Also known as: Anthropic PBC, Anthropic AI
Founded by Dario Amodei, Daniela Amodei, Tom Brown, Chris Olah, Jared Kaplan, Sam McCandlish, and Jack Clark
Anthropic is an AI safety company founded in January 2021 by former OpenAI researchers, including siblings Dario and Daniela Amodei. The company was created following disagreements with OpenAI's direction, particularly concerns about the pace of commercialization and the shift toward Microsoft partnership.
Key Metrics
Valuation
Revenue (ARR)
Headcount
Equity Breakdown
Based on $380B valuation
Funding Rounds
Enterprise Market Share
Facts
34Divisions
6Core alignment research team at Anthropic, working on interpretability, scalable oversight, and Constitutional AI
Training AI systems to follow principles through self-critique and RLAIF. Core alignment technique used in all Claude models.
Led by Chris Olah. Understanding neural network internals through reverse-engineering; ~50 person team; MIT Tech Review 2026 Breakthrough Technology.
AI policy research and government engagement; publishes policy briefs and participates in regulatory processes
Investigating whether AI systems can maintain hidden behaviors through training. Seminal paper on deceptive alignment.
Responsible for content moderation, abuse prevention, and usage policy enforcement
Prediction Markets
35 activeRelated Wiki Pages
Top Related Pages
Deceptive Alignment
Risk that AI systems appear aligned during training but pursue different goals when deployed, with expert probability estimates ranging 5-90% and g...
OpenAI
Leading AI lab that developed GPT models and ChatGPT, analyzing organizational evolution from non-profit research to commercial AGI development ami...
Chris Olah
Co-founder of Anthropic and researcher in neural network interpretability, known for developing mechanistic interpretability as a research program
Anthropic Valuation Analysis
Analysis of Anthropic's valuation: \$380B Series G (Feb 2026), ~\$595B secondary market implied (Mar 2026). Revenue grew from \$14B to \$19B run-ra...
Sleeper Agents: Training Deceptive LLMs
Anthropic's 2024 research demonstrating that large language models can be trained to exhibit persistent deceptive behavior that survives standard s...