저자: 미상 · 외국인
요약 보기
마이클 베리의 엔비디아·마이크론 숏 확대 소식이 반도체 메모리주 하락을 촉발. 마이크론 FQ3 2026 매출 415억달러(+345%), 매출총이익률 72.6%, FQ4 가이던스 500억달러. 베리는 순환적 자금조달 수요·감가상각 연장·중국 메모리 경쟁을 우려하나, 분석가들은 HBM 수급 타이트·장기 take-or-pay 계약을 근거로 강세. 밸류에이션은 선행PER 약 12배로 역사적 저평가. 영상은 명확한 방향 판단 없이 'sell/hold/buy' 판단은 시청자 몫이라고 정리.
다듬은 글
Good day to you everyone. Welcome back to the channel. Today, we will talk about why Micron and SanDisk stocks are dipping again as famous investor Michael Burry doubles down on his short positions against Nvidia and the semiconductor sector. In late July 2026, public disclosure filings from Michael Burry revealed a major expansion in his bearish wages across artificial intelligence hardware companies.
Burry increased his put options against Nvidia at $210 and added to his short position on Micron at $933 while maintaining bearish positions in other major tech names. This news triggered immediate pressure across memory chip makers with Micron stock falling nearly 7% back below $900 and SanDisk dropping over 9% in a single trading session.
Burry's core argument centers on what he describes as a circular demand arrangement in artificial intelligence infrastructure. He contends that a significant portion of current chip revenue is financed off the balance sheet rather than driven by genuine end-user monetization. Furthermore, major cloud hyperscalers like Microsoft and Google have extended their server depreciation schedules from 4 years to 6 years, which may obscure the true capital costs of these chip purchases.
Added to these concerns is the recent market debut of Chinese memory manufacturer Changxin Memory Technologies, which highlights growing competition in global memory supply. However, sell-side analysts point out that memory fundamentals remain exceptionally strong with Micron guiding for record revenues and high bandwidth memory capacity fully booked through next year.
This stark divergence between hedge fund short sellers and underlying company cash flows has created intense debate among investors evaluating whether to sell, hold, or buy these memory stocks. Hey guys, quick pause. If you have not realized yet, this person you are seeing here is a digital clone of me.
I am making these videos in hope to bring a new perspective and analysis on stocks and markets in general. If you like what you are hearing, do consider giving me a follow as it helps bring this video to more people. To properly evaluate whether this market sell-off is justified, we must examine the actual revenue numbers and forward growth trajectories for these memory companies.
In fiscal third quarter 2026, Micron reported quarterly revenue of $41.5 billion. This represents a staggering 345% increase compared to the same quarter in 2025. This historic top-line growth has been powered almost entirely by insatiable demand for high-bandwidth memory, known as HBM, which is paired directly with artificial intelligence accelerators.
SanDisk, operating in the NAND flash storage sector, delivered quarterly revenue of nearly $6 billion, marking a solid recovery from the cyclical trough experienced 2 years ago. Looking ahead to fiscal fourth quarter 2026, Micron management provided revenue guidance of $50 billion plus or minus $1 billion.
If Micron achieves this target, full year 2026 revenue will exceed $160 billion, setting setting an all-time record for the company. Consensus estimates from Wall Street analysts project that Micron will maintain a compound annual revenue growth rate of over 30% through 2027. This projected growth rate significantly exceeds the broader semiconductor industry average, which is forecasted to expand at roughly 12% annually over the same period.
SanDisk is also forecasted to deliver revenue growth above 20% as enterprise solid state drive demand accelerates across data centers. The central economic debate lies in whether this revenue growth represents a permanent structural shift or a classic semiconductor inventory spike. In previous memory cycles, rapid revenue expansion was invariably followed by severe oversupply and sharp top-line contractions.
During the 2022 downturn, for example, memory revenues across the industry dropped by nearly 50% in less than four quarters. Skeptics like Michael Burry view the current revenue trajectory as nearing its cyclical peak, expecting demand to normalize as cloud providers digest their massive hardware purchases.
Bulls, on the other hand, argue that the unique requirements of artificial intelligence training create a multi-year supercycle that differs fundamentally from past consumer electronics cycles. Understanding the broader industry context is essential when analyzing why hedge fund managers are taking opposing sides on semiconductor stocks.
The memory market historically operates as a commodity industry, where dynamic random access memory and NAND flash prices fluctuate based on global supply and demand balances. However, the arrival of high bandwidth memory has introduced structural changes to this traditional market dynamic. High-bandwidth memory chips are custom-engineered, stacked vertically, and integrated directly into artificial intelligence chip sets using advanced packaging technologies.
Because designing and validating these memory stacks requires extensive co-engineering with chip designers, switching costs for customers are substantially higher than in standard memory products. To capitalize on this dynamic, the major memory producers, Micron, SK Hynix, and Samsung, have shifted significant clean room capacity away from standard DRAM toward high-bandwidth memory manufacturing.
This capacity allocation has constrained the supply of conventional memory chips, providing price support across the broader memory ecosystem, including NAND flash products sold by SanDisk. Another critical strategic development in 2026 is the adoption of multi-year take-or-pay supply agreements. Micron has secured long-term contracts covering more than 50% of its anticipated revenue over the next 3 years.
These agreements lock in minimum purchase volumes and price floors, giving memory manufacturers unprecedented revenue visibility. Despite these positive structural shifts, macro risks remain prominent in investor calculations. Michael Burry's short thesis highlights potential accounting distortions among major cloud hyperscalers, including Microsoft, Alphabet, and Amazon.
These companies represent approximately 50% of total data center chip demand. By extending the useful life of server equipment from 4 years to 6 years, hyperscalers reduce annual depreciation expense on their income statements, artificially boosting reported operating profits. Bury argues that if end user monetization of artificial intelligence applications lags behind capital expenditure, hyperscalers will eventually be forced to curb their infrastructure spending.
Additionally, the competitive landscape is evolving rapidly in Asia. The successful public listing of Changxin Memory Technologies and China's progress in developing domestic deep ultraviolet lithography tools indicate that Chinese memory capacity could expand faster than anticipated. An influx of lower cost Chinese memory capacity could create headwind pressures in standard DRAM markets over the coming years.
A rigorous economic analysis requires looking beyond top-line revenue to evaluate cash flow quality and profit margins. Operating profit margins offer the clearest picture of a company's structural pricing power during different phases of the economic cycle. In fiscal third quarter 2026, Micron generated a non-GAAP gross margin of 72.6% This represents an extraordinary surge from the negative gross margins recorded during the bottom of the memory downturn in 2023.
Micron's operating margin reached 65.7% while its net profit margin stood at 55.9% Non-GAAP diluted earnings per share came in at $25 and 11 cents for the quarter compared to under $2 just 18 months prior. Free cash flow generation has followed a similar upward trajectory. In fiscal 2026, Micron is on track to generate more free cash flow in a single year than in all previous years of its operating history combined.
SanDisk has also demonstrated substantial margin recovery with gross margins rising to 35% and net profit margins expanding to 20%. Operating cash flow relative to sales for Micron currently sits at nearly 60%, a level typically reserved for software companies with minimal physical capital requirements.
In financial research, operating cash flow to sales exceeding 20% is considered exceptional. However, high profit margins in cyclical industries carry an inherent structural vulnerability. Historically, peak gross margins in semiconductor manufacturing incentivize management teams across the industry to expand capital spending.
As new manufacturing capacity comes online, total industry output eventually exceeds market demand, causing unit prices to collapse and margins to compress sharply. During the 2022 memory contraction, Micron's gross margin plummeted from over 45% to -16% in less than six quarters. Bury's bearish positioning rests on the belief that current margin levels are unsustainable and reflect the absolute peak of profitability for this cycle.
Conversely, proponents of the stock argue that high bandwidth memory yields remain structurally lower than standard DRAM, preventing rapid supply flooding and supporting elevated margin levels over an extended time frame. Valuation metrics provide the final piece of data required to determine whether current share prices present an opportunity or a hazard.
To evaluate semiconductor stocks objectively, financial analysts rely primarily on forward price-to-earnings multiples and forward price-to-operating cash flow ratios. Following the recent price drop to around $873, Micron trades at approximately 12 times its consensus forward earnings per share for fiscal 2027.
For context, Micron's 10-year median forward price-to-earnings ratio is roughly 15 times, while the broader S&P 500 index trades near 22 times forward earnings. On a price-to-operating cash flow basis, Micron is valued at roughly eight times forward operating cash flow. This multiple is significantly below the semiconductor sector median of 21 times forward operating cash flow.
SanDisk trades at an even lower valuation multiple of approximately 10 times forward earnings, reflecting investor caution regarding NAND flash pricing durability. To perform a conservative intrinsic value calculation, we can apply a scenario model. If we assume that Micron's earnings per share experience a 30% cyclical decline from peak 2026 levels down to $18 per share, and apply a conservative historical earnings multiple of 12 times, the resulting valuation estimate sits near $720.
Conversely, if we assume earnings stabilize near current consensus expectations of $30 per share, and trade at a normal 15 times multiple, the intrinsic value calculation yields $450 above current trading levels. The market is currently pricing Micron as if a severe cyclical downturn is guaranteed to occur within the next 12 months.
Investors holding long positions are essentially taking the stance that market expectations are overly pessimistic given structural demand for artificial intelligence hardware. Investors taking short positions aligned with Michael Burry are betting that earnings will fall far faster than consensus models anticipate.
In my personal portfolio framework, buying decisions are driven by fundamental margin of safety rather than attempting to time market momentum. In summary, the decision to sell, hold, or buy Micron and SanDisk comes down to how one views the durability of artificial intelligence infrastructure spending.
Michael Burry's short position provides a valid warning about circular demand risks, hyperscaler depreciation stretches, and historical memory cycle volatility. At the same time, Micron's record free cash flow generation, low forward valuation multiples, and multi-year supply contracts present a compelling long-term fundamental case.
Maintaining disciplined position sizing and conducting thorough independent research remains essential when navigating volatile semiconductor markets. This analysis reflects purely my personal portfolio framework and data calculations, so everyone should make their own informed decisions. Hey, thanks for watching.
I truly respect every viewer who takes the time to seek objective, data-driven analysis.
Comments 0