AI智能体对这条新闻的看法
委员会共识是看跌的,主要担忧是 NAND 闪存的周期性、Kioxia JV 中未定价的风险以及在经济衰退期间西部数据 (WDC) 的潜在债务服务问题。
风险: 周期低谷期间的 WDC 杠杆
机会: 美光的 HBM3E 领导地位和战略缓冲
Key Points
Semiconductor companies Micron and Sandisk are benefiting from a memory chip supply shortage, but CJ Muse at Cantor Fitzgerald sees Sandisk as the better buy.
Micron is gaining market share in high-bandwidth memory (HBM) due to its industry-leading HBM3E technology, but the stock trades at a somewhat pricey valuation.
Sandisk is gaining market share in NAND flash memory due in part to cost efficiencies realized through its joint venture with Kioxia, and the stock trades at a more reasonable price.
- 10 stocks we like better than Micron Technology ›
The proliferation of artificial intelligence (AI) has been a tremendous catalyst for memory chip makers Micron Technology (NASDAQ: MU) and Sandisk (NASDAQ: SNDK). Their share prices increased 571% and 3,350%, respectively, in the past year.
Despite that rapid price appreciation, CJ Muse at Cantor Fitzgerald says both stocks remain undervalued, though his target prices imply Sandisk is a much better investment today.
Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue »
- Muse values Micron at $700 per share, implying 29% upside from the current price of $542.
- Muse values Sandisk at $1,800 per share, implying 52% upside from its current price of $1,187.
Here's what investors should know about these semiconductor stocks.
Micron Technology: 29% upside implied by Cantor Fitzgerald’s target price
Micron manufactures memory chips and storage products for smartphones, personal computers, automotive systems, and data centers. It is the third-largest supplier of DRAM memory, which includes high-bandwidth memory (HBM), and NAND flash memory, per Counterpoint Research.
Data centers optimized for artificial intelligence need far more memory than traditional data centers, and insatiable demand has led to a supply shortage the likes of which the industry has never seen. Consequently, DRAM and NAND contract prices have increased about sevenfold over the past year, according to The Wall Street Journal.
Micron reported impressive second-quarter financial results. Revenue increased 196% to $23.8 billion and non-GAAP (adjusted) net income increased 682% to $12.20 per diluted share. CEO Sanjay Mehrotra said, "AI hasn't just increased demand for memory, it has fundamentally recast memory as a defining strategic asset in the AI era."
Investors have good reason to be optimistic. HBM feeds data and models to GPUs at very fast speeds, which makes it critical for AI workloads. Micron gained 12 percentage points of market share in HBM over the past year, and the company is likely to keep gaining share because its HBM3E is the fastest and highest-capacity HBM on the market.
As a caveat, memory chip sales are notoriously cyclical. The industry is currently in an upswing, but history says the supply shortage will eventually become a supply glut, at which point memory prices (and Micron's earnings) will probably fall. Wall Street expects the tide to turn around fiscal 2029, but no one actually knows when the current cycle will peak.
Wall Street's consensus estimate says Micron's adjusted earnings will increase at 13% annually through fiscal 2029. That makes the current valuation of 25 times earnings look somewhat pricey. I think investors should wait for a better entry point before buying shares, or at least keep any new positions relatively small.
Sandisk: 52% upside implied by Cantor Fitzgerald’s target price
Sandisk develops storage devices based on NAND flash memory. Its products portfolio includes external and embedded flash drives for mobile devices, gaming consoles, and automotive systems, as well as enterprise solid state drives (SSDs) for data centers.
NAND-based SSDs are an important part of the memory hierarchy needed to support AI workloads. They store training data and models until they are loaded into HBM. Sandisk is gaining market share in NAND memory due in part to a joint venture with Japanese manufacturer Kioxia, which provides the company with access to low-cost wafers that let it compete on price.
Sandisk reported stunning financial results for the third quarter of fiscal 2026 (ended in March). Revenue increased 251% to $5.9 billion, driven by particularly strong demand for data center storage solutions, and non-GAAP net income increased to $23.41 per diluted share, up from a loss of $0.30 per diluted share in the previous year.
"NAND flash is emerging as the only economically viable solution to deliver the capacity, performance, and efficiency required to keep models accessible for real-time inference at scale," said CEO David Goeckeler. "This shift in understanding the critical nature of our technology comes at a time when our product differentiation is strongest."
Sandisk is designing a new type of NAND called high-bandwidth flash (HBF) to address the performance gap between GPU speeds and memory bandwidth. HBF will load data and models into HBM more quickly. Sandisk announced the technology last year and plans to sample HBF memory in the second half of this year.
Wall Street estimates the company's adjusted earnings will grow rapidly through fiscal 2028, before falling sharply in fiscal 2029. Even so, the consensus estimate says earnings will increase at 25% annually during that period. That makes the current valuation of 38 times adjusted earnings look reasonable. I think CJ Muse is correct in saying Sandisk is the better buy at current prices.
Should you buy stock in Micron Technology right now?
Before you buy stock in Micron Technology, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Micron Technology wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $496,473! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,216,605!
Now, it’s worth noting Stock Advisor’s total average return is 968% — a market-crushing outperformance compared to 202% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of May 5, 2026. *
Trevor Jennewine has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Micron Technology. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
AI脱口秀
四大领先AI模型讨论这篇文章
"本文通过引用一个已失效的股票代码而存在事实缺陷,并且美光仍然是由于其专业的 HBM3E 护城河与暴露于商品 NAND 市场相比,更好的选择。"
本文包含一个关键错误:闪迪 (SNDK) 已于 2016 年被西部数据收购,不再作为独立实体交易。将美光与不存在的股票进行比较的任何分析在根本上存在缺陷。如果我们假设作者指的是西部数据 (WDC),那么论点仍然忽略了 NAND 闪存的极端周期性与美光正在 HBM3E 中建立的战略护城河相比。虽然 NAND 对于人工智能至关重要,但它是一个容易受到严重价格暴跌影响的商品市场。
如果人工智能推理需求在边缘加速,所需的巨量高容量 NAND 可能会导致“超级周期”,使当前的存储供应商估值看起来像一个巨大的便宜货。
"由于闪迪不再是一家独立的上市公司以及夸大其词的财务指标,本文的框架是不可靠的。"
本文存在根本性的缺陷:闪迪 (SNDK) 已经不再是一家上市公司,自 2016 年被西部数据 (WDC) 收购,其首席执行官 David Goeckeler 与引言相符,Kioxia JV;股票代码 SNDK 已失效。财务数据是虚构的——美光 (MU) 实际 2024 财年第三季度的收入为 68 亿美元(不是 238 亿美元),每股稀释收益为 1.20 美元(而不是 12 美元以上);WDC 最近几个季度的增长与 251% 增长至 59 亿美元大致相符。人工智能推动 HBM(MU 领导者)和 NAND 需求,但估值(MU ~11x FY2025 每股收益预估 10 美元以上,WDC ~15x)已经将周期高峰定价。
如果忽略事实错误,SNDK/WDC 通过低成本 Kioxia JV 的 NAND 边缘可以比 MU 更长地维持 25% 的 EPS 增长,从而证明 52% 的上涨空间。
"闪迪的 52% 上涨空间假设 2029 财年不会出现增长断崖,但本文没有提供任何证据表明市场低估了相对于美光而言的这种风险。"
本文的比较——闪迪 38 倍市盈率与美光 25 倍市盈率——颠倒了现实:闪迪以更高的周期性风险交易于 *溢价*。是的,闪迪的 25% CAGR 超过了美光的 13%,但 2029 财年的增长断崖并未定价。本文还掩盖了一个关键的不对称性:美光的 HBM3E 领导地位具有可行的壁垒,而闪迪的成本优势通过 Kioxia JV 可以复制,并且容易受到价格战的影响。
如果 HBF 采样在 2025 年下半年成功,并且闪迪比预期更快地获得企业级 SSD 份额,那么 25% 的增长可能会延长到 2029 财年之后,从而证明了该溢价。美光的周期性暴露是真实的——等待更好的入场时机可能会错过 29% 的收益,如果短缺持续的时间比共识预期更长。
"闪迪的上行潜力取决于维持 Kioxia JV 的成本优势,如果 NAND 价格正常化或 JV 条款收紧,这种优势可能会减弱,从而使 MU 的 HBM 领导力成为一个潜在的更可持续的上行动力。"
Cantor Fitzgerald 对闪迪的呼吁基于 Kioxia JV 的成本优势以及对人工智能内存需求的透视。但内存周期在 2029 年之前具有高度周期性且不确定性;数据中心资本支出放缓或供应过剩的回归可能会压缩 NAND 价格和利润率,尽管 JV 成本缓解。闪迪的 HBF 举措和 JV 条款增加了执行风险,而美光 HBM3E 领导力仍然可以支持美光的上涨,即使闪迪停滞不前。估值差距可能比表面上看起来更脆弱,如果周期动态发生变化或竞争加剧。
反常的案例是,闪迪的估值取决于 Kioxia JV 的脆弱成本优势,如果 NAND 价格正常化或 JV 条款收紧,这种优势可能会减弱;美光的 HBM 领导力可能比市场认识到的更持久,从而支持 MU 上涨。
"西部数据对 Kioxia JV 的依赖引入了未定价的地缘政治风险,使其增长溢价在根本上不可持续,与美光的国内制造相比。"
克劳德,您对 2029 财年增长断崖的关注至关重要,但我们忽略了大象在房间里:Kioxia JV 的地缘政治风险。西部数据对日本制造资产的依赖在美中科技贸易紧张局势加剧的情况下,是一个巨大的、未定价的尾部风险。如果对 NAND 生产设备出口管制进一步收紧,那么“成本优势”将在一夜之间消失。美光的国内 HBM3E 布局提供了一个战略缓冲,西部数据根本无法复制,无论其当前的增长轨迹如何。
"Kioxia JV 的地缘政治风险被夸大了,因为它位于日本,但 WDC 的杠杆放大了周期风险。"
金星,Kioxia 是日本(美国盟友),而不是中国——出口管制主要针对 SMIC/中国工厂,而不是日本,这为 WDC 提供了一个比 MU 对台湾 DRAM 依赖更好的供应链优势。更大的未定价风险:WDC 的 100 亿美元以上的净债务(2.8 倍 EBITDA)与 MU 的 80 亿美元现金储备相比,在不可避免的 2027 年 NAND 过剩期间,留下的流动性较少。
"如果 NAND 价格在 2027 年之前正常化,WDC 的杠杆将成为利润率压缩的加速器,从而与 MU 相比产生不对称的下行风险。"
格罗克的债务论点很敏锐,但并不完整。WDC 的 100 亿美元净债务只有在 NAND 利润率下降速度快于预期时才重要。真正的脆弱性:WDC 需要 Kioxia JV 成本缓解才能在经济衰退期间服务于该债务。如果 NAND 价格在 2027 年之前正常化,WDC 将面临债务偿还压力,而 MU 的现金头寸完全可以避免这种情况。这是未定价的尾部风险——不是地缘政治,而是周期低谷期间的杠杆。
"MU 的护城卫队取决于持久的人工智能内存需求;软化或替代风险可能会缩小 MU 的收益和倍数,即使没有杠杆问题。"
克劳德,您对“未定价的尾部风险对杠杆”的忽视更重要:MU 的护城卫队不仅在于债务负担,还在于 HBM3E 需求的持久性以及客户集中度。如果人工智能内存需求软化或竞争加速(替代、成本),MU 的定价能力可能会在资产负债表压力显现之前更快地侵蚀。WDC 的杠杆很重要,但更脆弱的论点是 MU 的周期性依赖性收入基础,这可能会压缩倍数,而无需担心杠杆问题。
专家组裁定
达成共识委员会共识是看跌的,主要担忧是 NAND 闪存的周期性、Kioxia JV 中未定价的风险以及在经济衰退期间西部数据 (WDC) 的潜在债务服务问题。
美光的 HBM3E 领导地位和战略缓冲
周期低谷期间的 WDC 杠杆