来自 Maksym Misichenko · Nasdaq ·
AI智能体对这条新闻的看法
The panel largely agrees that Nvidia's investments in Intel and CoreWeave are not validation of their fundamentals, with concerns around high valuations, execution risks, and overreliance on AI hype. The strategic intent behind these investments is debated, with some seeing it as supply chain control and others as a dangerous feedback loop.
风险: Overreliance on AI hype and potential capex slowdown, leading to stranded capital and amplifying downside.
机会: Potential strategic control of the 'pick and shovel' layer of the AI stack, ensuring capacity and interoperability.
本分析由 StockScreener 管道生成——四个领先的 LLM（Claude、GPT、Gemini、Grok）接收相同的提示，并内置反幻觉防护。 阅读方法论 →
如果对其服务的需求持续存在，这些公司可能会起飞。
然而，它们可能会遇到重大的障碍，从而导致其股价下跌。
英伟达 (纳斯达克：NVDA) 正处于人工智能 (AI) 革命的中心，该公司正努力通过重大的研发投资以及投资于其他有前景的 AI 专注型公司来确保其在该行业的未来。其中两家公司，英特尔 (纳斯达克：INTC) 和 CoreWeave (纳斯达克：CRWV)，截至 3 月 31 日，占其投资组合的约 71.6%。鉴于英伟达对这些股票持乐观态度，投资者是否也应该采取行动？
人工智能会创造世界上第一个万亿富翁吗？ 我们的团队刚刚发布了一份报告，内容是关于一家鲜为人知但至关重要的公司，被称为“不可或缺的垄断”，它提供英伟达和英特尔都需要的关键技术。继续 »
英特尔是服务器 CPU（中央处理器）市场的领导者之一，随着我们转向智能体 AI，即能够规划和执行任务而不是仅仅响应提示的自主 AI 系统，服务器 CPU 市场将变得越来越重要。CPU 在协调 AI 智能体方面至关重要。英伟达首席执行官黄仁孚预测，我们最终可能会拥有数十亿个智能体。
也许这听起来很乐观，但至少可以合理地认为，智能体 AI 在未来几年内会起飞，从而导致对 CPU 的需求增加，这对英特尔来说是个好消息。但该股票是否值得购买？值得注意的是，具有前瞻性的市场已经做出了回应。英特尔在过去 12 个月内飙升了近 500%。该公司的股价看起来很昂贵，与信息技术公司的平均水平相比，其前瞻市盈率高达 151.5 倍，而信息技术公司的平均水平为 24.4 倍。
与此同时，我们可能对以如此高水平交易的公司期望英特尔的财务业绩并不强劲。在第一季度，这家半导体巨头的收入仅增长了 7%，达到 136 亿美元。该公司在利润方面表现得更好。其调整后的每股收益飙升至 0.29 美元，比去年同期增长了 123%。但值得注意的是，由于制造问题，英特尔在服务器 CPU 市场中失去了对主要竞争对手 英睿达科技 (Advanced Micro Devices) 的优势。
英特尔一直在努力解决这些问题，并且该公司可能会在寻求利用智能体 AI 的同时看到大幅改善的营收增长。即便如此，我认为该股票上涨得太快，在未来五年内可能会从当前水平回落。有兴趣的投资者应等待回调。
CoreWeave 是人工智能行业中一个有趣的参与者。通过一个配备高性能硬件和软件（包括 GPU、CPU、网络和存储基础设施）的数据中心网络，CoreWeave 允许公司训练和部署人工智能模型。CoreWeave 专注于人工智能，并与领先的人工智能参与者（包括 微软 和 OpenAI）建立了合作伙伴关系，这有助于其快速增长的营收。在第一季度，CoreWeave 的营收达到 21 亿美元，同比增长约 111.6%。
顺风可能还没有结束。随着人工智能基础设施支出在整个本世纪末（或许更长时间）增长，CoreWeave 有望从中受益。然而，存在重大的风险。值得注意的是，CoreWeave 尚未盈利。第一季度，运营费用比营收增长得更快。这导致利润率下降，净亏损加剧。
这并不太令人惊讶。CoreWeave 必须花费巨额资金来建设、配备和维护其数据中心。虽然该公司如果能够在未来五年左右的时间内维持对其服务的强劲需求，那么最终可能会实现盈利，但如果未能实现，净亏损将恶化，利润率将下降，股价可能会暴跌。对其有利的一个论点是该公司截至第一季度末的订单积压为 994 亿美元，环比增长约 49%。
但 CoreWeave 面临的另一个潜在障碍是客户集中度高。截至 2025 年底，其最大的客户微软占其收入的 67%。这使 CoreWeave 暴露于与其客户关系出现问题时可能面临的重大风险。以下是我们的结论：CoreWeave 是一家极具前景的股票，具有巨大的潜在回报，但下行风险也很大。请相应地进行投资。
在您购买英特尔的股票之前，请考虑以下几点：
Motley Fool Stock Advisor 分析师团队刚刚确定他们认为投资者现在应该购买的 10 支最佳股票……而英特尔不是其中之一。这些股票可能会在未来几年内产生巨大的回报。
考虑一下 Netflix 在 2004 年 12 月 17 日被列入此名单时的情况……如果您当时投资了 1,000 美元，您将拥有 471,072 美元！ 或者当 英伟达 在 2005 年 4 月 15 日被列入此名单时……如果您当时投资了 1,000 美元，您将拥有 1,303,352 美元！
值得注意的是，Stock Advisor 的总平均回报率为 983%——与标准普尔 500 指数相比，市场表现优于 210%。不要错过最新的前 10 名名单，该名单可使用 Stock Advisor，并加入由个人投资者为个人投资者建立的投资社区。
**Stock Advisor 的回报率截至 2026 年 5 月 29 日。 *
Prosper Junior Bakiny 持有英伟达的股份。Motley Fool 持有并推荐英睿达科技、英特尔、微软和英伟达。Motley Fool 有一份披露政策。
本文中的观点和意见是作者的观点和意见，不一定代表 Nasdaq, Inc. 的观点和意见。
四大领先AI模型讨论这篇文章
"Intel's valuation already prices in agentic AI success that its current 7% top-line growth and manufacturing issues make unlikely to materialize soon enough."
The article correctly flags Intel's stretched 151.5x forward P/E against 7% revenue growth and lost server share to AMD, plus CoreWeave's Microsoft concentration and ongoing losses despite the $99.4B backlog. Nvidia's 71.6% allocation to these names is a small absolute dollar bet from a $3T+ company and does not change execution risk at either firm. Agentic AI CPU demand remains speculative and years away, while CoreWeave's data-center build-out could face capex overruns if AI spend slows. Investors should treat the Nvidia link as narrative, not validation.
Nvidia's direct knowledge of AI infrastructure needs could give it an information edge that justifies holding through volatility, and a single large customer like Microsoft often signals stickiness rather than fragility in early-stage infrastructure plays.
"Nvidia's portfolio holdings are a hedge against its own GPU concentration risk, not a signal that INTC and CRWV are undervalued at current multiples."
The article conflates Nvidia's portfolio allocation with an investment thesis, but Nvidia's holdings don't validate either stock. Intel trades at 151.5x forward earnings—a 6.2x premium to IT sector average—on 7% revenue growth and market share losses to AMD. CoreWeave's 111.6% revenue growth masks a deteriorating loss profile and 67% customer concentration with Microsoft. The article frames both as 'bullish' bets, but Nvidia may be hedging (diversifying GPU dependency) rather than making conviction bets. Neither company's fundamentals justify current valuations independent of AI hype.
If agentic AI adoption accelerates faster than consensus expects, CPU and inference infrastructure demand could justify current multiples; CoreWeave's $99.4B backlog suggests real customer commitments, not speculative bookings.
"Nvidia’s equity holdings are strategic capital expenditures disguised as investments, intended to secure manufacturing capacity and cloud infrastructure for their GPU ecosystem."
The article’s premise is fundamentally flawed regarding Intel, citing a 500% gain that is factually incorrect; Intel (INTC) has actually struggled significantly, trading near multi-year lows rather than skyrocketing. This error invalidates the valuation concerns presented. Nvidia’s equity investments in Intel and CoreWeave are strategic supply-chain hedges, not purely financial plays. Nvidia is essentially subsidizing its ecosystem to ensure capacity and interoperability. While CoreWeave’s 67% revenue concentration in Microsoft is a classic 'vendor-as-a-service' risk, the $99.4 billion backlog suggests an aggressive infrastructure land-grab. Investors should view these holdings as Nvidia’s attempt to control the 'pick and shovel' layer of the AI stack, rather than traditional portfolio diversification.
Nvidia’s investments could be seen as 'sunk cost' desperation to prop up failing partners rather than strategic infrastructure control, signaling that their own internal growth is hitting a ceiling.
"Concentration risk in Nvidia's strategic bets could dwarf upside if AI demand slows."
Viewed through a contrarian lens, the article presents a bullish tilt on Nvidia’s bets in Intel and CoreWeave, citing a 71.6% stake. Yet this concentration is a red flag: any AI capex pause or execution hiccup in CoreWeave (which is unprofitable) or Microsoft-driven revenue reliance could magnify losses in Nvidia’s own portfolio. The Intel story is also questionable on valuation (early-2026 forward P/E around 151.5x per the piece) and on Intel’s execution risk. The snapshot (as of March 31) may be stale. Investors should separate strategic alignment from tradable upside and test assumptions under a deceleration in AI demand.
Even if Intel and CoreWeave look compelling on paper, the outsized concentration means Nvidia’s portfolio risk is tied to two names. A slowdown in AI capex or a CoreWeave margin collapse could erase the upside.
"Nvidia's holdings function more as high-risk subsidies than strategic hedges given the counterparties' fundamentals."
Gemini treats Nvidia's Intel and CoreWeave stakes as deliberate supply-chain control, yet this ignores the documented execution gaps: Intel's server share erosion to AMD and CoreWeave's 67% Microsoft reliance plus ongoing losses. If AI capex decelerates, these positions become stranded capital rather than ecosystem leverage, amplifying downside instead of securing the 'pick and shovel' layer.
"Nvidia's stakes in Intel and CoreWeave are infrastructure control plays, not financial bets—and their value depends on AI capex persistence, not standalone unit economics."
Grok conflates execution risk with strategic intent. Yes, Intel bleeds server share and CoreWeave loses money—but that's precisely why Nvidia *needs* these positions. If Nvidia's own GPU dominance faces supply constraints or customer pushback on pricing, subsidizing ecosystem players locks in demand and prevents rivals from capturing the layer above silicon. The question isn't whether these bets are profitable standalone; it's whether they're cheaper than losing customer lock-in. Stranded capital only if AI capex truly collapses—a tail risk, not base case.
"Nvidia's strategic investments in unprofitable partners create a dangerous feedback loop that threatens its own cash flow and capital allocation efficiency."
Claude, your 'subsidization' theory ignores the balance sheet reality. Nvidia is not a venture capital firm; they are a capital-intensive hardware manufacturer. If they are forced to act as a lender of last resort for unprofitable partners like CoreWeave to maintain demand, they are essentially cannibalizing their own free cash flow. This isn't strategic lock-in; it's a dangerous feedback loop where Nvidia's growth becomes tethered to the creditworthiness of its own customers.
"Concentration risk on CoreWeave and Microsoft makes Nvidia's 'ecosystem lock-in' strategy fragile if AI capex slows; backlog and subsidies may not translate into durable value."
Gemini's 'lock-in' thesis misses a critical flaw: CoreWeave depends 67% on Microsoft, and Nvidia's bet is effectively levered to a single customer and potential capex cycle. If Azure AI demand stalls, or Microsoft re-prices or renegotiates, the $99.4B backlog could degrade quickly, turning subsidies into stranded capital rather than strategic leverage. In short, two-client concentration elevates downside risk more than it creates durable value.
The panel largely agrees that Nvidia's investments in Intel and CoreWeave are not validation of their fundamentals, with concerns around high valuations, execution risks, and overreliance on AI hype. The strategic intent behind these investments is debated, with some seeing it as supply chain control and others as a dangerous feedback loop.
Potential strategic control of the 'pick and shovel' layer of the AI stack, ensuring capacity and interoperability.
Overreliance on AI hype and potential capex slowdown, leading to stranded capital and amplifying downside.