美光重挫23%触发抄底潮供给潮与周期均值回归压制长线回报

浑水摸鱼清源

美光当期虽录得87%毛利率与极低市盈率，但长鑫存储等竞对扩产与传统DRAM份额流失，预示其未来三至五年毛利率将不可避免地均值回归至常态区间。

AI-assisted summary of 【The Motley Fool】半导体崩盘，美光MU暴跌23%，还能买吗？ — original source. Check the original for full context.

Source published: 2026-10-17

Key ideas

美光（MU）股价自历史高点回撤23%至每股约92.20美元，前瞻市盈率降至仅约5倍；最新季度营收达54.00亿美元（同比激增380%），毛利率与营业利润率分别飙升至87%与81%的历史峰值。 (0:00)
高带宽内存（HBM）的晶圆消耗量约为传统DRAM的3倍，引发消费级存储供应严重紧缺与价格暴涨500%至600%，但这为长鑫存储（CXMT）等外部竞争者填补常规DRAM市场空白提供了战略窗口。 (1:11)
2027年末至2029年全球将迎来三星、SK海力士、美光及中国厂商的大规模产能集中释放，美光与日韩三巨头的寡头定价默契面临瓦解，87%的超高毛利率长期极可能面临均值回归至30%至40%。 (10:40)
美光年内已录得45次单日波动超5%的极端交易日，展现极强动量属性；尽管短期业绩受下游科技巨头强劲AI开支与股票回购支撑，但资本市场已开始提前18至24个月定价远期供需拐点。 (10:00)

Detailed notes

## Background 本内容由Bilibili博主“浑水摸鱼清源”于2026年10月17日发布。视频基于The Motley Fool的投研分析，聚焦半导体存储龙头美光科技（MU）在自高点暴跌23%、前瞻市盈率跌至约5倍背景下，深入探讨当前属于低估值抄底良机还是大周期见顶信号，并结合供给侧剧变解析其多空逻辑与操作策略。

## TL;DR 节目围绕美光科技股价回撤23%后的多空分歧展开全方位剖析。尽管美光凭借AI热潮实现单季营收暴增380%、毛利率飙升至历史罕见的87%，并在极低估值下拥有强劲的做多动能与回购预期，但行业基本面正面临严峻挑战。随着高带宽内存（HBM）挤占产能推高常规DRAM价格，长鑫存储等新入局者正趁机切入，打破美光、三星与SK海力士以往的默契寡头格局；加上2027至2029年全球巨量新增产能集中释放，存储行业的强周期商品属性与均值回归风险，将使美光中长期的超额盈利与估值承受巨大下行压力。

## Atmosphere & style 整体氛围理性克制、警醒深刻，充满专业华尔街投研对话的冷静审视与辩证思维。分析师们并未被当下炸裂的财务数字冲昏头脑，而是在赞叹惊人毛利率的同时，以冷峻的口吻层层剥离繁荣表象，直击周期律与全球竞争格局演变的残酷本质。收听过程中既能感受到动量交易与暴利数字带来的短期脉冲刺激，又会随着时间轴推移油然而生一种对周期顶部反噬的警惕感与敬畏心。

## Bulletin - [00:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=0s) 美光（MU）股价自历史高点回撤23%至每股约92.20美元，前瞻市盈率降至仅约5倍；最新季度营收达54.00亿美元（同比激增380%），毛利率与营业利润率分别飙升至87%与81%的历史峰值。 - [01:11](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=71s) 高带宽内存（HBM）的晶圆消耗量约为传统DRAM的3倍，引发消费级存储供应严重紧缺与价格暴涨500%至600%，但这为长鑫存储（CXMT）等外部竞争者填补常规DRAM市场空白提供了战略窗口。 - [10:40](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=640s) 2027年末至2029年全球将迎来三星、SK海力士、美光及中国厂商的大规模产能集中释放，美光与日韩三巨头的寡头定价默契面临瓦解，87%的超高毛利率长期极可能面临均值回归至30%至40%。 - [10:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=600s) 美光年内已录得45次单日波动超5%的极端交易日，展现极强动量属性；尽管短期业绩受下游科技巨头强劲AI开支与股票回购支撑，但资本市场已开始提前18至24个月定价远期供需拐点。

## Key takeaways - [00:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=0s) 美光股价较高位回落23%至92.20美元附近，前瞻市盈率回落至5倍左右，在AI需求爆发的大背景下引发散户左侧抄底热潮。 - [00:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=0s) 中国存储龙头长鑫存储（CXMT）公开宣布今明两年大幅扩充DRAM产能，正在逐步打破存储市场的原有供求平衡预期。 - [00:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=0s) 2026至2027年初全球新增供给有限，但2027年末至2029年美光、三星、SK海力士及中国厂商的大批新产能将集中入场。 - [01:11](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=71s) 制造HBM需消耗约3倍于传统DRAM的晶圆资源，对常规晶圆形成强烈虹吸效应，导致手机和PC常规芯片供应短缺、成本飙升500%至600%。 - [01:11](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=71s) 美光与日韩巨头全力倾斜资源生产高溢价HBM所腾出的传统DRAM空缺，正被长鑫存储等厂商迅速填补并可能借此打入全球供应链。 - [01:11](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=71s) 面对高昂的元器件采购成本，苹果和英伟达等下游巨头具备在2027至2028年重新设计硬件架构或转向第三方供应商的动机，美光面临长期客户流失风险。 - [05:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=300s) 美光单季毛利率高达87.00%、营业利润率达81.00%，相较2023年11月季度的负毛利率创下历史极值，反映本轮AI行情的极度繁荣。 - [05:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=300s) 存储芯片本质上仍属于高周期性、商品化市场，超高利润率必然刺激全行业盲目扩产，毛利率大概率在未来3至5年回归至30%至40%的常态。 - [07:34](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=454s) 多头逻辑立足于强劲的即期基本面：最新季度营收同比暴增380%至54.00亿美元，下季指引达61.00亿美元，年内即将重启低估值股票回购。 - [10:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=600s) 半导体厂商在周期繁荣期大幅拉升资本开支（CapEx），但大举扩张的产能最终将转化为远期巨量供给，推动行业走向下行周期。 - [10:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=600s) 美光今年已录得45次单日涨跌幅超5%的剧烈震荡，属于典型的高波动动量标的，而非适合防御型长期持有的平稳资产。 - [10:40](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=640s) 美光、三星与SK海力士以往形成的“默契控产”三寡头格局正在被打破，中国厂商愿意以较低毛利抢夺普通DRAM市场份额。 - [10:40](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=640s) 美光在全力追逐HBM暴利的同时，存在永久丧失传统大宗DRAM市场份额的战略风险，重蹈昔日电子硬件代工转移的历史覆辙。 - [11:55](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=715s) 华尔街机构普遍提前18至24个月定价周期见顶，美光当下的逢低买入更适合短期动量交易，长期持有面临显著的均值回归风险。

## Detailed walkthrough ### [00:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=0s) 暴跌23%与供给端变局：长鑫存储与巨头产能扩张隐忧

美光科技（Micron）股价自近期高点已回落 23%，目前的交易价格大约在 92.20 美元附近。面对仅约 5 倍的前瞻市盈率（Forward P/E），市场上许多投资者将其视为难得的低估值“抄底”良机。然而，讨论的核心在于：这究竟是一个绝佳的买入点，还是美光本轮大周期的顶部已然出现？尽管美光过去一年多因全球数据中心对 AI 算力近乎“无限的需求”而暴涨超过 500%，但供给端的结构性变化正在重塑整个行业预期。

一方面，中国本土存储大厂长鑫存储（CXMT）公开宣布，不仅在今年，更将在明年大幅提升存储产品的产能。虽然长鑫存储目前未必能在所有高端领域与巨头展开直接竞争，但存储市场是一个整体蛋糕，任何边缘份额的流失都会引发市场对供需平衡的重新定价。针对这一潜在冲击，Lou 指出：

> “If there is a whole other player that it's at least competing for part of this, that has to change the economics of the business, period.”

另一方面，国际巨头也绝非坐以待毙。在 2026 年第三季度乃至 2027 年初，市场上确实极少有大规模新增产能投产；但一旦步入 2027 年末至 2028、2029 年，美光自身、SK 海力士（SK Hynix）、三星电子（Samsung）以及一到两家中国存储供应商的大量新产能将集中释放。届时，困扰行业数年的“供不应求”方程式恐将被彻底改写。

```mermaid flowchart TD accTitle: 存储行业供需与产能释放动态演变 accDescr: 展示近期供需格局与中长期大批新增产能上线对产业链的冲击机制。

subgraph NearTerm ["近期格局（2026 - 2027年初）"] A["传统晶圆产能有限"] --> B["晶圆大规模倾斜至 HBM"] B --> C["传统 DRAM 严重供给短缺"] C --> D["存储产品价格暴涨 500% - 600%"] end

subgraph LongTerm ["远期格局（2027年末 - 2029年）"] E["长鑫存储 (CXMT) 扩产"] --> H["全球供给过剩风险"] F["日韩巨头 (三星/SK海力士)"] --> H G["美光自建产能释放"] --> H H --> I["均值回归：毛利率回落至 30%-40%"] end ```

- 股价自高点回撤 23%，前瞻市盈率回落至 5 倍左右。 - 过去一年受益于 AI 算力狂潮，美光股价涨幅曾超过 500%。 - 2026 年至 2027 年初供给增量极为有限，但 2027 年底至 2029 年将迎来全球性产能释放高峰。

---

[01:11](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=71s) HBM 虹吸效应、DRAM 缺口与大客户的架构替代风险

从供给技术细节来看，高带宽内存（HBM）的生产正在严重挤占常规 DRAM 的晶圆供给。简单来说，制造 HBM 需要消耗约三倍于传统 DRAM 的晶圆资源。正是这种对基础晶圆的大量虹吸，直接导致全球智能手机（如 iPhone）和电脑的常规芯片供给短缺、终端设备价格普遍攀升。

然而，美光与日韩巨头全力转向生产利润更丰厚的 HBM 所留下的常规 DRAM 市场空白，正给外部竞争对手创造进入良机。长鑫存储等厂商正在快速填补这一空缺，且这些产品不仅会在中国本土消化，长远来看极有可能打入全球供应链体系。正如节目中警示的那样，如果苹果（Apple）等核心大客户因为供应链短缺而与第三方建立起替代采购关系，这部分核心业务未来恐怕再也无法回流：

> “If Apple sets up relationships and is able to source with third parties, I don't think there's a given that that business ever comes back.”

这意味着，美光股东如今不仅是在享受 AI 带来的爆发式红利，更是将整个公司的命运单一押注在 AI 之上。与此同时，需求端的能动性不容小觑。虽然苹果、英伟达（NVIDIA）无法在短短几个月内更改产品硬件规范，但面对暴涨 500% 至 600% 的存储芯片成本，各大科技巨头具备在 2027、2028 年重新配置处理架构的动力与手段，届时硬件工程师很可能会设法减少高昂 HBM 的用量或采用常规 DRAM 替代方案。

> “Again, the higher the price, the more the scarcity, the more you are incentivizing smart people to put their brains to work in solving the problem that is your margin.”

- HBM 晶圆消耗量约为传统 DRAM 的 3 倍，引发全行业消费级存储供应紧张。 - 部分内存元器件采购成本已狂飙 500.00% 至 600.00%。 - 硬件设计存在数年滞后期：2026 年确立的研发调整将在 2027 至 2029 年的产品迭代中集中显现。

---

[05:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=300s) 87% 毛利率的历史极值与商品化大周期的“均值回归”

在 AI 数据中心建设的推动下，美光的利润率表现达到了令人咋舌的历史极值。回顾 2023 年 11 月季度，美光的毛利率甚至还是负数；而如今，公司季度毛利率（Gross Margin）已飙升至 87.00%，营业利润率（Operating Margin）亦高达 81.00%——这意味着美光每产生 1 美元营收，就能斩获 87 美分毛利和 81 美分营业利润。将数据纵向回溯至 2015 年，这种盈利水平在整个半导体存储行业都是绝无仅有的。

面对这种近乎暴利的财务报表，主持人们直指其背后潜藏的周期律风险。半导体存储在本质上依然是一个高度“商品化”（Commoditized）的市场，历史经验表明，超高毛利必然会招致资本无节制的扩产，进而引发均值回归（Mean Reversion）。长期来看，毛利率重回 20%、30% 或 40% 的常态是大概率事件。Lou 对此给出了明确的冷思考：

> “For me to believe that numbers that they have never been able to consistently generate in the past are the new normal just because they got there, that is probably the least likely scenario in my head.”

在此背景下，对于紧盯长线供需平衡的投资者而言，美光当前更接近于一段基于业绩爆发的“动量交易”（Momentum Trade），而非可以高枕无忧的超长线买入并持有标的。

---

[07:34](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=454s) 多空大博弈：天量回购与 380% 营收爆发对决供给大顶

讨论随后深入到了针对美光的多空逻辑对比。多头的核心底牌在于无可匹敌的当期基本面表现（“Scoreboard”）。在最新财报中，美光单季度营收达到了 54.00 亿美元，同比暴增 380.00%；而公司给出的下一季度指引更为激进，预计营收将攀升至 61.00 亿美元，大幅击败了华尔街原先预期的 57.00 亿美元。此外，受此前美国政府芯片法案（CHIPS Act）相关资助条款的约束，美光在今年晚些时候才被允许恢复股票回购，而在 5 倍市盈率的低估值区间推行大规模回购，将对每股收益产生惊人的提振作用。Lou 总结了多头的核心心态：

> “With margins like that, you can buy back your shares. So the bull case for now is, may this last.”

从需求方来看，Alphabet、亚马逊（Amazon）、英伟达等科技巨头在 AI 算力集群的投资上依旧不断加码，消费电子端对于更大容量内存的需求也在同步增加。在未来 12 个月内，全球几乎没有显著的新增供应能够入场竞争，美光的业绩顺风期仍有望维持。

然而，空头的逻辑同样坚硬且不可忽视：繁荣的表象之下，全球竞争版图正在被重构。除了三星和 SK 海力士紧随其后的技术追赶，来自中国存储芯片厂商的实质性扩产，正在让市场重新认识到全球供给版图绝非坚不可摧。一旦这股竞争暗流在未来几年汇聚成潮，美光当前的定价权与超级周期便将面临严峻的下行拐点。

- 最新季度营收达 54.00 亿美元，同比增幅高达 380.00%。 - 下一季度营收业绩指引达 61.00 亿美元，显著超出分析师预期的 57.00 亿美元。 - 多头核心支柱：供需链极度紧张、下游科技巨头（Alphabet、Amazon、NVIDIA）资本开支不减、年内即将重启的低估值股票回购。 - 空头核心逻辑：存储行业商品属性未变、中韩竞对产能全面加速，繁荣期或在未来数年遭遇供求逆转。

[10:00](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=600s) Extreme Volatility, Surging CapEx, and the Inevitable Memory Cycle

The discussion opens with a stark reminder of the inescapable economic reality governing the memory chip market: memory is, and always has been, deeply cyclical. Right now, capital expenditures (CapEx) across the industry are surging through the roof alongside runaway revenue growth. Companies can certainly afford this aggressive spending at the moment, and from an operational standpoint, this flush cash environment is precisely when management teams ought to be reinvesting heavily into their business. However, that flood of capital inevitably guarantees that substantially more supply will come online down the road, setting the stage for an eventual cyclical downshift as industry-wide capacity catches up with demand.

For individual investors, navigating this environment requires a strong stomach because high volatility is simply par for the course. Looking at the data, the host points out an eye-popping statistic: Micron has already logged 45 distinct single-day price moves of greater than 5% this year alone. That degree of whipsaw trading makes it abundantly clear that this is a high-beta momentum vehicle rather than a defensive buy-and-hold staple. Emphasizing the wild swings, the analyst underscores: “That is not a widows and orphans stock.”

- **Surging Capital Expenditures**: CapEx is climbing alongside top-line revenue, as semiconductor manufacturers deploy peak-cycle cash flow into long-term fabrication expansion. - **Extreme Price Volatility**: Micron has recorded 45 separate single-day price moves exceeding 5% in 2026 alone, reflecting relentless momentum swings. - **Cyclical Reality**: Expanding capacity will inevitably increase total market supply over time, challenging today's peak pricing power.

[10:40](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=640s) The Breaking Oligopoly and the Threat of Chinese Competition

The conversation pivots to the competitive landscape, highlighting how the market's traditional structure is undergoing structural disruption. Entering 2026, the DRAM and memory space operated essentially as a three-player oligopoly, closely resembling the domestic U.S. telecom sector. The three entrenched incumbents—Micron, Samsung, and SK hynix—could comfortably maintain an unspoken understanding, an industry dynamic where everyone could “wink, wink, nudge, nudge” to keep global supply constrained, protect record-breaking margins, and ensure that everyone enjoyed immense profitability.

```mermaid flowchart TD accTitle: Memory Market Structural Disruption accDescr: Diagram illustrating how the traditional memory oligopoly is being disrupted by new entrants targeting commoditized DRAM.

subgraph Oligopoly ["Traditional Memory Oligopoly"] M["Micron"] S["Samsung"] SK["SK hynix"] M --- S --- SK Oligopoly -->|"Collusive pricing & discipline"| PM["High Margins (87% Peak)"] end

subgraph Disruption ["Emerging Threat & Market Shift"] CX["Chinese Manufacturers"] CX -->|"Willingness to take lower margins"| DRAM["Standard DRAM Market"] M -->|"Pivots wafer capacity toward"| HBM["High-Bandwidth Memory (HBM)"] end

DRAM --> Loss["Erosion of Core DRAM Market Share"] HBM --> Normal["HBM Margins Normalize (e.g., 65%)"] ```

That comfortable equilibrium is now under serious threat from emerging Chinese memory manufacturers who are entirely outside that cartel-like dynamic. Seeing massive profits being generated in standard memory products, these new entrants operate under a fundamentally different incentive structure. Unlike their American and South Korean counterparts, they are far less focused on optimizing short-term shareholder returns or defending fat profit margins, and are fully willing to undercut the market to capture volume. Drawing a historical parallel to consumer hardware—where PC assembly migrated from places like South Dakota overseas to Asia, or how Foxconn built its empire on thin margins—the hosts warn that Micron risks ceding its core commodity DRAM business while chasing premium HBM products. As the analyst explains: “They're not in that same wink wink nudge nudge place. They're in the, oh my gosh, we see profits over here, we can go get some of those profits.”

- **Cartel Dynamics Challenged**: The historical three-firm oligopoly had maintained tight supply discipline and sky-high gross margins near 87%. - **Divergent Incentives**: Chinese memory makers are willing to operate at significantly lower profit margins to aggressively capture DRAM market share. - **The Long-Term Hazard**: While Micron may retain a lucrative High-Bandwidth Memory (HBM) business at strong ~65% margins, it risks permanently surrendering market share in standard DRAM to lower-cost competitors.

[11:55](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=715s) Long-Term Bearish Headwinds vs. Short-Term Momentum Plays

Wrapping up the analysis, the speakers contrast the near-term trading reality against the multi-year fundamental outlook. Over the immediate horizon of the next quarter or two, Micron's operational numbers are projected to remain spectacular, and management is not expected to post a weak earnings report anytime soon. However, equity markets constantly price in expectations 18 to 24 months forward, looking right past upcoming blowout quarters to assess where the cycle will peak and normalize. When extending the forecast horizon out three to five years, the case for remaining bullish becomes far more difficult to defend.

The analyst notes that while short-term momentum traders may very well capture profitable bounces off recent pullbacks, history suggests that holding through a cyclical top eventually leads to downside mean reversion. While acknowledging that the current dip might present an attractive short-term momentum entry, the host delivers a cautious concluding rule for long-term investors: “Look, I think my best advice to investors anyway is to focus on the future, but be mindful of the past. And in this case, I think the past suggests the long-term trajectory is lower.”

- **Near-Term Performance**: Quarterly results are projected to stay strong through early next year, insulating the company from immediate earnings misses. - **Forward-Looking Horizon**: Institutional markets look 18 to 24 months ahead, prioritizing supply additions and margin compression over current-quarter beats. - **Trading Strategy**: While short-term momentum buyers may capitalize on current dip-buying setups, multi-year buy-and-hold investors face an increasingly bearish risk-reward profile.

## Notable quotes > [01:11](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=71s) The higher the price, the more the scarcity, the more you are incentivizing smart people to put their brains to work in solving the problem that is your margin.

> [01:11](https://www.bilibili.com/video/BV1oDJQ6ZEPp?t=71s) If Apple sets up relationships and is able to source with third parties, I don't think there's a given that that business ever comes back.

## Mentioned 美光科技 (Micron), 长鑫存储 (CXMT), 三星电子 (Samsung), SK海力士 (SK Hynix), 苹果 (Apple), 英伟达 (NVIDIA), 谷歌 / Alphabet, 亚马逊 (Amazon), 富士康 (Foxconn), The Motley Fool, Lou, 浑水摸鱼清源, CHIPS Act (美国芯片法案)

## Video description 【The Motley Fool】半导体崩盘，美光MU暴跌23%，还能买吗？ Micron Stock Is Falling - Should You Buy the Dip？

低估值诱惑与暴跌风暴大家好，近期我在复盘半导体龙头美光科技（Micron）时发现，其股价自历史高点已重挫23%，目前前瞻市盈率（Forward P/E）仅为5倍左右，当前交易价位回落至每股92.20美元附近。表面上看，这样一个极其廉价的估值吸引了大量散户试图左侧抄底。然而，回顾过去一年，美光凭借全球数据中心对AI算力近乎无限的追捧，股价曾爆涨超过500%。这波惊人涨幅的背后，实际上隐藏着市场对长线供给侧剧变以及半导体大周期魔咒的深深担忧。

炸裂财报与历史级毛利率从最新公布的季度财报数据来看，美光的业绩表现堪称现象级。上一季度公司实现了高达54.00亿美元的营业收入，同比暴增380.00%；而管理层给出的下一季度营收指引进一步攀升至61.00亿美元，轻松超越了华尔街分析师原先预期的57.00亿美元。在盈利指标方面，美光的毛利率恐怖地跃升至87.00%（即每1美元销售额中产生87美分毛利），营业利润率更是高达81.00%。对比2023年11月季度毛利率为负值的暗淡时光，当前的盈利水平创下了行业历史新高。此外，公司计划在今年晚些时候启动规模高达350.00亿美元的股票回购计划，在极低市盈率下，短期基本面与做多动能依然非常强劲。

供给侧产能释放与寡头格局打破我在深入拆解盘面供给端时注意到，华尔街看空美光的核心逻辑在于芯片产能的长期扩张。当前美光、三星和SK海力士三巨头掌控的垄断格局正迎来外溢压力。中国本土存储厂商长鑫存储（CXMT）已明确宣布将在今明两年大幅提升DRAM产能，而到2027年、2028年至2029年期间，三星、SK海力士以及美光自身都将有巨量新产能集中释放。虽然目前美光将大量DRAM晶圆转向生产高带宽内存（HBM），导致普通DRAM供给吃紧、价格暴涨500.00%至600.00%，但这留下的传统DRAM市场空白正在被竞争对手快速填补。一旦苹果、英伟达、谷歌或亚马逊等大客户在2027年或2028年调整硬件架构或寻找替代供应商，美光恐将彻底失去这部分长期客户。

大周期回归与极端波动风险从宏观周期来看，半导体存储行业向来是高度大周期性的商品化市场。历史规律表明，87.00%的毛利率和81.00%的营业利润率极大概率面临均值回归，未来数年内毛利率极可能重新回落至30.00%至40.00%的常态区间。在资本市场上，美光今年已录得45次单日涨跌幅超过5.00%的极端波动，展现出极强的投机动能属性，绝非适合长期无脑躺平的稳健标的。对于看重未来3到5年长期回报的投资者而言，供需平衡点的逆转将是股价最大的下行隐患。

三条可执行性操作指引控制仓位切忌盲目重仓抄底：鉴于美光单日波动超5.00%的次数高达45次，即便5倍前瞻市盈率诱人，也应将持仓严格控制在总投资组合的5.00%以内，避免左侧重仓承压。密切追踪2027年产能释放节点：重点监控2027至2029年间长鑫存储及日韩巨头的产能上线进度，一旦全球供给过剩信号显现，及时锁定短期反弹利润。熟练运用跟踪止损限价单：在交易软件中设置回撤比例为6.00%至8.00%的跟踪止损限价单（Trailing Stop-Limit），在享受动能上涨的同时有效防御突发重挫。

简洁时间轴 [00:00] 美光暴跌23%与5倍市盈率背景 [01:11] 长期供应链压力与长期产能规划 [03:41] 高带宽内存HBM与DRAM需求变迁 [05:36] 惊人的87%毛利率与历史财务对比 [07:34] 美光股票的多头逻辑与回购计划 [09:37] 竞争加剧、周期性及极端波动风险 [11:00] 寡头格局冲击与中国厂商威胁分析

[Source](https://www.bilibili.com/video/BV1oDJQ6ZEPp)

中文简报

美光重挫23%触发抄底潮供给潮与周期均值回归压制长线回报美光当期虽录得87%毛利率与极低市盈率，但长鑫存储等竞对扩产与传统DRAM份额流失，预示其未来三至五年毛利率将不可避免地均值回归至常态区间。节目围绕美光科技股价回撤23%后的多空分歧展开全方位剖析。尽管美光凭借AI热潮实现单季营收暴增380%、毛利率飙升至历史罕见的87%，并在极低估值下拥有强劲的做多动能与回购预期，但行业基本面正面临严峻挑战。随着高带宽内存（HBM）挤占产能推高常规DRAM价格，长鑫存储等新入局者正趁机切入，打破美光、三星与SK海力士以往的默契寡头格局；加上2027至2029年全球巨量新增产能集中释放，存储行业的强周期商品属性与均值回归风险，将使美光中长期的超额盈利与估值承受巨大下行压力。美光（MU）股价自历史高点回撤23%至每股约92.20美元，前瞻市盈率降至仅约5倍；最新季度营收达54.00亿美元（同比激增380%），毛利率与营业利润率分别飙升至87%与81%的历史峰值。 (0:00) 高带宽内存（HBM）的晶圆消耗量约为传统DRAM的3倍，引发消费级存储供应严重紧缺与价格暴涨500%至600%，但这为长鑫存储（CXMT）等外部竞争者填补常规DRAM市场空白提供了战略窗口。 (1:11) 2027年末至2029年全球将迎来三星、SK海力士、美光及中国厂商的大规模产能集中释放，美光与日韩三巨头的寡头定价默契面临瓦解，87%的超高毛利率长期极可能面临均值回归至30%至40%。 (10:40) 美光年内已录得45次单日波动超5%的极端交易日，展现极强动量属性；尽管短期业绩受下游科技巨头强劲AI开支与股票回购支撑，但资本市场已开始提前18至24个月定价远期供需拐点。 (10:00)

Reading summary

节目分析认为，美光科技（Micron）近期股价虽回撤23%、前瞻市盈率仅约5倍且当期业绩爆发，但其正处于存储芯片大周期的历史极值点；随着资本开支剧增、外部新产能释放及寡头格局松动，长期将面临显著的均值回归与结构性竞争风险。当期财务狂飙与短期动量支撑 (7:34) 美光最新季度营收达到54.00亿美元，同比暴增380.00%，并给出了下一季度61.00亿美元的强劲指引，远超分析师预期的57.00亿美元。受益于AI数据中心建设狂潮，美光毛利率飙升至87.00%的历史极值，营业利润率达到81.00%，相较于2023年11月季度的负毛利和亏损呈现爆发式扭转。此外，受芯片法案限制的股票回购计划预计在年内重启，科技巨头Alphabet、亚马逊与英伟达在算力集群上的资本开支未减，加之未来12个月内全球极少有显著新增产能上线，为多头提供了强劲的短期业绩护航。 HBM挤占效应与大客户架构替代风险 (1:11) 制造高带宽内存（HBM）消耗的晶圆资源约为传统DRAM的3倍，这种对基础产能的虹吸导致消费级存储芯片严重短缺，终端采购成本飙升500%至600%。节目警示，美光重注高利润HBM所留下的标准DRAM市场空白，正给长鑫存储（CXMT）等外部竞争对手创造进入良机。长远来看，若苹果等核心大客户因缺货而与第三方建立起替代采购关系，外流的核心业务恐难以回流；同时，面对暴涨的存储成本，硬件巨头在2027至2029年的芯片架构迭代中存在减少HBM用量或采用常规替代方案的研发动力。三寡头平衡松动与长鑫扩产冲击 (10:40) 存储行业长期由美光、三星和SK海力士三巨头维持相对默契的限产保利格局，但以长鑫存储为代表的中国厂商正打破这种平衡。与追求短期股东回报和极高利润率的传统巨头不同，新兴厂商更倾向于接受较低利润率以抢占市场份额，其公开宣布的今明两年大幅扩产计划将重构行业供需。尽管美光在高端HBM领域可能维持约65%的利润率，但在标准DRAM这一商品化大盘中面临永久流失市场份额的长期隐患。资本开支激增与周期均值回归 (10:00) 存储芯片本质上仍是强周期、高度商品化的行业。当前全行业激增的资本开支虽然受充沛现金流支撑，但必然转化为远期产能的过剩。分析指出，2026年至2027年初虽极少有大规模新产能上线，但2027年末至2029年美光、三星、SK海力士以及中国厂商的新增产能将集中释放。分析师Lou直言，将87%的极端毛利率视为“新常态”是最不可能的情景，未来3至5年行业毛利率大概率回落至30%至40%的历史常态区间。美光年内已录得45次单日波动超5%的极端行情，属于典型的高波动交易标的而非长线防御资产。尽管极度紧俏的供需和强劲指引有望在未来几个季度支撑动量交易，但股市通常提前18至24个月反映周期见顶与产能过剩预期，长期持有的投资者需警惕估值与利润率均值回归带来的下行风险。

Source quotes

[1:11] The higher the price, the more the scarcity, the more you are incentivizing smart people to put their brains to work in solving the problem that is your margin. [1:11] If Apple sets up relationships and is able to source with third parties, I don't think there's a given that that business ever comes back.

Original transcript

[0:00] Micron stock is down 23% and a lot of investors think that this is a great value trading for just five times forward earnings but Lou, the big question for us today is, Is this a great buy right now or... Have we already hit the top with Micron? Let's get into it. First of all, from a supply standpoint, one of the things that's driving Micron stock short term is some of the added supplies coming on into the market. There's a Chinese company called CXMT that said that they're gonna be producing product, more product not only this year, but also next year. That goes into the supply side of things. Do you think that a Chinese memory supplier is even something to think about if you're looking at Micron stack? Yeah, I think you have to almost inevitably. They may not be able to compete for all of the business, but we're talking about the total pie. And any part of the total pie that's under pressure is going to raise question about demand. I mean, yeah, Micron is down, but what? They're up more than 500% over the last year on this almost unlimited demand. If there is a whole other player that it's at least competing for part of this, that has to change the economics of the business, period. Yeah. Well, and the other piece from a supply standpoint is no one's standing still. You know, the quarter that we're in right now, we're in the third quarter of 2026, maybe even into early 2027. Nobody's really adding a lot of supply at that period of time. But you start to get to the end of 2027, into 2028, 2029. suddenly, Micron's going to add supply, SK Hynix is going to add supply, Samsung, you have at least one or two Chinese suppliers that are going to bring more supply to the market. So, you know, we always hear about there being more demand than supply, but it seems like from a supply side, there's a lot more coming online now. Which... fundamentally is going to change that formula. Right. And you know, it gets worse if you think long term, too, because some of this I mean, look, AI demand could last forever and let's hope it does, right, for the sake of the stock markets and all these stocks. But to some extent, the core business from Micron is this lesser memory, whether it's iPods or all of these other products that need the memory. DRAM. Yeah, yeah. And to the extent that even if CXMT just takes that out, well, that is your long-term memory. customer that might be shifting. I don't think Apple's going to say, okay, my bad. We'll just come back to you if you want us. If Apple sets up relationships and is able to source with third parties, I don't think there's a given that that business ever comes back. So increasingly, you're not just benefiting from the AI upside if you're a micron shareholder, you're depending on it if some of that other business is going elsewhere, which kind of changes the setup for the stock long term, I think. Yeah, the other thing that I want to highlight just from that supply standpoint is we talk a lot about HBM or high bandwidth memory and what that effectively is is you take a three wafers, you know, from a supply standpoint, the supply is constrained by the number of wafers that these companies are able to make you take three of them. From DRAM, you turn that into HBM, oversimplifying it, but from just a strictly supply-demand standpoint, that's why we don't have enough chips to make more iPhones, why prices for computers are going up, all that kind of stuff. There's less... DRAM potentially available, it's because that DRAM is going to HBM. And that's what's in such high demand. So what you're saying here is that gap that Micron and other companies are leaving behind in DRAM is going to be filled by somebody else. Yes, it is going to be filled in China first. But those products could eventually make their way into the products that we know today. The other thing that I want to highlight, because we'll get to things like marginal, what their potential profitability will be in just a second, but... The demand side, I think, is also something that we're not talking enough about. We know that we're supply constrained at this point, but every single one of these demand sources, whether you're Apple, whether you're Nvidia, has agency. OK, so you can't make a change in March of 2026 that is going to impact how much RAM is going to go into or how much memory is going to go into your product in September 2026. But you can make a change. that will impact your product in 2027, 2028, 2029. What do you think about that from a demand side as these prices have gone up in some cases 500, 600 percent? Are we going to see that demand? start to moderate a little bit as these companies say you know what we maybe maybe we'll put a little bit more DRAM instead of HBM into our products maybe we'll move we'll reconfigure uh our processor a little bit differently so that we're using less of this very very expensive component To the extent they can. And look, I'll be honest, I don't think I know enough to predict the future here and say that will happen. But I think the point you're making and one that investors should think about is, is that as the cost goes up, you're incentivizing smart people to find workarounds. [5:00] And can they totally, I doubt it, to be honest. I think there's a reason they're using these high-bandwidth products, and I think you need that to some extent. But can you find ways to need a little less of it or to let the margins change what you're doing? Again, the higher the price, the more the scarcity, the more you are incentivizing smart people to put their brains to work in solving the problem that is your margin. think investors should be mindful of at least. So One of the things that... We often talk about is margins here with these kinds of products. And we've seen in the AI build out that a lot of companies have seen their margins go absolutely crazy. This is a chart just going back to November 2023. These are quarterly numbers. In blue is the gross profit margin. So we're now up to 87%, meaning. For every dollar worth of sales that Micron has, 87 cents of that, is... gross profit 81 cents of that is an operating profit. Just. A couple of years ago, November 2023 quarter... even the gross profit was negative. We can rewind this all the way back to 2015 These are margin levels that we've really never seen. When you think about these margin levels long term, and you can bring a bunch of companies into this like NVIDIA, Do you think that there's going to be a mean reversion where eventually – maybe not next year, but three, four, five, six years out, we do go back to maybe a 20, 30, 40% gross margin. Maybe those operating profits go down. Isn't this typically what happens in these cyclical industries? It feels like it. I mean, I don't want to be the guy saying it's different this time because I don't think it might be different for the next few quarters. It might be different for now, but long term, I can't imagine it's different. You already said it. This is a commoditized business. Traditionally, there is more. The bull case for now is a lot of that new supply won't be online until 2028. The bear case is 2028 is right around the corner and it is happening. This becomes a momentum trade and not a long term investment in that world. me to believe that numbers that they have never been able to consistently generate in the past are the new normal just because they got there. That is probably the least likely scenario in my head. Okay, let's... We'll get to the valuation a second, but let's do let's do a bull case in a bear case. What is the bull case for Micron from here? Is it just about operations? There should be a buyback component. I don't think they can start that until later this year because of some funding they got from the U.S. government. But heck, if you're buying back stock at five times earnings, That could be pretty compelling. When you look at... If this is going to go from about $922 a share as we're recording to $1,500, $2,000 per share, what has to happen? So the bull case is scoreboard, right? I mean, the bull case is more of the same. They did $54 billion in revenue in the most recent quarter. That was up 380%, I think, year over year. Adjusted earning per share is just going crazy with these margins. Guidance is for even more. I think they're guiding for $61 billion against about, what, 57 that analysts had expected. You still have tight supply chains. You still have massive demands. to come online in the near future. With margins like that, you can buy back your shares. So the bull case for now is, may this last. Yeah, and I think to just kind of I reemphasize that most of these companies that we're talking about in the AI build out don't yet show any signs of slowing down. We'll maybe critique that in a moment, but if you look at Alphabet and Amazon, NVIDIA, all these companies that are either building or buying chips, they're saying, hey, more, more, more, more, more. And that means more memory. The other piece of that, even if HBM is not super strong, there's still a huge consumer demand and there's more consumer devices coming all the time that all need memory. So at least, you know, over the next 12 months or so, not a lot of new supplies coming online. Some of that supply will come online. Maybe that even provides growth long term for these companies, long term meaning, you know, in the next three to five years. I think it's pretty hard to figure out what's going to happen five plus years from now, but this should be a pretty good run for Micron. Okay, let's look at the bear case. If the stock continues to fall, what is the reason for that in your mind? So we've talked about a couple of things, but I think just to emphasize the points we've made, competition is growing. And the Chinese players, if anything, I think this week it's just been a reminder of... that competition is out there more than it is a real surprise but whether it's Samsung whether it's Hynix whether it's Chinese or wherever else there's just going [10:00] be over time more supply in the system. We already said it, memory is cyclical. It always has been. I think it will be. Part of it is because more supply can come on. As everybody is building up supply, CapEx is going through the roof along with revenue. Now, they can afford it. Now is the time to be investing in your business. The other thing too is, is that as an investor, you have to understand that volatility right now is just par for the course. I looked it up real quick. Micron has had 45 moves greater than 5%. This year. 45 single day moves greater than 5%. That is not a widows and orphans stock. The competition piece, I think... we really need to highlight because what was unique coming into 2026 was that this was an oligopoly. This is a little bit like the US telecom industry, right? There's three major suppliers. They can all kind of wink, wink, nudge, nudge. You know what? Let's keep supply where it's at right now. Let's keep these margins really high. Things are going really well for all of us. The reason I think these Chinese suppliers are so interesting in this space right now and such a threat from a bear side is they're not in that same wink wink nudge nudge place. They're in the, oh my gosh, we see profits over here, we can go get some of those profits. And those Chinese manufacturers have a very different incentive structure than a lot of the U.S. and South Korean companies. that have kind of built out the industry that we know today. So they are maybe less... short-term or long-term profit motivated you know they're willing to drive those margins down we've seen that with all kinds look at computers right computers that used to be made in South Dakota are now made in Asia because. they're willing to do it for far less profit margin. Foxconn, this is how that business was built. So I think this is something really to keep an eye on because We may look back in a few years and say, hey, Micron still has this HBM business. Maybe it's not quite the 87% margin that we have today, but it's a really great business at 65% margins. But they completely gave up. the DRAM business because they allowed some of these other suppliers to come in I tend to fall on that bearish side. with these companies because the run has been so strong the market looks out 18 to 24 months those markets not doesn't really care about what's going to happen next quarter numbers are going to be great let's let's just double emphasize that this This is not a company that's going to report a bad quarter next quarter or even early next year. Do you fall on that? bullish side or bear side if you look out three to five years. I think it's harder to be bullish the longer term your outlook is. I don't doubt that. momentum traders can benefit from here. But yeah, the longer you look out, I do think, look, I think my best advice to investors anyway is to focus on the future, but be mindful of the past. And in this case, I think the past suggests the long-term trajectory is lower. But look, it works until it doesn't. And in the near term, I think this could be a momentum buying opportunity this week. Yeah, we will see. I think there's some good bear and bull cases for investors. But let us know where you fall in the comment section below. Don't forget to subscribe to the Motley Fool's channel here on YouTube. Thanks for watching, everybody. See you here next time. . .

美光重挫23%触发抄底潮 供给潮与周期均值回归压制长线回报