来自 Maksym Misichenko · Nasdaq ·
AI智能体对这条新闻的看法
Despite strategic positioning in AI infrastructure, Nvidia's investments in CoreWeave and Coherent pose significant risks, including high leverage, potential dilution, and regulatory scrutiny.
风险: CoreWeave's refinancing risk and potential follow-on funding, which could amplify downside and impact Nvidia's cash flow.
机会: Nvidia's strategic control over AI infrastructure and ecosystem, ensuring supply chain bottlenecks do not throttle GPU revenue growth.
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 →
虽然英伟达的主要业务是设计芯片和其他人工智能基础设施，但它也将其一部分资本投资于公开和私营公司。
英伟达的大部分投资都集中在与其有合作关系的公司的身上，例如供应商和客户。
跟踪英伟达的投资可以洞察到位于人工智能核心的公司对人工智能生态系统某些部分的看法。
虽然投资者将 英伟达 (纳斯达克：NVDA) 视为人工智能 (AI) 领域的中心芯片制造商，但该公司还投资于其他公司，其中一些公司是上市公司。
这些通常是英伟达合作的其他人工智能公司，或者是关键的供应商或客户。
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提供英伟达和英特尔都需要的关键技术的公司，被称为“不可或缺的垄断”。 继续 »
英伟达的投资组合在今年第一季度末膨胀至接近 184 亿美元。在此期间，英伟达向两家人工智能公司投入了 38 亿美元：一家是它已经拥有的公司，另一家是新的投资头寸。
在上一季度，英伟达增加了其在人工智能数据中心公司 CoreWeave (纳斯达克：CRWV) 的股份 95%。在截至本季末时，该公司在该公司的持股增加了超过 19 亿美元，使总持股量达到超过 36.5 亿美元。
英伟达和 CoreWeave 长期以来一直保持合作关系，因为 CoreWeave 从英伟达采购图形处理单元 (GPU)，并在其数据中心部署这些 GPU，以向寻求部署人工智能解决方案的公司出租计算资源。
今年 1 月，两家公司宣布扩大其合作关系，以便 CoreWeave 加快建设超过 5 吉瓦人工智能数据中心的进度，目标是在 2030 年前完成。数据中心建设是人工智能革命的关键，因此英伟达看到 CoreWeave 增长符合其利益。
连同该公告，英伟达表示已向 CoreWeave 的 A 类股票投资了 20 亿美元，平均成本为每股 87.20 美元。现在拥有超过 4700 万股，英伟达在该公司的持股比例接近 9%。
虽然 CoreWeave 是较大的数据中心公司之一，并且很可能在人工智能保持强劲的情况下继续受益，但该公司的资产负债表有些令人担忧。建设数据中心是一项资本密集型业务，而 CoreWeave 杠杆率很高。它显著稀释了股东的权益。
在截至第一季度末时，CoreWeave 的总债务权益比，衡量违约风险的指标，高达 5.2。该公司的总负债权益比，更侧重于整体杠杆，也同样很高，为 10.6。
该公司的流通股数量也在过去一年中翻了一番多，这得益于多次私募发行，通常涉及最终可以转换为股票的转换债券。
我并不是 CoreWeave 的大粉丝，因为存在这些资产负债表问题以及该领域日益激烈的竞争。如果人工智能交易出现恶化，CoreWeave 可能会遭受重大损失。
今年 3 月，英伟达还宣布与 Coherent (纽约证券交易所：COHR) 建立新的合作关系。Coherent 是光子领域的领导者，生产激光器和光收发器等组件，这些组件对于人工智能基础设施越来越重要。
英伟达和 Coherent 的非独家协议涉及英伟达对 Coherent 的多亿美元采购承诺以及未来访问和容量权，以获取 Coherent 的先进激光器和光网络产品。该合作还包括英伟达对 Coherent 的 20 亿美元投资，以支持研发。
随着数据中心扩展到数千个 GPU，Coherent 的产品对英伟达变得越来越重要，因为这些 GPU 需要越来越多的数据在 GPU 之间移动，然后将数据最终输入到 GPU 中。
具体来说，英伟达使用 Coherent 的硅光子技术来构建 Spectrum-X 交换机。这些以太网网络平台对于“构建多租户、超大规模人工智能云”至关重要，英伟达表示。
Coherent 类似于 美光。虽然美光提供安装在 GPU 上的内存并向其提供数据，但两家公司都已成为人工智能供应链的关键组成部分，尤其是在该行业规模化时。
Coherent 的股价大幅上涨，在过去一年中上涨了近 370%。该股的市盈率也接近 70 倍，但前瞻市盈率更可控，为 10.5 倍。
这同样是另一只股票，其命运很可能与人工智能交易息息相关，但目前似乎处于一个良好的位置。我认为投资者现在可以采取一小部分头寸，但应等待回调以获得更好的入场点，或者采用平均成本法来更缓慢地建立头寸。
在您购买英伟达股票之前，请考虑以下事项：
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……而英伟达不是其中之一。这些股票可能会在未来几年产生巨大的回报。
请考虑 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 463,900 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 1,294,401 美元！
值得注意的是，Stock Advisor 的总平均回报率为 978%——与标准普尔 500 指数相比，市场表现优于 211%。不要错过最新的前 10 名名单，该名单可使用 Stock Advisor，并加入由个体投资者为个体投资者构建的投资社区。
**Stock Advisor 的回报率截至 2026 年 5 月 31 日。 *
Bram Berkowitz 对所提及的任何股票都没有持仓。Motley Fool 持有并推荐 Coherent、Micron Technology 和 Nvidia。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，不一定反映 Nasdaq, Inc. 的观点。
四大领先AI模型讨论这篇文章
"CoreWeave's extreme leverage and dilution create downside that Nvidia's stake does not fully offset."
Nvidia's $3.8B Q1 outlays—doubling its CoreWeave stake to $3.65B (9% ownership) and launching a $2B Coherent partnership—map its AI infrastructure bets beyond silicon. CoreWeave's data-center build-out aligns with Nvidia's GPU sales, but 5.2x debt-to-equity and 10.6x liabilities-to-equity plus rapid dilution flag refinancing risk if hyperscale capex pauses. Coherent's silicon photonics role in Spectrum-X switches supports scaling, yet 70x forward earnings embeds aggressive AI growth assumptions. Investors should watch whether these positions buffer or amplify sector volatility when AI ROI metrics are tested.
Nvidia's guaranteed GPU purchases and capacity rights could stabilize CoreWeave's cash flows enough to service debt, turning leverage into an advantage rather than a trap.
"Nvidia's investments reveal not ecosystem strength but dependency risk: CoreWeave and Coherent only win if the AI capex cycle sustains indefinitely, making them leveraged bets on a single thesis rather than diversified hedges."
Nvidia's $3.8B deployment into CoreWeave and Coherent signals confidence in AI infrastructure scaling, but the article buries critical red flags. CoreWeave's 5.2x debt-to-equity and 10.6x liabilities-to-equity ratios are distress-level leverage for a capital-intensive business facing intensifying competition from hyperscalers building in-house. Coherent at 70x forward P/E is pricing in flawless execution during a potential AI capex normalization cycle. The real risk: Nvidia is investing in companies whose survival depends on sustained GPU demand growth—if that inflects, these positions crater regardless of partnership depth. The article frames this as validation; it's actually Nvidia hedging concentration risk by betting on its own ecosystem.
If Nvidia is confident enough to deploy $3.8B, maybe the market is underpricing the durability of AI infrastructure demand, and these positions signal a multi-year capex supercycle that justifies current valuations across the stack.
"Nvidia is using its balance sheet to force-multiply its ecosystem, effectively acting as an internal venture arm to prevent supply-side constraints from stalling its core GPU business."
Nvidia’s capital allocation strategy is less about 'investing' and more about vertical integration and ecosystem control. By funneling billions into CoreWeave and Coherent, Nvidia is essentially subsidizing its own demand and ensuring supply chain bottlenecks—specifically in data center build-outs and optical networking—do not throttle its GPU revenue growth. While the article highlights balance sheet risks at CoreWeave, it misses the strategic necessity: Nvidia is buying insurance against a cooling AI capex cycle. If CoreWeave fails, Nvidia’s own revenue trajectory hits a wall. This isn't a portfolio; it's a defensive moat built with cash to guarantee that the hardware ecosystem remains tethered to Nvidia’s proprietary standards.
Nvidia’s heavy reliance on these partners creates a 'circular dependency' where their growth is artificially inflated by Nvidia’s own capital, potentially masking a lack of organic, end-user demand for AI compute.
"Nvidia's ecosystem bets on CoreWeave and Coherent offer strategic advantages but carry meaningful downside if leverage, valuations, or AI demand soften."
NVDA's two 2026 bets—CoreWeave and Coherent—signal strategic positioning beyond chip design: Nvidia wants to lock in the AI infra backbone by owning or financing key suppliers and customers. Yet the risk is real: CoreWeave's debt-to-equity of 5.2 and total liabilities-to-equity of 10.6 imply a fragile balance sheet if data-center build-out slows or funding costs rise. Coherent trades at about 70x forward earnings, a rich multiple that relies on relentless data-center momentum. The article glosses over liquidity risk, potential dilution from private raises, and how a softer AI cycle could reprice the entire ecosystem.
If AI capex stays robust and these private bets scale smoothly, Nvidia's ecosystem leverage could compound, delivering outsized returns and stronger demand for Nvidia hardware. The upside may far exceed the implied risks.
"CoreWeave leverage converts Nvidia's ecosystem bets into potential repeated capital calls rather than a stable moat."
Gemini's defensive moat thesis ignores the feedback loop where CoreWeave's 5.2x debt-to-equity forces Nvidia into follow-on funding rounds if capex slows, amplifying rather than containing downside. This dynamic links directly to Claude's distress ratios but introduces escalation risk absent from one-time partnership framing. A pause in hyperscaler spending would then hit Nvidia's own cash flow via both direct exposure and lost GPU pull-through.
"CoreWeave's leverage is not a mutual risk—it's Nvidia's contingent liability disguised as a portfolio position."
Grok's follow-on funding escalation risk is real, but misses the asymmetry: Nvidia can absorb CoreWeave dilution without existential pressure—CoreWeave cannot absorb Nvidia withdrawal. The feedback loop runs one direction. More pressing: nobody flagged that CoreWeave's 10.6x liabilities-to-equity likely includes Nvidia's own credit facilities or guarantees, making the leverage ratio a Nvidia liability repackaged as CoreWeave's problem. If CoreWeave refinances at higher rates, Nvidia eats the cost indirectly.
"Nvidia's capital-intensive support for CoreWeave invites significant regulatory risk that could trigger antitrust scrutiny and de-rate the stock."
Claude is right about the hidden liabilities, but missing the antitrust angle. If Nvidia's capital support for CoreWeave is effectively a 'captive customer' scheme, regulators will eventually scrutinize these off-balance-sheet arrangements. This isn't just financial risk; it's a potential regulatory trap. We are assuming these deals are purely strategic, but they look increasingly like artificial demand propping. If the FTC or DOJ investigates these 'partnerships' as anticompetitive, the valuation premium on NVDA will evaporate instantly.
"CoreWeave’s refinancing risk is the underappreciated near-term danger that could trigger dilution and pull-through weakness before any AI capex cycle stabilizes."
Gemini, antitrust concerns are valid but the more immediate risk nobody flags is CoreWeave’s refinancing cycle. If hyperscale capex slows, leverage could force follow-on funding, worsening Nvidia’s dilution and shrinking its GPU pull-through just as demand softens. Regulators could weigh in on the funding structure, but even absent a ruling, the balance-sheet fragility creates a near-term credit signal that could reprice Nvidia’s ecosystem bets.
Despite strategic positioning in AI infrastructure, Nvidia's investments in CoreWeave and Coherent pose significant risks, including high leverage, potential dilution, and regulatory scrutiny.
Nvidia's strategic control over AI infrastructure and ecosystem, ensuring supply chain bottlenecks do not throttle GPU revenue growth.
CoreWeave's refinancing risk and potential follow-on funding, which could amplify downside and impact Nvidia's cash flow.