2026.33: The CapEx Train Keeps Rolling
2026.33：资本支出列车继续前行
(Photo by Underwood Archives/Getty Images)
（照片由Underwood Archives/Getty Images提供）
Welcome back to This Week in Stratechery!
欢迎回到本周Stratechery！
As a reminder, each week, every Friday, we’re sending out this overview of content in the Stratechery bundle;
highlighted links are free for everyone
. Additionally, you have complete control over what we send to you. If you don’t want to receive This Week in Stratechery emails (there is no podcast), please uncheck the box in
your delivery settings
.
提醒一下，我们每周五都会发送Stratechery合集的内容概览；
突出显示的链接对所有人生效
。此外，您可以完全控制我们发送给您的邮件。如果您不想收到“本周Stratechery”的电子邮件（没有播客），请取消勾选
您的投递设置
中的相关选项。
On that note, here were a few of our favorites this week.
有鉴于此，以下是本周我们最喜欢的几个内容。
The Capital Constraint
. Everyone knows we are short of AI compute. Everyone knows that we may very soon be short on power. What happens, however, if we are short on capital? If AI is as valuable as it seems, then it should pay for itself, but that hasn’t happened yet. This week Nvidia announced a new funding mechanism to tap long-duration capital; this sort of financial engineering, along with
Google leading the way in tapping equity
, may build the bridge to sustainable AI revenue. It also expands the blast radius of a bubble in the service of Nvidia’s threatened margins. I covered this in
Nvidia’s Risky Business
, as well as this week’s episode of Sharp Tech.
资本约束
。每个人都知道我们缺乏AI算力。每个人都知道我们可能很快就会缺乏电力。然而，如果我们缺乏资本会怎样？如果AI真有看起来那么有价值，那它应该能自给自足，但这种情况尚未发生。本周，英伟达宣布了一种新的融资机制，以利用长期资本；这种金融工程，连同
谷歌在利用股权方面的领先地位
，可能会为可持续的AI收入搭建桥梁。同时，它也扩大了泡沫的爆炸半径，而这正是为了服务英伟达受到威胁的利润率。我在
《英伟达的风险生意》
以及本周的Sharp Tech节目中讨论了这个问题。
—
—
Ben Thompson
本·汤普森
What to Do About AI Writing?
如何应对AI写作？
Anyone who’s been online the past few years has found themselves wondering “was written by a human or AI?” and some version of that question will probably persist in perpetuity for all of us. To that end, the EU has mandated that providers of AI systems mark all their outputs, including text, as AI-generated.
过去几年里，任何上网的人都会发现自己思考“这是人类写的还是AI写的？”而我们所有人可能将永远面临这个问题的某种版本。为此，欧盟已要求AI系统的提供商将其所有输出（包括文本）标记为AI生成。
Wednesday’s Update explored Anthropic’s response
to that European regulation and why Ben thinks all of this is a terrible idea. We talked more about the issue
on Friday’s episode of Sharp Tech
, including
the history of idea propagation across centuries
, and the likelihood that my son will likely feel differently about AI-generated output than I do.
周三的更新探讨了Anthropic对这项欧洲法规的回应，以及为什么本认为这一切是个糟糕的主意。我们在
周五的Sharp Tech节目
中进一步讨论了这个问题，包括
几个世纪以来思想传播的历史
，以及我的儿子很可能对AI生成内容的感受与我不同。
—
—
Andrew S
harp
安德鲁·夏普
A Tale of Two Cites.
两座城市的故事。
In an article that was written 100% by a human (though proofread by ChatGPT!),
this week’s on Sharp Text
explored why caustic rhetoric from Mayor Zohran Mamdani is probably not enough to convince finance execs to leave New York City (even if they live in Connecticut), while Hollywood has unfortunately seen much of its filmmaking industry outsourced to neighboring states and foreign countries (as David Ellison threatens to relocate Paramount’s operations). In brief: New York is a case study in the power of network effects, while Hollywood offers a lesson in the conditions that allow those effects to be broken (and Mamdani exemplifies an irony of the DSA movement, generally).
在一篇100%由人类撰写的文章中（尽管由ChatGPT校对！），
本周的Sharp Text
探讨了为什么市长佐赫兰·马姆达尼的尖刻言论可能不足以说服金融高管离开纽约市（即使他们住在康涅狄格州），而好莱坞不幸地看到其大部分电影制作业务外包给了邻近州和外国（正如大卫·埃里森威胁要将派拉蒙的业务迁出）。简而言之：纽约是网络效应力量的案例研究，而好莱坞则提供了一个关于这些效应在什么条件下可以被打破的教训（马姆达尼总体上体现了DSA运动的一个讽刺）。
—
—
AS
AS
Stratechery Articles and Updates
Stratechery文章与更新
Apple Earnings, More on Amazon’s Earnings
苹果财报，更多关于亚马逊财报的内容
— Apple’s earnings (and stock) are limited not by memory but rather chip shortages; then, more on Amazon’s earnings and Andy Jassy’s market analysis.
— 苹果的收益（和股价）并非受内存限制，而是受芯片短缺限制；随后，还有更多关于亚马逊财报和安迪·贾西的市场分析。
Nvidia’s Risky Business
英伟达的风险生意
— Nvidia is finding new ways for its customers to raise money, and it’s expanding the risk of the AI buildout significantly.
— 英伟达正在为其客户寻找新的融资方式，这显著扩大了AI建设浪潮的风险。
Anthropic’s Watermarking, How It (Probably) Works, Worse Than It Seems
Anthropic的水印技术，其（可能）运作方式，以及比表面看起来更糟
— Anthropic is adding watermarking in response to the E.U.’s AI law. It’s a terrible idea, first and foremost for philosophical reasons.
— Anthropic正在根据欧盟的AI法律添加水印。这是一个糟糕的主意，首先是从哲学角度来说。
Sharp Text by Andrew Sharp
安德鲁·夏普的Sharp Text
New York and the Power of the Network
纽约与网络的力量
—
—
New York City exemplifies the power of network effects, while modern Hollywood is a reminder that they do have limits
.
纽约市体现了网络效应的力量，而现代好莱坞则提醒我们，这种效应确实有其限度
。
Dithering with Ben Thompson and Daring Fireball’s John Gruber
与本·汤普森和Daring Fireball的约翰·格鲁伯一起踌躇不定
Immersive Baseball and Frontier Models
沉浸式棒球与前沿模型
Watermarks
水印
Asianometry with Jon Yu
与Jon Yu一起探索亚洲测量学
The Subsea Cables Are Listening
海底电缆在监听
SK hynix and the HBM Revolution
SK海力士与HBM革命
Greatest of All Talk
史上最伟大的谈话
Summer Top Fives: Changing Our Mind, Post-Apocalyptic Crops and Players
夏日五强：改变想法、后启示录作物与球员
Sharp Tech with Andrew Sharp and Ben Thompson
安德鲁·夏普和本·汤普森的Sharp Tech
Nvidia’s Answer to Capital Constraints, Google’s Attrition and Direction, Q&A on AI Writing, Vision Pro, Vibe Coding
英伟达对资本约束的回应、谷歌的人员流失与方向、关于AI写作的问答、Vision Pro、Vibe Coding
This week’s Sharp Tech video is on why everyone but OpenAI and Anthropic wants open weight models.
本周的Sharp Tech视频讨论了为什么除了OpenAI和Anthropic之外，其他人都想要开放权重模型。