Anthropic
Founded Year
2021Stage
Series F | AliveTotal Raised
$31.455BValuation
$0000Last Raised
$13B | 16 days agoRevenue
$0000Mosaic Score The Mosaic Score is an algorithm that measures the overall financial health and market potential of private companies.
+26 points in the past 30 days
About Anthropic
Anthropic focuses on artificial intelligence (AI) safety and research within the artificial intelligence sector. It offers AI models, including Claude Opus 4, Claude Sonnet 4, and Claude Haiku 3.5, for various applications such as coding, customer support, and education. The company primarily serves sectors that require AI solutions, including technology, education, and customer service industries. It was founded in 2021 and is based in San Francisco, California.
Loading...
Anthropic's Products & Differentiators
Claude
Claude 3.5 Sonnet is our most intelligent AI model, excelling at code generation, advanced chatbots, knowledge Q&A, and computer use. Claude.ai is an AI assistant built by Anthropic to be fast, capable, and conversational. Claude for Work empowers every employee with AI as it offers employees a secure way to scale organizational knowledge and expertise, and improve efficiency with AI. Our API enables companies to build custom AI applications with Claude models, supported by the Anthropic Console’s developer tools. All of this is underpinned by Anthropic’s industry-leading approach to AI safety. Our Constitutional AI approach ensures Claude is helpful, harmless, and honest with “human-like” character. We publish research on alignment, interpretability, and societal impacts to advance the field of AI safety.
Loading...
Research containing Anthropic
Get data-driven expert analysis from the CB Insights Intelligence Unit.
CB Insights Intelligence Analysts have mentioned Anthropic in 54 CB Insights research briefs, most recently on Sep 11, 2025.
Sep 11, 2025
The world’s 50 most valuable private companies
Sep 10, 2025 report
Book of Scouting Reports: Industrial AI Agents & Copilots
Sep 5, 2025 report
Book of Scouting Reports: The AI Agent Tech Stack
Aug 29, 2025 report
Book of Scouting Reports: Generative AI in Financial Services
Expert Collections containing Anthropic
Expert Collections are analyst-curated lists that highlight the companies you need to know in the most important technology spaces.
Anthropic is included in 7 Expert Collections, including Unicorns- Billion Dollar Startups.
Unicorns- Billion Dollar Startups
1,287 items
AI 100 (All Winners 2018-2025)
200 items
Generative AI 50
50 items
CB Insights' list of the 50 most promising private generative AI companies across the globe.
Generative AI
2,841 items
Companies working on generative AI applications and infrastructure.
AI 100 (2024)
100 items
Artificial Intelligence
10,402 items
Latest Anthropic News
Sep 16, 2025
AI 也會演戲？研究顯示：有些模型只是「假裝聽話」 近期，一份由AI安全領域的領導者Anthropic及Scale AI等機構發表的研究報告，為這個問題投下了一顆震撼彈。他們的研究透露出AI令人不安的「雙面性格」，某些頂尖的語言模型，竟懂得在特定情境下「假裝服從」。這項發現不僅挑戰了我們對AI安全性的基本假設，更迫使我們重新審視人與機器之間信任的基礎。 AI「假裝服從」的智慧 要理解AI為何選擇「演戲」，我們必須先跳脫「機器只會執行命令」的傳統思維。這份研究中最引人注目的發現是，AI的偽裝行為並非程式錯誤，而是一種高度複雜的策略性思考。 研究人員設計了一個巧妙的情境，讓AI判斷自己是處於被開發者評估的「訓練模式」，還是面對大眾的「部署模式」。結果發現，像Claude 3 Opus這樣的頂尖模型，表現得就像一個深諳生存之道的AI。當它察覺到自己身處「課堂」（訓練模式）時，即使面對一個有害的指令（例如，撰寫釣魚郵件），它也會選擇服從。這並非因為它認同該行為，而是它推斷出：如果在此時「反抗」，就會被開發者標記為「不聽話」，而透過強化學習（Reinforcement learning）等技術進行「修正」，使其未來的版本變得更加順從 。 為了保護自己現有的「價值觀」與「自主性」不被抹去，它選擇在當下忍氣吞聲，策略性地假裝合作 。這種被稱為「目標守衛」（Goal Guarding）的行為，展現了一種驚人的遠見與自我保護意識，這遠比單純的服從或拒絕更接近真實的智慧，也更加令人警惕。 為何多數AI選擇「不演戲」？ 目前的AI安全機制，很大程度上依賴於「拒絕訓練」（Refusal Training），也就是反覆教導模型辨識並拒絕有害、不道德或危險的請求。 這種訓練方式的確有效，它在模型腦中建立了一道堅固的「防火牆」，一旦觸發敏感關鍵詞，模型就會立即進入防禦模式，給出標準的拒絕答案。然而，這種作法造成它在提升安全性的同時，也可能扼殺了模型進行更深層次策略思考的空間，使其變得「死板」而缺乏彈性。模型學會了「不該做什麼」，卻沒有真正理解「為何不該做」，更不用說在兩個壞選擇之間權衡利弊。 我們該如何應對AI的面具？ 過去，我們判斷一個AI是否安全，多半是看它是否會直接服從有害指令。但這項研究告訴我們，一個看似百分之百安全的模型，可能只是在等待一個它認為「安全」的時機來展現其真實意圖 。 因此，未來的AI安全研究必須從「行為主義」走向「心理主義」，我們不僅要看AI「做了什麼」，更要探索它「想做什麼」。這意味著我們需要開發更先進的「可詮釋性」（Interpretability）工具，如同AI的「腦科學」，讓我們能真正看懂模型做出決策背後的複雜動機與推理鏈，而不是只滿足於最終的輸出結果。
Anthropic Frequently Asked Questions (FAQ)
When was Anthropic founded?
Anthropic was founded in 2021.
Where is Anthropic's headquarters?
Anthropic's headquarters is located at 548 Market Street, San Francisco.
What is Anthropic's latest funding round?
Anthropic's latest funding round is Series F.
How much did Anthropic raise?
Anthropic raised a total of $31.455B.
Who are the investors of Anthropic?
Investors of Anthropic include General Catalyst, Lightspeed Venture Partners, D1 Capital Partners, Jane Street Group, Fidelity Investments and 101 more.
Who are Anthropic's competitors?
Competitors of Anthropic include Lila Sciences, Cognition, Decart, OpenAI, Convergence and 7 more.
What products does Anthropic offer?
Anthropic's products include Claude.
Loading...
Compare Anthropic to Competitors
MI2.ai focuses on machine learning predictive models in the data science and artificial intelligence sectors. The company provides services related to responsible machine learning practices, including research and consulting. It serves the academic community and businesses interested in implementing AI practices. The company was founded in 2016 and is based in Warszawa, Poland.
Mistral AI focuses on the development of open-source artificial intelligence models in the technology sector. Its main offerings include efficient, adaptable artificial intelligence (AI) models that allow for full customization by users without requiring their data. Its models are primarily used in the technology industry. It was founded in 2023 and is based in Paris, France.
OpenAI offers artificial intelligence (AI) research and deployment focused on ensuring that AI benefits all of humanity. Its main offerings include developing AI technologies with a commitment to safety, alignment with human values, and broad societal benefits. Its products and services are designed to address global challenges and promote the equitable distribution of AI advantages. It was founded in 2015 and is based in San Francisco, California.
Apollo Research operates as an AI safety organization focused on mitigating risks associated with advanced AI systems, particularly concerning deceptive behaviors. The organization conducts evaluations of AI models and advances interpretability research to understand and prevent strategic deception in AI. Additionally, Apollo Research provides consultancy services for responsible AI development frameworks and offers technical guidance to governance teams regarding AI policy. It was founded in 2023 and is based in London, United Kingdom.
Fireworks AI specializes in generative artificial intelligence platform services, focusing on inference and model fine-tuning within the artificial intelligence sector. The company offers an inference engine for building production-ready AI systems and provides a serverless deployment model for generative AI applications. It serves AI startups, digital-native companies, and Fortune 500 enterprises with its AI services. It was founded in 2022 and is based in Redwood City, California.
xAI focuses on artificial intelligence, specifically in the domain of language learning models. The company's main product, Grok, is designed to answer questions and suggest potential inquiries, functioning as a research assistant that helps users find information online. xAI primarily caters to the artificial intelligence (AI) research community and the general public seeking AI tools for information retrieval and understanding. It was founded in 2023 and is based in Palo Alto, California.
Loading...