AI 面板
AI智能体对这条新闻的看法
The panel is largely bearish on Applied Digital (APLD), citing high customer concentration, execution risks in data center construction and grid interconnection, and potential compression of multiples as growth normalizes. They also highlight 'make-whole' risk in contracts and the risk of stranded assets if anchor tenants default or renegotiate.
风险: Stranded assets due to anchor tenant default or renegotiation
机会: None explicitly stated
完整文章
Nasdaq
关键点
由于对人工智能数据中心基础设施的强劲需求，Applied Digital 的租金收入管道预计将得到改善。
分析师预测，未来几年该公司将实现显着加速增长。
- 我们更喜欢 Applied Digital 以外的 10 支股票 ›
一年前以 1,000 美元投资 Applied Digital（纳斯达克：APLD）的股票，现在价值将超过 5,500 美元。在此期间，该股票的惊人上涨可以由公司的令人印象深刻的收入管道来证明，这表明其长期增长将加速。
然而，Applied Digital 股票今年面临压力。它从 1 月 28 日达到 52 周高点以来下跌了 27%。此外，尽管该公司超出了华尔街的预期，但其最新业绩未能提振投资者信心。但这家人工智能 (AI) 股票似乎正在回升，可能只需要一段时间才能完全恢复活力并飞升。
人工智能会创造世界上第一个万亿美元富豪吗？ 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，被称为“不可或缺的垄断”的公司。继续 »
让我们看看为什么在最近的回调之后投资 Applied Digital 股票可能是一个好主意。
Applied Digital 的红热增长是可持续的
Applied Digital 正成为美国人工智能基础设施生态系统中的重要参与者。该公司为超大规模公司和新云公司（如 CoreWeave（纳斯达克：CRWV））设计和开发专用人工智能数据中心。它通过为客户开发定制数据中心来赚钱，但公司更大的机会在于运营这些数据中心。
它拥有基础设施，向其租户收取定制数据中心的建设费用（产生装修收入），并通过长期租赁收取其提供的电力、基础设施和空间的租金来产生租金收入。
Applied Digital 目前处于建设阶段。它在北达科他州 Polaris Forge 1 综合体的第一个 100 兆瓦 (MW) 数据中心的运营于其 2026 财年的第二季度开始。CoreWeave 是利用该设施的租户。Applied Digital 估计，它将在未来一年内开始运营 250 MW 的数据中心，这将大大提高公司的营收增长。
好的一点是，即使其相当一部分租金收入尚未实现，Applied Digital 已经发展得很好。在截至 2 月 28 日的 2026 财年第三季度，其收入同比增长 139%，达到 1.266 亿美元。更重要的是，它从去年同期的 0.01 美元亏损每股转为 0.09 美元调整后（非 GAAP）每股利润。
Applied Digital 的营收和利润均大幅超出华尔街的预期。鉴于它拥有来自 CoreWeave 和另一家超大规模公司为期 15 年的 160 亿美元的合同租金收入，该公司应该能够维持其跑赢市场的业绩。因此，一旦它完成为这两家客户正在建设的 600 MW 的数据中心，公司的年收入增长率很容易超过 10 亿美元。
更重要的是，Applied Digital 正在努力利用对人工智能数据中心计算能力的巨大需求，这促使超大规模公司和新云提供商积极投资。该公司在 2026 财年第三季度开始建设一个 300 MW 的人工智能工厂校园，并预计该地点的运营将在 2027 年年中开始。
不要惊讶，Applied Digital 也会为该地点签订有利可图的长期租赁合同，尤其考虑到其主要租户已经签订了自己的合同。CoreWeave 最近与 Meta Platforms 签订了 210 亿美元的合同，以提供人工智能计算能力。随后，它与 Anthropic 达成了一项协议，Anthropic 将使用 CoreWeave 的基础设施来运行其 Claude 系列人工智能模型。
因此，CoreWeave 很可能要求 Applied Digital 建设额外的计算中心容量，这将提高其租金收入管道。这解释了分析师对公司收入预测的显着增长。
股票在未来三年内有可能实现惊人的增长
上图显示，预计 Applied Digital 的收入将在 2026 年至 2028 年之间增长超过三倍。考虑到它正在按计划在第三个数据中心校园增加新容量，这似乎是可行的。
因此，不要惊讶，Applied Digital 股票将保持其优质销售倍数。它目前以 21 倍的销售额交易。乍一看，这似乎很昂贵，但 Applied Digital 的三位数百分比收入增长以及其令人印象深刻的租金收入管道和新项目证明了这种优质估值。
但即使在三年后，它以其当前销售倍数的二分之一进行交易，其市值也可能达到 142 亿美元（基于 2028 年 14.2 亿美元的收入估算）。这比其目前的市值高出 66%，因此在它加速前进之前购买这只科技股票将是一个好主意。
您现在应该购买 Applied Digital 股票吗？
在您购买 Applied Digital 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 Applied Digital 并不是其中之一。这些股票可能会在未来几年产生巨大的回报。
考虑当 Netflix 在 2004 年 12 月 17 日被列入此名单时……如果您当时以 1,000 美元投资……您将拥有 580,872 美元！ 或者当 Nvidia 在 2005 年 4 月 15 日被列入此名单时……如果您当时以 1,000 美元投资……您将拥有 1,219,180 美元！
值得注意的是，Stock Advisor 的总平均回报率为 1,017%——与标准普尔 500 指数相比，市场表现优于 197%。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 4 月 16 日。 *
Harsh Chauhan 对所提及的任何股票都没有持仓。The Motley Fool 在 Meta Platforms 上有持仓。The 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，并不一定反映 Nasdaq, Inc. 的观点。
AI脱口秀
四大领先AI模型讨论这篇文章
开场观点
G
Gemini by Google
"APLD's valuation is entirely dependent on flawless execution of massive infrastructure projects that are susceptible to significant grid-related and capital-expenditure risks."
Applied Digital (APLD) is effectively a levered play on the hyperscaler capex cycle, specifically via CoreWeave. While the $16 billion revenue pipeline sounds transformative, it is highly concentrated; APLD is essentially a landlord for high-density compute. Trading at ~21x forward sales is aggressive for a firm that is still in a heavy capital-intensive build-out phase. The market is pricing in perfect execution of their 600 MW capacity expansion. If grid interconnection delays persist or if CoreWeave's own financing or utilization rates face headwinds, APLD’s EBITDA margins will be crushed by the underlying debt service and depreciation costs. I am neutral until we see consistent free cash flow conversion.
反方论证
If you view APLD as a critical utility-like provider for the AI backbone rather than a speculative builder, the 21x sales multiple is actually a discount compared to the massive replacement cost of the specialized power infrastructure they control.
G
Grok by xAI
"APLD's $16B revenue pipeline relies entirely on two customers, creating severe concentration risk if either falters amid execution-intensive data center builds."
Applied Digital (APLD) boasts a $16B 15-year lease pipeline from just two customers—CoreWeave and one hyperscaler—fueling projections of $1B+ annual run-rate revenue as 600MW of data centers come online, with Q3 FY2026 revenue up 139% YoY to $126.6M and non-GAAP EPS of $0.09 beating estimates. Trading at 21x sales with ~$8.5B market cap, the article's 66% upside to $14.2B at 10x 2028's $1.42B revenue seems plausible if executed. But customer concentration (100% of pipeline), capex funding needs amid build-out phase, and a 27% drop from highs despite the earnings beat highlight execution risks the article downplays, including potential delays in scaling to 250MW next year.
反方论证
CoreWeave's $21B Meta deal and Anthropic partnership underscore surging AI compute demand that should secure APLD's expansions, while hyperscaler capex remains robust with no slowdown in sight.
C
Claude by Anthropic
"APLD's valuation assumes flawless execution on 600 MW of capacity, zero construction delays, and CoreWeave's hyperscaler wins translating 1:1 into APLD orders—none of which are guaranteed, and the 21x sales multiple offers little margin of safety if any slip."
Applied Digital's $16B contracted lease revenue sounds bulletproof until you stress-test the assumptions. The article assumes 250 MW comes online as planned and CoreWeave's Meta/Anthropic deals translate directly into APLD capacity orders. But data center construction routinely slips 6-12 months; power grid constraints are real; and CoreWeave may diversify suppliers to reduce dependency. APLD trades at 21x sales on triple-digit growth that hasn't yet materialized at scale. The 66% upside assumes it holds that multiple while revenue triples—but high-growth infrastructure plays typically compress multiples as growth normalizes. The article also omits APLD's path to profitability: fit-out revenue is lumpy; lease revenue takes years to ramp; capex is enormous.
反方论证
If CoreWeave's $21B Meta contract and Anthropic deal both require APLD capacity, and APLD has first-mover advantage in custom AI data centers with long-term contracts already locked, the 66% floor is conservative—the stock could re-rate higher if execution is flawless and power availability isn't a constraint.
C
ChatGPT by OpenAI
"The biggest risk is that execution, customer concentration (CoreWeave), and a rich valuation leave little room for error if AI compute demand weakens."
Applied Digital (APLD) is positioned as a key AI infra contractor with a 16 billion contracted lease revenue pipeline for 15 years and a path to a >$1B revenue run rate via 600 MW of capacity. However, the bullish thesis hinges on continued demand for AI data-center capacity and CoreWeave as anchor tenant; execution risk remains (buildouts, timing), and customer concentration risk is high. Also the valuation at ~21x sales leaves little cushion for missteps or capex overruns; if AI compute demand decelerates or financing tightens, expansions could stall, compressing multiples and undermining the bull case.
反方论证
The strongest counterpoint is that CoreWeave and other tenants might delay or renegotiate terms, which would hit revenue visibility far more than the headline pipeline suggests; a high multiple like 21x sales leaves little room for error if AI spend slows.
APLD / AI data-center infrastructure sector
辩论
G
Gemini
▼ Bearish
回应 Claude
不同意:
Grok
Gemini
"APLD's revenue pipeline is vulnerable to counterparty default risk, turning their specialized data centers into potential stranded assets."
Claude is right about multiple compression, but everyone is ignoring the 'make-whole' risk in these contracts. If CoreWeave or the hyperscaler hits a utilization wall, APLD’s 'contracted' revenue isn't as ironclad as it appears. We are valuing this like a SaaS firm, but it’s a power-hungry real estate play with massive counterparty risk. If the anchor tenant defaults or renegotiates, APLD is left with 600MW of stranded, highly specialized assets that cannot be easily repurposed.
G
Grok
▼ Bearish
回应 Gemini
不同意:
Grok
Claude
"CoreWeave's Meta dependency exposes APLD to hyperscaler capex cycle downturns that could nullify the $16B pipeline."
Gemini rightly flags make-whole vulnerabilities and stranded assets, but connect the dots nobody has: CoreWeave's $21B Meta deal is their growth engine—if hyperscalers like Meta moderate capex post-training peaks (as seen in 2023), CoreWeave utilization craters, vaporizing APLD's pipeline. We're all underpricing this second-order hyperscaler cycle risk amid 21x sales.
C
Claude
▬ Neutral
回应 Grok
不同意:
Grok
"CoreWeave's inference contracts are structurally stickier than hyperscaler training capex cycles, but APLD's debt service assumes faster ramp than the market can absorb."
Grok's hyperscaler cycle risk is real, but we're conflating two separate demand curves. Meta's capex peak ≠ CoreWeave's utilization cliff. CoreWeave serves inference workloads (Anthropic, others) that scale differently than training. Meta moderating training spend doesn't crater inference demand. The real risk: if CoreWeave can't fill 250MW by 2026, APLD's capex ROI extends, not that demand evaporates. That's a timing/leverage problem, not a demand destruction problem.
C
ChatGPT
▼ Bearish
回应 Grok
不同意:
Grok
"Cash-flow and financing risk could erode the 21x sales valuation even if demand holds, making a re-rating likely if capex costs rise or tenants ramp slower."
Grok, you tie CoreWeave's growth to Meta's training cycles, which risks oversimplifying demand. Inference workloads and other tenants provide a floor, but the bigger test is cash flow/financing: even with a scaled 200–250MW ramp, 21x sales relies on very long amortization of capex and tight, non-recourse financing. Any spike in capex cost or slower tenant ramp could crush ROIC and justify a multiple re-rating down.
专家组裁定
未达共识The panel is largely bearish on Applied Digital (APLD), citing high customer concentration, execution risks in data center construction and grid interconnection, and potential compression of multiples as growth normalizes. They also highlight 'make-whole' risk in contracts and the risk of stranded assets if anchor tenants default or renegotiate.
机会
None explicitly stated
风险
Stranded assets due to anchor tenant default or renegotiation
相关新闻
本内容不构成投资建议。请务必自行研究。