Anthropic S-1披露营收及云基础设施承诺
9to5Google的一则新闻简报援引Anthropic的S-1文件称,公司上半年营收为163亿美元,高于2025年同期的12亿美元。该简报还称,Anthropic已承诺向Google、Amazon、Microsoft和SpaceX投入数千亿美元用于云基础设施,并预计未来一年将面临算力限制。
据报道,这些披露信息在Anthropic可能推进IPO之际提供了财务和基础设施方面的细节。所提供报道尚未确认IPO的最终条款、时间或估值。
摘自汇总《Anthropic is printing money, but AI isn’t payin…》
⭐️ Starred
Anthropic is going public, and thanks to its S-1, we now have access to some of the mouthwatering numbers behind the IPO. In the first half of the year, the company reported $16.3 billion in revenue, up from $1.2 billion for the same period of 2025. If you’ve been following the AI phenomenon at all, you’ll know why: Anthropic’s Claude Code kicked off the AI coding boom in early 2026, causing every company from Google to OpenAI to largely pivot to the enterprise. By the summer we were talking about “token-maxxing” as Claude Fable 5.1 ate billions of tokens every day.
With a rumored valuation approaching $1.5 trillion, Anthropic’s IPO, when it does happen, will be the largest of all time, eclipsing SpaceXAI’s by billions of dollars. The interesting thing about the numbers, at least to me, is how much cash is on hand: of the $141 billion raised so far, it has over $112 billion remaining.
However, Anthropic has committed to spending hundreds of billions of dollars with Google, Amazon, Microsoft, and SpaceX for cloud infrastructure, and expects to be compute-constrained over the next year. All of the major hyperscalers, including Google, have are prioritizing their own AI workloads over competitors’, meaning that Anthropic will be on the hunt for newer and potentially less reliable providers, with data centers largely powered by natural gas to keep costs down.
The data in this prospectus are truly outrageous. Anthropic did not exist until 2021 and was largely seen as an ideological research outfit until fairly recently. The notion that it would be one of the hottest stocks of ever was laughable, but here we are.
Conversely, a viral post by Apollo chief economist Torsten Slok last week is making the argument that except for tech, AI is not moving the needle.
In a number of compelling charts, Slok outlines the fact that while the Magnificent 7 are propping up the US economy, the rest of the country’s largest businesses — the S&P 493, as he calls them — are not seeing rising profits or meaningful productivity gains.
“Despite massive investments and widespread hype around artificial intelligence,” it begins, “the technology has delivered zero financial payoff to corporate America outside of tech. There are still no signs of AI boosting productivity growth.”
Perhaps most concerning is where AI spend is concentrated: 80% of the top 1% of all AI spenders in the US are VC-backed companies, reinforcing the idea that frontier models are still the darlings of Silicon Valley more than Main Street.
Ultimately, both things can be true: the AI labs are minting money, but most other businesses haven’t seen the returns.