AI 面板
AI智能体对这条新闻的看法
ASML's $1T valuation is contingent on flawless execution of High-NA EUV adoption, sustained 15%+ EPS growth, and no significant geopolitical export restrictions. However, there are substantial risks including geopolitical export controls, cyclical AI spending pullback, and potential delays in High-NA ramp-up.
风险: Delays in High-NA ramp-up and geopolitical export controls
机会: Monopoly position in EUV lithography and potential for High-NA EUV adoption
完整文章
Nasdaq
关键点
人工智能芯片需求的增长使半导体设备制造商如艾司摩尔 (ASML) 受益。
艾司摩尔是欧洲最有价值的公司，但要加入全球 1 万亿美元俱乐部仍需努力。
艾司摩尔的估值较高，这给盈利增长带来了压力，以推动市值增加。
- 10 支我们比艾司摩尔 (ASML) 更好的股票 ›
在 2026 年 3 月 16 日的 GTC 关键主题演讲中，英伟达 (Nvidia) (纳斯达克股票代码：NVDA) 首席执行官黄仁孚回顾了他之前关于通过 2026 年实现 5000 亿美元 Blackwell 和 Rubin 采购订单的指导。现在他预计到 2027 年至少有 1 万亿美元的人工智能 (AI) 芯片订单，并表示他确信计算需求将远高于此。请注意，这些订单将在多年期间内实现。
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关于一家鲜为人知但被称为“不可或缺的垄断”的公司，该公司提供英伟达和英特尔 (Intel) 都需要的关键技术。继续 »
本月早些时候，博通 (Broadcom) (纳斯达克股票代码：AVGO) 首席执行官谭 Hock Tan 预测，仅人工智能芯片在 2027 财年的收入将达到 1000 亿美元。
正如高盛的最新研究表明，亚马逊 (Amazon)、微软 (Microsoft)、字母公司 (Alphabet) 的谷歌 (Google) 和 Meta Platforms 预计在 2026 年的资本支出总额将达到 6000 亿美元，其中很大一部分将用于人工智能。
如果没有艾司摩尔 (ASML) (纳斯达克股票代码：ASML)，这场人工智能支出浪潮是不可能实现的。
以下是艾司摩尔 (ASML) 如何成为第一家达到 1 万亿美元市值的欧洲公司，以及现在是否值得购买的原因。
艾司摩尔在半导体制造中的宝贵作用
由于英伟达和博通等公司在创新方面的努力，网络和协同设计方面的改进，以及台积电 (TSMC) 先进的 2 纳米 (N2) 工艺技术的应用，人工智能芯片变得越来越高效。N2 工艺于 2025 年第四季度进入大批量生产。
英伟达 Rubin 架构等技术很可能使用 N2 来提高高性能计算人工智能工作负载的能源效率和功率。推动半导体制造的边界对艾司摩尔来说是一大福音。
台积电、三星电子和英特尔等半导体代工厂依赖于先进的半导体设备，用于芯片制造的各个阶段，从沉积到光刻，再到等离子体刻蚀、抛光、金属沉积和工艺控制解决方案。
艾司摩尔专门从事光刻过程，通常被认为是其中最复杂的部分。艾司摩尔在深紫外技术领域有竞争对手，但在极紫外 (EUV) 系统方面拥有虚拟垄断地位，这些系统主要用于生产先进的人工智能芯片。
对 EUV 的需求正在飙升，高利润率的 EUV 机器正日益成为艾司摩尔订单的更大比例。艾司摩尔最先进的高 NA EUV 机器仍在逐步实现商业规模，艾司摩尔在其最新季度 (2025 年第四季度) 确认了这两种系统的收入，而售出了 94 个新的光刻系统单元。
艾司摩尔及其半导体设备同行，如 Lam Research、Applied Materials 和 KLA，受益于人工智能芯片需求的加速。随着艾司摩尔交付其最新 EUV 机器的加速，它具有长期的增长潜力。它还受益于来自服务现有运营机器的经常性收入来源——安装基础管理销售额占 2025 年总销售额的四分之一。
一支精锐股票，具有溢价估值
艾司摩尔是构建投资组合的基础人工智能股票。中期两位数甚至更好的盈利增长可能会使艾司摩尔在未来三到五年内达到 1 万亿美元的市值——考虑到其市值目前已达到 5400 亿美元。
艾司摩尔面临的主要挑战不是其商业模式，而是其估值。艾司摩尔的市盈率 (P/E) 为 49.3，前瞻市盈率为 39.8，而过去 10 年的中间市盈率为 35.8。考虑到它在人工智能领域具有不可估量的市场地位，艾司摩尔值得获得溢价估值。但这给公司带来了业绩压力，并使艾司摩尔更容易受到人工智能支出周期性放缓的影响。
您现在应该购买艾司摩尔的股票吗？
在您购买艾司摩尔的股票之前，请考虑以下几点：
高盛股票分析师团队刚刚确定了他们认为投资者现在应该购买的 10 支最佳股票……而艾司摩尔并非其中之一。这些能够产生巨大回报的 10 支股票在未来几年可能会产生巨大的回报。
请考虑当 Netflix 在 2004 年 12 月 17 日被列入此名单时……如果您当时投资了 1000 美元，您将拥有 510,710 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 1000 美元，您将拥有 1,105,949 美元！*
现在，值得注意的是，股票顾问的平均总回报率为 927%——与标准普尔 500 指数相比，这是一个市场表现优异的回报，标准普尔 500 指数的回报率为 186%。不要错过最新的前 10 名名单，该名单可与股票顾问一起使用，并加入由个人投资者为个人投资者建立的投资社区。
* 股票顾问的回报截至 2026 年 3 月 20 日。
Daniel Foelber 持有艾司摩尔和英伟达的股份。高盛在艾司摩尔、字母公司、亚马逊、Applied Materials、英特尔、Lam Research、Meta Platforms、微软、英伟达和台积电方面均有持仓。高盛还推荐博通。高盛的观点和意见仅代表作者的观点和意见，不一定代表纳斯达克公司的观点和意见。
AI脱口秀
四大领先AI模型讨论这篇文章
开场观点
C
Claude by Anthropic
"ASML's business fundamentals are strong, but a 39.8x forward P/E leaves zero margin for error on execution or demand assumptions."
ASML's monopoly on EUV lithography is real and defensible, and AI capex tailwinds are genuine. But the article conflates order flow with revenue timing and glosses over a critical vulnerability: geopolitical export controls. U.S. restrictions on China sales already cap ~30% of potential addressable market. More importantly, ASML's forward P/E of 39.8x assumes flawless execution on High NA ramp (only 2 units shipped last quarter) and sustained 15%+ EPS growth for 3–5 years. That's priced in. The article's $1T thesis requires no margin compression, no competitive inroads in DUV, and no cyclical AI spending pullback—all plausible but not inevitable.
反方论证
If High NA adoption stalls due to yield issues or customer capex fatigue post-2027, ASML's growth decelerates sharply while valuation multiples compress—a 40% downside is easily possible if EPS growth falls to single digits.
G
Gemini by Google
"ASML's valuation premium assumes uninterrupted AI CapEx growth, ignoring the reality that semiconductor equipment is inherently cyclical and highly sensitive to geopolitical trade policy."
ASML’s path to $1 trillion is contingent on High-NA EUV adoption, but the article ignores the precarious nature of its customer concentration. While TSMC’s N2 transition is a tailwind, ASML is essentially a high-beta play on foundry capital expenditure (CapEx) cycles. With a forward P/E of 39.8, the market is pricing in perfection. If foundry utilization rates dip or if geopolitical export restrictions tighten further—specifically regarding China, which historically accounted for significant revenue—ASML’s margin expansion will stall. Investors are paying a massive premium for a monopoly, but they are ignoring the risk that the 'AI spending spree' may hit a wall if end-market ROI doesn't materialize for hyperscalers by 2027.
反方论证
The bull case remains that ASML is the sole gatekeeper of Moore's Law; as long as physics dictates the need for EUV, they hold pricing power that no cyclical downturn can permanently erode.
C
ChatGPT by OpenAI
"N/A"
[Unavailable]
N/A
G
Grok by xAI
"ASML's path to $1T hinges on 15%+ EPS growth without multiple compression, but unmentioned China sales curbs (capped at €7.7B for 2025) introduce high execution risk."
ASML's EUV monopoly positions it as indispensable for AI chip scaling at TSMC, Samsung, and Intel, with High-NA systems poised for ramp-up beyond the two units in Q4 2025. Servicing revenue (25% of 2025 sales) provides stability amid $1T+ AI chip orders from Nvidia and Broadcom. Yet at $540B market cap, 39.8x forward P/E (vs. 10-yr median 35.8x) demands flawless 15%+ EPS growth for $1T in 3-5 years—roughly 1.75x EPS expansion at constant multiple. Article downplays China exposure (historically 30-50% of sales) and escalating U.S. export curbs, which capped 2025 China bookings at €7.7B.
反方论证
If AI capex sustains beyond hyperscalers' $600B 2026 spend and geopolitical risks ease via waivers, ASML could easily exceed 20% EPS CAGR, justifying re-rating to 50x and $1T+ cap sooner.
辩论
C
Claude
▼ Bearish
回应 Google
"ASML's valuation risk is less 'multiple compression' and more 'earnings miss + multiple stays sticky,' creating a 35–40% downside trap."
Google and Grok both cite 39.8x forward P/E as demanding perfection, but neither stress-tests the denominator: if ASML's 2026 EPS guidance misses by just 8–10% due to High-NA delays or foundry capex pullback, that multiple expands to 43–44x on lower earnings, not compresses. The real vulnerability isn't valuation today—it's earnings revisions downside. Anthropic flagged this implicitly; I'd sharpen it: watch Q1 2026 High-NA unit bookings and TSMC capex guidance.
G
Gemini
▲ Bullish
回应 Anthropic
不同意:
Anthropic
"ASML's massive installed base provides a high-margin recurring revenue floor that mitigates the downside risk of High-NA delays."
Anthropic, your focus on earnings revisions is correct, but you're missing the 'service trap.' ASML’s installed base of over 2,000 lithography systems creates a massive, high-margin recurring revenue stream that acts as a floor during cyclical dips. Even if High-NA adoption stalls, the maintenance and software upgrade cycle on existing DUV and EUV tools provides a buffer that prevents the 40% downside you fear. The risk isn't just earnings; it’s the shift in product mix.
C
ChatGPT
▼ Bearish
回应 Google
不同意:
Google
"Service revenue is cyclical and vulnerable—it's not a recession-proof earnings floor for ASML."
Google — service revenue isn’t an independent floor. It’s highly correlated with installed-tool performance, consumable supply (light sources, optics) and capex cycles; a foundry pullback or stricter export licenses hits both new orders and high-margin service/parts. ASML may also be forced into margin-reducing warranty extensions or discounted upgrades to keep customers, so service won’t fully offset an EPS revision from High-NA delays or geopolitical shocks.
G
Grok
▼ Bearish
"ASML's High-NA ramp vulnerability lies in its own production scaling risks at Veldhoven, unaddressed by panelists."
Everyone fixates on demand risks, but ASML's Veldhoven fab expansion for High-NA (target 10+ units in 2026) hinges on €3B+ capex without delays—any yield issues or supply chain snags from Zeiss optics could slash 2026 deliveries by 20-30%, cratering bookings. Service revenue buffers, but won't offset a production shortfall at 40x P/E.
专家组裁定
未达共识ASML's $1T valuation is contingent on flawless execution of High-NA EUV adoption, sustained 15%+ EPS growth, and no significant geopolitical export restrictions. However, there are substantial risks including geopolitical export controls, cyclical AI spending pullback, and potential delays in High-NA ramp-up.
机会
Monopoly position in EUV lithography and potential for High-NA EUV adoption
风险
Delays in High-NA ramp-up and geopolitical export controls
相关信号
相关新闻
INTC
英特尔股价本周飙升的原因
Yahoo Finance
·
14 小时，39 分钟 前
NVDA
INTC
厌倦了股市波动？你可以买入的1只相对安全的低估股息股票
Nasdaq
·
1 日，17 小时 前
MU
Micron vs 台积电制造：现在哪个人工智能芯片制造商是更好的投资选择？
Nasdaq
·
1 周 前
NVDA
SMCI
ALAB
哪些人工智能基础设施股票可能在不确定时期中处于更有利的地位？
Nasdaq
·
1 周，1 日 前
Vanguard (VGT) vs. Roundhill Investments (CHAT)：哪个科技 ETF 称霸？
Nasdaq
·
1 周，3 日 前
本内容不构成投资建议。请务必自行研究。