• 亿邦公众号
    • APP下载

三次濒临破产 如今市值破万亿 “亚洲新巨头”一季净赚超千亿

作者:天下网商 文章来源:天下网商

2026-06-09 11:56

作者:天下网商 文章来源:天下网商

2026-06-09 11:56

邦小白快读

EN
全文速览

本文核心讲了韩国SK海力士从三次濒临破产到市值突破1万亿美元,成为AI产业链隐形赢家的发展历程,干货要点如下:

1. AI产业当前的财富逻辑已经变化,淘金热中最赚钱的不一定是做模型的淘金者,而是做核心基础设施的送水人,SK海力士就是靠给AI巨头供应核心内存赚到超额收益,一年股价暴涨10倍,单季净利润超过千亿,净利润率超过了英伟达。

2. SK海力士的成功不是偶然,它十五年前就布局了HBM高带宽内存技术,哪怕早期没有应用场景、持续亏损也没有砍掉该产品线,最终熬到AI大模型风口,凭借超过一半的全球市场份额获得垄断性收益。

3. 当前该赛道已经出现估值过热信号,行业周期性波动的风险依然存在,普通投资者需要注意规避追高风险。

本文为科技品牌展现了AI浪潮下的产业趋势,总结了可借鉴的发展经验与风险提示,核心干货如下:

1. 当前产业趋势已经明确,AI产业从拼大模型参数的竞争阶段,进入拼基础设施的竞争阶段,高端核心元器件成为稀缺资源,价值会被市场重新定价,品牌需要提前抓住基础设施赛道的机会。

2. 产品研发层面,SK海力士的案例印证了长期布局前瞻性核心技术的价值,它早在2013年就推出全球首款HBM内存,持续迭代十五年,最终获得超过52%的全球HBM市场份额,下一代产品份额将提升到54%,实现了对行业龙头的弯道超车。

3. 风险层面,存储芯片的周期性属性未变,当前三星、美光都在加速扩产,供给缺口未来会逐步收窄,行业已经出现估值过热信号,同时行业受出口管制、政策变动影响大,品牌需要提前应对供需波动。

本文为AI产业链相关卖家梳理了当前赛道的增长机会、可借鉴经验与风险提示,核心干货如下:

1. 当前AI基础设施赛道存在大量增量机会,除了核心的HBM内存,液冷设备、先进封装材料、服务器电源管理芯片、光模块等过去不起眼的细分赛道,都因为AI需求爆发成为刚需,获得资本市场重新估值,卖家可以切入这些配套赛道寻找增长机会。

2. 可参考的发展经验:SK海力士的案例说明,只要押对符合产业趋势的长期核心技术,哪怕短期没有需求、企业陷入经营危机,也要坚持布局,最终风口来临时就能获得超额收益,SK海力士靠前瞻布局实现一年股价涨10倍,单季净利润率达到77%,超过全球AI芯片龙头英伟达。

3. 风险提示:当前存储赛道已经出现估值过热,头部厂商加速扩产,未来供给缺口会收窄,同时行业受出口管制、政策变动影响大,卖家需要警惕周期波动带来的经营风险。

本文对制造工厂,尤其是半导体及配套领域工厂的转型发展,梳理了明确的需求方向、商业机会与发展启示,核心干货如下:

1. 当前产品生产设计的需求方向已经明确,AI大模型发展带来“内存墙”痛点,原有平铺式普通内存带宽不足,无法满足大模型数据调用需求,垂直堆叠的HBM高带宽内存成为刚需,技术路线清晰,高端芯片制造工厂可以向这个方向转型布局。

2. 全产业链都存在新的商业机会,AI基础设施建设带动了大量配套环节的需求增长,除了高端存储,液冷设备、先进封装材料、服务器电源芯片等细分配套零部件的需求都在快速提升,相关制造工厂可以抓住本轮产业升级的机会切入新赛道。

3. 发展启示:SK海力士三次濒临破产都没有砍掉前瞻性的HBM研发线,最终熬到AI风口获得超额收益,制造工厂需要有布局长期核心技术的耐心,同时也要警惕存储行业的周期性波动风险,合理安排产能扩张节奏。

本文为AI产业链相关服务商梳理了行业发展趋势、客户痛点与市场机会,核心干货如下:

1. 当前AI行业的发展趋势已经明确,AI产业已经从前期拼大模型参数的竞争阶段,进入拼基础设施的落地竞争阶段,产业链对核心基础零部件和配套服务的需求大幅爆发,稀缺环节会被市场重新定价,给各类服务商带来大量新增需求。

2. 当前AI产业的核心痛点与技术方向明确,AI大模型参数量不断提升,运行过程中需要从内存大量调取数据,普通内存带宽不足形成“内存墙”,限制了算力发挥,对应的成熟解决方案就是HBM高带宽内存,目前该技术仅被少数厂商掌握,供应链相关配套服务商有大量切入空间。

3. 除核心存储领域,液冷、先进封装、光模块等下游配套环节的需求也在快速增长,相关技术服务、配套供应服务商都可以抓住本轮AI基础设施建设的红利,同时需要关注行业周期和政策变动带来的波动风险。

本文对AI相关产业平台的招商布局、运营优化和风险规避都有参考价值,核心干货如下:

1. 当前产业对平台的需求已经发生变化,AI产业发展带动基础设施类产业链的需求大幅提升,原来不受关注的存储芯片、各类配套零部件赛道快速升温,平台可以围绕AI基础设施赛道调整自身的招商方向与布局重心,挖掘高成长的优质企业。

2. 平台可以跟进的最新方向:当前HBM等高端AI元器件供给缺口明显,谷歌、英伟达等头部科技企业都主动投资相关厂商扩产,平台可以围绕AI基础设施产业链推出专项对接服务、扶持政策,打通供需两端,抓住本轮产业升级的发展机会。

3. 风险规避要点:存储等AI基础设施赛道的周期性较强,当前行业已经出现估值过热信号,头部厂商都在加速扩产,未来供给缺口会逐步收窄,同时行业受出口管制、政策变动影响大,平台在运营中需要提示相关风险,避免过度布局过热赛道。

本文呈现了AI浪潮下全球半导体存储产业的最新动向,为产业研究提供了典型案例与新的研究方向,核心干货如下:

1. 产业发展新动向:AI发展正在重构全球半导体产业链的议价权,AI产业已经从大模型研发竞争转向基础设施落地竞争,HBM等高带宽存储成为核心稀缺资源,SK海力士凭借十五年的前瞻布局,占据全球超过一半的HBM市场份额,成为AI产业链的隐形赢家,单季净利润率甚至超过AI芯片龙头英伟达,彻底改变了原有存储芯片低利润的行业格局。

2. 产业需要关注的新问题:存储芯片原有的周期性规律并未消失,当前赛道已经出现估值过热,三星、美光等龙头都在加速扩产,未来供需关系可能出现反转,同时行业受出口管制、产业补贴政策变动等外部变量影响大,波动风险仍然存在。

3. 研究与政策启示:AI淘金热中,基础设施配套环节的价值长期被低估,“送水人”商业模式的回报率在当前阶段已经超过台前的技术龙头,这为产业研究提供了新视角,也为政策支持前瞻性核心技术研发提供了典型案例参考。

声明:快读内容全程由AI生成,请注意甄别信息。如您发现问题,请发送邮件至 run@ebrun.com 。

我是 品牌商 卖家 工厂 服务商 平台商 研究者 帮我再读一遍。

Quick Summary

This article outlines the growth journey of South Korea's SK hynix, which went from three near-bankruptcy events to a market capitalization exceeding $1 trillion, emerging as a hidden winner in the AI industry chain. Key takeaways are as follows:

1. The wealth logic of the AI industry has shifted: in an AI "gold rush", the biggest winners are not always the model-building "gold diggers", but the "water carriers" that build core infrastructure. SK hynix has earned outsized profits by supplying core memory to AI giants: its stock has surged 10x in a year, it posted over 100 billion RMB in net profit for a single quarter, and its net profit margin even outpaces NVIDIA.

2. SK hynix's success is no accident. It began investing in High Bandwidth Memory (HBM) 15 years ago, and kept the product line despite early lack of use cases and sustained losses. When the AI large model boom arrived, it captured monopoly-level profits with over 50% of the global HBM market.

3. Valuation overheating has already emerged in this sector, and the risks of cyclical industry fluctuation remain. Retail investors should be cautious of chasing prices at their current highs.

This article outlines AI industry trends for technology brands, summarizes actionable development lessons and risk alerts. Key insights are as follows:

1. The current industry trend is clear: AI competition has shifted from a race for larger model parameters to a race for infrastructure. High-end core components have become scarce assets that will be revalued by the market, and brands should seize opportunities in infrastructure-focused tracks early.

2. On product R&D, the SK hynix case proves the value of long-term布局 of forward-looking core technologies. It launched the world's first HBM product as early as 2009, has iterated on the technology for 15 years, and now holds over 52% of the global HBM market share, a figure set to rise to 54% for its next-generation product, allowing it to overtake the longtime industry leader.

3. On risk, the cyclical nature of memory chips remains unchanged. Samsung and Micron are both accelerating capacity expansion, which will gradually narrow the current supply gap. Valuation overheating has already appeared in the sector, and the industry is highly exposed to export controls and policy shifts. Brands should prepare in advance for supply and demand volatility.

This article sorts out growth opportunities, actionable lessons and risk alerts for sellers along the AI industry chain. Key takeaways are as follows:

1. There is massive incremental opportunity in the AI infrastructure track. Beyond core HBM, previously overlooked niche segments including liquid cooling equipment, advanced packaging materials, server power management chips and optical modules have now become essential amid the AI demand boom, and are being revalued by capital markets. Sellers can enter these supporting segments to pursue growth.

2. Key lesson from SK hynix: if you bet on the right long-term core technology aligned with industry trends, you should stick to your布局 even if there is no near-term demand and your business faces crisis. When the market boom arrives, you will capture outsized returns. SK hynix's 10x stock surge in one year and 77% single-quarter net profit margin -- which outpaces AI chip leader NVIDIA -- proves this strategy.

3. Risk alert: Valuation overheating has emerged in the memory track, and leading players are accelerating capacity expansion that will narrow the supply gap over time. The industry is also highly sensitive to export controls and policy shifts, so sellers should watch for operational risks brought by cyclical fluctuations.

This article outlines clear demand directions, business opportunities and development insights for manufacturing factories, especially those in the semiconductor and supporting sectors. Key takeaways are as follows:

1. The demand direction for production and design is now clear: the growth of large AI models has created a "memory wall" bottleneck. Conventional planar memory lacks sufficient bandwidth to meet the data retrieval requirements of large models, making vertically stacked HBM an essential product. The technical path is clear, and high-end chip manufacturing factories can transform and布局 this segment.

2. New business opportunities exist across the entire industry chain. AI infrastructure build-out has boosted demand for a wide range of supporting links. Beyond high-end memory, demand for niche supporting components including liquid cooling equipment, advanced packaging materials, and server power chips is growing rapidly. Relevant manufacturing factories can seize this industry upgrade opportunity to enter new tracks.

3. Development insight: SK hynix did not cut its forward-looking HBM R&D line even when it was on the brink of bankruptcy three times, and eventually captured outsized returns when the AI boom arrived. Manufacturing factories need the patience to布局 long-term core technologies, while also remaining alert to the cyclical fluctuation risk of the memory industry and arranging capacity expansion at a reasonable pace.

This article sorts out industry development trends, customer pain points and market opportunities for service providers along the AI industry chain. Key takeaways are as follows:

1. The current AI industry trend is clear: AI has moved from an early stage of competition over large model parameters to a stage of implementation competition focused on infrastructure. Industry demand for core basic components and supporting services has surged, and scarce segments are being revalued by the market, bringing massive new demand for all types of service providers.

2. The core pain points and technical directions of the current AI industry are clear: as large AI models grow in parameter size, they need to retrieve large volumes of data from memory during operation. Insufficient bandwidth of conventional memory creates the "memory wall" bottleneck that limits computing power performance, and HBM is the proven mature solution. Currently, the technology is only mastered by a small number of manufacturers, leaving massive room for entry for supporting supply chain service providers.

3. Beyond core memory, demand for downstream supporting links such as liquid cooling, advanced packaging and optical modules is also growing rapidly. Relevant technical service and supporting supply service providers can all capture the dividend from this round of AI infrastructure build-out, while also needing to watch for volatility risks brought by industry cycles and policy changes.

This article offers valuable reference for industry platform investment positioning, operation optimization and risk mitigation in AI-related sectors. Key takeaways are as follows:

1. Industry demand for platforms has changed: AI growth has driven a sharp rise in demand for infrastructure-focused industrial chains, and previously underfollowed segments including memory chips and various supporting components are heating up rapidly. Platforms can adjust their investment positioning and sourcing focus around AI infrastructure tracks to discover high-growth, high-quality enterprises.

2. Key new directions for platforms: There is currently a clear supply gap for high-end AI components such as HBM, and leading tech players including Google and NVIDIA are actively investing in capacity expansion at relevant manufacturers. Platforms can launch specialized matching services and support policies around the AI infrastructure industrial chain to connect supply and demand, and seize the development opportunity from this round of industrial upgrading.

3. Key risk mitigation takeaways: AI infrastructure tracks such as memory have strong cyclicality, and valuation overheating has already emerged in the current market. Leading players are accelerating capacity expansion, which will gradually narrow the supply gap over time. In addition, the industry is highly exposed to export controls and policy changes, so platforms should highlight relevant risks in operations and avoid overinvesting in overheated tracks.

This article presents the latest developments in the global semiconductor memory industry amid the AI boom, and offers a典型 case and new research directions for industrial research. Key insights are as follows:

1. New industry development trends: AI growth is restructuring bargaining power across the global semiconductor supply chain. AI competition has shifted from large model R&D to infrastructure implementation, and HBM and other high-bandwidth memory products have become core scarce resources. Thanks to 15 years of forward-looking布局, SK hynix holds more than half of the global HBM market, making it a hidden winner in the AI industry chain. Its single-quarter net profit margin even outpaces AI chip leader NVIDIA, completely rewriting the original low-profit landscape of the memory chip industry.

2. New issues for industry attention: The original cyclical pattern of memory chips has not disappeared. Valuation overheating has already emerged in the current sector, and leading players including Samsung and Micron are accelerating capacity expansion, which may lead to a reversal of supply and demand dynamics in the future. The industry is also heavily affected by external variables including export controls and changes to industrial subsidy policies, so volatility risk remains.

3. Research and policy implications: In the AI gold rush, the value of infrastructure supporting links has been undervalued for a long time. The return on the "water carrier" business model now outpaces that of front-stage tech leaders, which provides a new perspective for industrial research and a典型 case reference for policy support for forward-looking core technology R&D.

Disclaimer: The "Quick Summary" content is entirely generated by AI. Please exercise discretion when interpreting the information. For issues or corrections, please email run@ebrun.com .

I am a Brand Seller Factory Service Provider Marketplace Seller Researcher Read it again.

AI算力需求爆发,这家公司靠给巨头“送水”赚到盆满钵满。

文|杨洁

AI淘金热里,最赚钱的不一定是淘金者,而是给淘金者送水的人。

在这轮AI狂欢里,定义了算力的英伟达无疑是站在聚光灯中央的角色,在它成为这轮全球技术革新中的标志性符号时,另一家公司正在英伟达身后跃升为“隐形赢家”。

它不做大模型,也不直接生产GPU(图形处理器),而是卖内存,一种在过去几十年里常被视作周期性标准品、价格起起落落、利润忽高忽低的基础元器件。

这家公司,就是来自韩国的SK海力士。

2026年5月27日,韩国股市开盘后,SK海力士股价一路上冲,盘中涨幅一度接近15%,最高达到235.8万韩元,市值首次突破1万亿美元(约合人民币6.76万亿),成为继三星电子之后亚洲第二家、全球第三家市值达万亿美元的存储芯片企业。

更夸张的是,在一年前,SK海力士的股价还只有20.3万韩元(约898元人民币)。一年暴涨10倍,放在任何一家公司身上,这样的走势都足够吸睛。

股价之外,SK海力士还因为向员工发放巨额绩效奖金而出了圈。据韩国媒体报道,其部分团队拿到的年终奖高达年薪的数倍,平均年终奖金达到60多万人民币。一家曾经差点破产、一度被银行列为不良资产的公司,如今成了整个科技行业令人羡慕的“印钞机”。

但市值跻身万亿美元俱乐部、员工奖金刷屏,这些都只是故事浮出水面的一角。真正让全球资本市场重新理解SK海力士的,是水面之下,一场横跨十五年、几度接近失败的产业豪赌。

天下网商

内存价格暴涨,海力士卡住了AI的咽喉

如果只看表面,SK海力士的爆发像是一场资本市场的情绪狂欢。

股价一年十倍,市值冲上万亿美元,员工奖金刷屏,韩国股市跟着沸腾,所有故事都像是突然发生的,但半导体行业从来没有真正的突然。

想要理解SK海力士为什么突然火成这样,得先搞清楚一个技术上的变化:在AI服务器里,内存的地位彻底变了。

过去几十年,内存一直是电脑里最没有存在感的零件之一。消费者买电脑看的是CPU型号、显卡性能,内存条往往只是装机单末尾顺手加上的一个标准件。行业里拼的是谁家便宜、谁家容量大,价格战打了一轮又一轮,厂商的利润非常薄。

但在AI大模型里,游戏规则彻底颠覆了。

今天的AI大模型,参数量动辄几千亿甚至上万亿。每一次用户跟AI对话,它背后的GPU都要从内存里疯狂调取数据。如果内存带宽跟不上,再强的芯片也只能干等。业内人士给这个问题取了个形象的名字:内存墙,这堵墙不推倒,算力再多也是白搭。

推倒这堵墙的工具,就是HBM,即“高带宽内存”。

和普通内存条不同,普通内存芯片是一颗颗平铺在主板上的,数据从内存走到处理器要跑一段不短的路。HBM是把多层芯片像叠积木一样垂直堆起来,再用一种微型导线上下打通。这样一来,数据通道数量也成倍增加。

而SK海力士,恰恰是内存行业最关键的供应商之一。它掌握着全球一半以上的HBM市场份额,英伟达、谷歌、亚马逊排着队采购,甚至主动提出要直接投资它的扩产项目。

根据集邦咨询TrendForce的数据,2026年SK海力士在HBM市场的出货量份额约为52%,三星电子约为39%,美光约为8%。据Counterpoint Research预测,下一代HBM4量产后,SK海力士将占据全球市场54%的份额。换句话说,全球超过一半的高端AI加速器,都得等着它供应的内存才能满血工作。

这种近乎垄断的地位,直接反映在了其财务数字上。

2026年一季度,SK海力士单季营收达到52.58万亿韩元,同比增加198%,营业利润37.61万亿韩元(约合人民币1661亿元),同比增加405%,净利润率更是达到惊人的77%。做个对比更直观:2027财年一季度(截至2026年4月26日),英伟达的净利润率是71.48%。

在比谁更会赚钱这件事上,这家卖内存的公司甚至一度超过全球最牛的AI芯片公司。

SK海力士不是唯一尝到甜头的存储厂商。今年以来,整个存储赛道随之沸腾,美光科技股价同步走强,多家产业链公司跟着涨停。

美国存储巨头美光科技同样在这一轮AI浪潮中赚得盆满钵满。美光科技2026年的HBM产能已经被英伟达、亚马逊、微软、谷歌云提前锁定,其股价今年以来同样走势强劲,资本市场对整个高端存储赛道都在重新估值。

除此之外,2026年一季度,三星电子设备解决方案部门的营收和营业利润双双大幅增长,内存业务创下季度销售纪录。

三大存储巨头中,SK海力士的站位更为特殊,相比三星和美光科技,它的业务集中在高带宽内存上,HBM及其相关高端产品贡献了它营收的绝对大头。

当AI需求爆发时,海力士自然而然涨势最猛。

天下网商

三次濒临破产,却等来了最大的风口

不过,海力士今天能成为产业链的关键角色,靠的不仅是运气。往回看,这是一场用十五年时间下的赌注。

1983年,韩国现代集团创始人郑周永力排众议,成立了现代电子,主攻内存芯片。这家公司就是SK海力士的前身。当时的韩国半导体产业几乎一片空白,日本和美国企业牢牢把持着技术和市场。现代电子从最基础的SRAM(静态随机存取存储器)和DRAM(动态随机存取存储器)起步,一脚踩进了全球制造业战场。

存储芯片这个行业有一个残酷的规律,玩家必须不停地砸钱搞研发、建产线,但新产品一上市就可能撞上价格暴跌。赚一年,亏三年,是常态。没有足够厚的家底,随时可能被一波下行周期直接带走。

现代电子的第一次危机来得很快。1997年,现代集团自身陷入债务泥潭,现代电子也背负着巨额负债苦苦支撑。最终债务重组和银行输血,才勉强活了下来。1999年,现代电子并购了LG半导体,完成了一次重要的产业整合。

伤口还没愈合,第二次风暴又来了。

2001年,全球存储芯片价格崩盘,刚刚脱离现代集团、改名海力士的公司,账面上已经资不抵债。债权银行不得不再次出手,进行第二轮大规模债务重组。那一年,海力士在全球DRAM市场的份额虽然排在前列,但在财务上已经是一只脚踩在破产线上了。

苦撑几年后,海力士又一次被逼到绝境。2008年金融危机后,海力士负债率飙升到危险水位,被银行接管,成了账本上的不良资产。银行四处寻找买家,韩国各大企业没有一家愿意伸手。理由很直白:内存是一个烧钱的无底洞,而且前面还挡着一个不可战胜的三星。

转折点出现在2012年。SK集团会长崔泰源决定出手,以约30亿美元的价格拿下海力士的控股权。他当时讲了一句话:我赌的不是海力士,我赌的是信息会成为下一个时代的石油。

在签约两周后,2012年2月27日,日本DRAM巨头尔必达申请破产保护。全球第三大内存制造商的倒下,直接收缩了供给端。内存价格触底反弹,SK海力士当年就实现了扭亏为盈。

但真正改变公司命运的,是2013年的一颗芯片。

那一年,SK海力士发布了全球首款基于硅通孔技术的HBM内存。刚发布时,HBM几乎找不到买家。应用场景稀少,成本高企,良率爬坡缓慢。

从当时的商业角度看,这的确吃力不讨好。但SK海力士没有砍掉这条产品线,而是持续迭代。

真正的爆发需要等待一个足够大的场景。2022年底,ChatGPT横空出世,全球科技公司突然意识到大模型的威力,HBM的价值被重新定义。

例如英伟达在设计每一代新平台时,都不得不把HBM的规格和供应放在最优先的位置。而在这个赛道上跑得最早的SK海力士,顺理成章地成为了最大的受益者。

三次接近破产,三次死里逃生,命运最终眷顾了这个还留在牌桌上的玩家。

天下网商

AI时代的送水人,开始闷声发财

放眼整个行业,可以看到一个更大的变化正在发生:AI产业正在从拼模型的阶段,进入拼基础设施的阶段。

第一阶段的关键词是大模型本身,谁能做出参数最多的模型,谁就占据叙事优势。而进入第二阶段,关键词正在变成数据中心、芯片、内存、存储这些基础设施。

SK海力士在2026年英伟达GTC大会(GPU技术大会)后的技术观察中给出过一个判断:AI竞争正在从训练最大、最聪明的模型,转向如何以更低成本部署大模型,并让它们进入物理制造和平台化场景。

推断一旦大规模铺开,内存和算力的需求就会像水和电一样必要。

为这个趋势买单的,正是全球最大的科技新贵们。

英伟达2026年5月20日发布的最新财报显示,公司一季度收入达到816亿美元,数据中心收入752亿美元,同比增长92%。谷歌、亚马逊、微软此前都已宣布千亿美元级别的AI基础设施投资计划。这些钱会沿着产业链扩散,流进每一个环节。

除了SK海力士、美光科技以及长鑫存储等,台积电也是这条链上另一个典型的送水人。GPU需要最先进的制造工艺,而台积电几乎垄断了全球AI芯片的高端代工。台积电2025年二季度财报明确指出,业务持续受AI和HPC相关需求支撑。

往下游看,AI带火的产业链环节还在扩散,例如液冷设备、先进封装材料、服务器电源管理芯片,这些过去被视为配角甚至不起眼的细分赛道,正在被资本市场逐一重新估值。

博通、台达电、鸿海、纬创、工业富联、光模块厂商、液冷厂商,都在不同位置上扮演着“送水人”的角色,它们未必站在舞台中央,但它们卖的是AI时代无法绕开的基础设施。

所以,SK海力士股价一年暴涨十倍,背后并不是一家韩国芯片公司偶然撞上风口的老套故事,而是AI产业议价权正在重新分配,真正稀缺的东西会被重新定价。

不过,存储芯片的周期性基因从未消失。6月5日,SK海力士股价在冲上高点后出现明显回调。资本市场的担忧并非空穴来风,三星和美光的扩产计划正在加速推进,HBM的供给缺口可能在未来几个季度逐步收窄。

更宏观来看,部分华尔街投资机构已经开始对存储股的仓位进行调整,有基金经理公开表示内存板块估值已出现过热信号,正在降低相关持仓比例。与此同时,存储芯片行业容易受到出口管制、关税调整、产业补贴政策变动,任何一项变量的变化都可能打乱供需节奏。

这个道理,一百多年前的加州淘金热里被验证过一次:卖水的人赚得盆满钵满,但当金矿的产量开始波动、新的水源被发现,最先受到冲击的往往也是他们。今天在AI半导体产业链上,同样的逻辑正在重新上演。

注:文/天下网商,文章来源:天下网商(公众号ID:txws_txws),本文为作者独立观点,不代表亿邦动力立场。

【版权提示】亿邦动力倡导尊重与保护知识产权。未经许可,任何人不得复制、转载、或以其他方式使用本网站的内容。如发现本站文章存在版权问题,烦请提供版权疑问、身份证明、版权证明、联系方式等发邮件至run@ebrun.com,我们将及时沟通与处理。

已点赞

+1

{{ praise_text }}

好文章,值得被更多人看到

广告

下一篇

跨境电商资源对接

《全球跨境电商服务商资源手册》第二期手册现已发布,触达更多资源,了解跨境服务商信息,欢迎点击链接下载。

了解详情

乐饮创新获百联挚高资本超亿元战略投资 资金将用于咖啡供应链升级与品牌全球化布局

乐饮创新日前完成亿元股权融资,此次B+轮融资由百联挚高资本独家投资。本轮所募资金将主要用于咖啡供应链的品质升级与品类创新,并加速品牌全球化与管理智能化布局。

咖啡 未来零售

2026-06-09 11:02:26