed here are not intended to be an exhaustive summary of all the risks and assumptions involved.
The strategy might suffer from look-ahead bias which occurs due to the use of information or data in a study or simulation that would not have been known or available during the period being analyzed. This can lead to inaccurate results in the study or simulation.
Future price movements may not be exactly the same as the historical price movements and this could lead to variation in performance.
Testing can sometimes lead to over-optimization. This is a condition where performance results are tuned so high to the past they are no longer as accurate in the future.
The model assumes no slippages in trading. Slippage refers to the difference between the expected price of a trade and the price at which the trade is actually executed.
The back-tested strategy might be at risk of data dredging, which is the behavior of testing multiple hypotheses at one time, resulting in picking the data that best supports your main hypothesis.
Drawdowns in actual trading can be higher than the tested system and losses could be significant in the event of leverage.
Unforeseen events can lead to variation in performance from the tested trading strategy.
The tested result has been computed with price feeds available from Bloomberg.
The testing environment has not considered transaction or any other costs.
Trading indicators used for the purpose of testing has been provided by Bloomberg.
The strategy might suffer from data mining fallacy, selection bias and backfill bias.
A trading strategy that performs well on multiple datasets from one market (e.g., forex) might not perform as well in another market (e.g., stocks).
The strategy may not depict accuracy in terms of spread changes due to the spread-widening events.

Disclaimer:Century Financial Consultancy LLC (CFC) is licensed and regulated by the Capital Market Authority (CMA) of the UAE under license numbers 20200000028 and 301044 to carry out the activities of Financial Products dealer, Trading Broker in international markets, Trading Broker of OTC derivatives and currencies in the spot market, Introduction, Financial Consultations, and Promotion. CFC is incorporated under UAE law, registered with the Dubai Economic Department (No. 768189), with its office at 601, Level 6, Building No. 4, Emaar Square, Downtown Dubai, UAE, PO Box 65777.

Terms and Conditions of Access
By accessing and continuing to use the Publication (which includes this document, flyer, charts, diagrams, illustrations, images, calculations, scenario analysis, and related data or content), you confirm that you have read, understood, and agreed to the terms of this Disclaimer.
CFC reserves the right to amend or update the Publication and this Disclaimer at any time without prior notice. Continued use following any such update constitutes your acceptance of the revised terms. If you do not agree with these terms, please discontinue use of the Publication.

Purpose and Intended Use
This Publication is classified as marketing material and should not be regarded as independent investment research. It is provided for informational, educational, and illustrative purposes only and does not constitute investment advice, a recommendation, an offer, or a solicitation to buy or sell any financial instruments or services. All views expressed are general market commentary and may not reflect the opinions of CFC as a whole.

Risk Disclosures and Limitations
The information presented does not cover all the risks associated with the products or scenarios discussed. Please refer to the full Risk Disclosure Statement available on our website.
This Publication reflects information available at the time of preparation and does not account for subsequent developments. Any forward-looking statements involve assumptions and uncertainties; actual outcomes may differ materially. CFC does not guarantee the accuracy, completeness, or reliability of the information and disclaims liability for any action taken based on it.

No Offer or Contractual Commitment
No part of this Publication constitutes an offer, agreement, or commitment to enter into any transaction. Distribution of this Publication does not oblige CFC to engage in any trade or provide any services. Product names or terms may differ across platforms or providers. This material should not be interpreted as legal, regulatory, tax, accounting, or credit advice. Recipients should seek independent professional advice and assess their own financial situation, objectives, and risk profile before making investment decisions.

Data Sources and Interpretation
This Publication may rely on publicly available data, third-party information, or model-based assumptions. CFC makes no representation or warranty as to their accuracy or completeness. Data limitations, errors, or outdated inputs may impact the reliability of projections or scenarios. Names of financial products may differ from those used on trading platforms.

Use, Reproduction, and Analyst Disclosure
This Publication is intended solely for the recipient’s informational use. It may not be copied, transmitted, or distributed in any form, wholly or partially, without prior written permission from CFC.

Analyst Declaration: The Analyst(s) certifies that all opinions expressed in this Publication represent their own independent views and that reasonable care was taken to ensure objectivity. They do not hold securities in the companies mentioned, and their compensation is not linked to the views expressed. CFC’s research and marketing divisions operate independently.

Trading Risk Warning:
Trading in financial products involves significant risk. Leveraged OTC derivatives, such as Contracts for Difference (CFDs) and spot forex contracts, carry a high risk of loss that can potentially exceed initial deposits and may not be suitable for all investors. These instruments do not confer ownership of underlying assets. Investors must carefully evaluate their investment objectives and risk tolerance, and consult independent advisors where appropriate.