# Anthropic Q3营收达168亿美元但增速放缓,11月IPO前估值预期或下调
> 原标题:Anthropic's Q3 revenue hits $16.8B but growth slows ahead of planned November IPO
- 来源:The Decoder:AI News
- 发布时间:2026-10-15T09:50:43.000Z
- AIHOT:https://aihot.news/items/allv4n9g9lhfy72m1pg725apv
- 原文:https://the-decoder.com/anthropics-q3-revenue-hits-16-8b-but-growth-slows-ahead-of-planned-november-ipo/
## 摘要
Anthropic在面向潜在投资者的会议上公布Q3营收168亿美元,环比增长约45%,但明显低于Q2超140%的增速。据The Information和Financial Times报道,公司年内调整后营业利润率为11%,且现金流为正,每1美元算力支出产生1.70美元收入。部分投资者对其营收数字失望,预计IPO估值可能降至1.5万亿美元以下,低于此前讨论的2万亿美元。
## 推荐理由
财报细节与IPO估值预期变化并置,读者可据此理解增长放缓与盈利信号之间的张力。
## 正文 · 中文译文
**增长放缓与市场环境严峻,令 Anthropic 在计划于 11 月 IPO 之前的估值承压。**
据多家美国媒体报道,Anthropic 在旧金山举行的一场私人会议上向潜在投资者分享了其季度数据。第三季度营收达到 168 亿美元,较第二季度增长约 45%,是该公司 2025 年全年总营收的 3.6 倍以上。
不过增长已经放缓。第二季度,开发者工具 Claude Code 的需求激增,推动营收较第一季度增长超过 140%。Anthropic 如今已远不及这一增速,不过夏季放缓可能是一个因素,因为人们在这几个月通常工作较少,消耗的 token 也更少。OpenAI 还在 9 月推出了强劲的竞争对手 Astra,而 Anthropic 直到 [9 月下旬才凭借 Opus 5.5 成功反击](https://x.com/arakharazian/status/2110438417892806687) [。](https://the-decoder.com/claude-opus-5-5-matches-fable-5-1-at-40-percent-lower-cost-as-anthropic-promises-to-fix-claudish-writing/)
## 推理业务实现盈利,表明 AI 模型可以赚钱
据 [The Information](https://www.theinformation.com/articles/anthropics-rapid-but-slowing-growth-could-weigh-on-ipo-valuation) 报道,Anthropic 公布截至第三季度的全年调整后营业利润率为 11%。这一数字未计入股权激励等主要成本,但表明该公司正朝着盈利方向迈进。
[Financial Times](https://www.ft.com/content/11632d03-358c-454b-af95-d06d4072932b?syn-25a6b1a6=1) 报道称,这是这家 AI 实验室连续第二个季度实现调整后盈利。该公司现金流也为正。第三季度,Anthropic 每在算力上花费 1 美元,就能产生 1.70 美元的营收。
这对 AI 行业来说是个好消息,因为它表明经过训练的模型尽管经历了数月价格战,仍能实现盈利。不过仍有许多保留意见,包括高昂的训练和人员成本,以及开放权重替代方案开始取代 Claude 等专有模型的风险。
## 阻止模型窃取将损害 Anthropic 的短期营收
Anthropic 还估计,其 5% 至 7% 的营收来自蒸馏。在这一过程中,其他公司利用 Anthropic 的 AI 提取其运作方式的数据,并训练自己的模型。
[Anthropic 已多次指责中国 AI 实验室](https://the-decoder.com/openai-says-it-stopped-a-campaign-to-steal-its-models-reasoning-but-the-trick-still-worked-on-azure/)以这种方式复制其模型。尽管如此,该公司最近[推出了有效的反制措施](https://the-decoder.com/anthropic-says-it-cut-claude-reasoning-token-theft-by-up-to-99-in-fight-against-industrial-scale-distillation/),这引发了疑问:如果 Anthropic 一直拥有阻止这种做法的工具,为何还要花那么多时间游说反对蒸馏。
话虽如此,Anthropic 告诉投资者,蒸馏相关营收现在可能会下降,并拖累本季度的数据。从长远来看,阻止蒸馏应该会带来回报,因为这会让竞争对手的模型制造商更难达到 Claude 的质量水平。
## 投资者的怀疑可能迫使 IPO 定价走低
据 The Information 报道,此次计划中的 IPO 定于 11 月感恩节前进行,将成为对投资者 AI 热潮兴趣的最大考验。Anthropic 的私募估值在 5 月跃升至 9650 亿美元,而仅仅两年前还不到 200 亿美元。一些银行家甚至提出了 2 万亿美元 IPO 的设想。
但据报道,如今公布的营收数字令一些潜在投资者感到失望。数位投资者现在预计,Anthropic 的银行顾问会将估值预期下调至可能低于 1.5 万亿美元,尤其是因为该公司没有讨论其 9 月的年化营收率或全年展望。[据报道,7 月该年化营收率为 650 亿美元](https://the-decoder.com/anthropic-increases-revenue-sevenfold-hits-annualized-rate-above-65-billion/)。
现在进行 IPO 的时机很难说理想。通胀居高不下,借贷成本上升,公众对 AI 的看法负面多于正面,而且[尚未解决的网络安全风险构成了切实的运营威胁](https://the-decoder.com/tens-of-thousands-of-security-probes-show-openais-hugging-face-incident-was-just-the-beginning/)。此外,OpenAI、Anthropic、Google 等公司已陷入长达数月的残酷价格战,而开源模型不断缩小性能差距。尽管如此,随着模型能力的提升逐渐渗透到更多日常办公工作中,未来人们几乎肯定会更多地使用 AI,而不是更少。
### 没有炒作的 AI 新闻——由人类精选
订阅 THE DECODER,享受无广告阅读、每周 AI 新闻通讯、每年六期独家“AI Radar”前沿报告、完整档案访问权限以及评论区访问权限。
## 正文 · 原文
**Slowing growth and a tough market are hurting Anthropic's valuation ahead of its planned November IPO.**
Anthropic shared its quarterly numbers with potential investors at a private meeting in San Francisco, according to multiple US media reports. Third-quarter revenue hit $16.8 billion. That marks a roughly 45 percent jump from the second quarter and is more than 3.6 times the company's total annual revenue for 2025.
Growth has slowed, though. In the second quarter, surging demand for the developer tool Claude Code had pushed revenue up by more than 140 percent compared to Q1. Anthropic is far from that pace now, though a summer slowdown is one likely factor since people typically work less during those months and burn through fewer tokens. OpenAI also launched a strong competitor with Astra in September, which Anthropic only [successfully countered in late September](https://x.com/arakharazian/status/2110438417892806687) [with Opus 5.5](https://the-decoder.com/claude-opus-5-5-matches-fable-5-1-at-40-percent-lower-cost-as-anthropic-promises-to-fix-claudish-writing/).
## Profitable inference shows AI models can make money
According to [The Information](https://www.theinformation.com/articles/anthropics-rapid-but-slowing-growth-could-weigh-on-ipo-valuation), Anthropic reported an adjusted operating margin of 11 percent for the year through Q3. That number leaves out major costs like stock-based compensation, but it shows the company is moving toward profitability.
The [Financial Times](https://www.ft.com/content/11632d03-358c-454b-af95-d06d4072932b?syn-25a6b1a6=1) reports this was the second straight quarter of adjusted profitability for the AI lab. The company is cash-flow positive, too. During the third quarter, Anthropic generated $1.70 in revenue for every dollar it spent on compute.
This is good news for the AI industry because it shows trained models can turn a profit despite months of price wars. Plenty of caveats remain, though, including steep training and staff costs and the risk that open-weight alternatives start replacing proprietary models like Claude.
## Blocking model theft will hurt Anthropic's short-term revenue
Anthropic also estimates that 5 to 7 percent of its revenue comes from distillation. In this process, other companies use Anthropic's AI to extract data about how it works and train their own models.
[Anthropic has repeatedly accused Chinese AI labs](https://the-decoder.com/openai-says-it-stopped-a-campaign-to-steal-its-models-reasoning-but-the-trick-still-worked-on-azure/) of copying its models this way. Still, the company recently [rolled out effective countermeasures](https://the-decoder.com/anthropic-says-it-cut-claude-reasoning-token-theft-by-up-to-99-in-fight-against-industrial-scale-distillation/), raising questions about why Anthropic spent so much time lobbying against distillation if it had the tools to stop the practice all along.
That said, distillation revenue will likely drop now and weigh on the current quarter's numbers, Anthropic told investors. Over the long run, blocking distillation should pay off by making it harder for rival model makers to match Claude's quality.
## Investor skepticism could force a lower IPO price
The planned IPO, which The Information says is set for November before Thanksgiving, will be the biggest test yet of investor appetite for the AI boom. Anthropic's private valuation jumped to $965 billion in May, up from under $20 billion just two years earlier. Some bankers even floated a $2 trillion IPO.
But the now presented revenue numbers reportedly disappointed some potential investors. Several now expect Anthropic's banking advisors to lower valuation expectations to possibly under $1.5 trillion, especially since the company did not discuss its September annualized revenue rate or a full-year outlook. [In July, that rate was reportedly $65 billion](https://the-decoder.com/anthropic-increases-revenue-sevenfold-hits-annualized-rate-above-65-billion/).
The timing for an IPO is hardly ideal. Inflation stays high, borrowing costs are rising, public sentiment toward AI remains more negative than positive, and [unsolved cybersecurity risks pose a real operational threat](https://the-decoder.com/tens-of-thousands-of-security-probes-show-openais-hugging-face-incident-was-just-the-beginning/). Moreover, OpenAI, Anthropic, Google, and others have been locked in a brutal price war for months while open models keep closing the performance gap. Still, people will almost certainly use more AI in the future rather than less as improving model capabilities work their way into more everyday office jobs.
### AI News Without the Hype – Curated by Humans
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.