Anthropic

Anthropic is an AI company based in San Francisco, founded in 2021 by a group of researchers who left OpenAI, among them the siblings Dario and Daniela Amodei. It builds Claude, a family of large language models, and safety is its stated reason to exist: the founders' bet is that powerful AI is coming either way, and that it is better to have safety-focused labs at the frontier than to leave it to others.

What it sells is access to Claude. People and teams use the Claude app (web, desktop, mobile); developers call the models through an API, paying per token, and can also reach them inside Amazon Bedrock and Google Cloud. Claude Code, a tool that lets Claude work inside a software project (reading files, editing code, running tests), became one of its best-known products. The models come in sizes with different price and speed, named Haiku, Sonnet and Opus.

Anthropic is also a research lab. It published Constitutional AI, a method for training a model to follow a written set of principles; it is a leading group in interpretability, the effort to understand what happens inside a neural network; and it maintains a Responsible Scaling Policy that defines which safety measures each level of capability requires. In late 2024 it released the Model Context Protocol (MCP), an open standard for connecting AI models to tools and data that other companies adopted.

Why it matters: it is one of a handful of labs training frontier models, Claude is widely used for coding and for enterprise work, and its research and policy positions shape how governments and other labs think about AI risk. Amazon and Google are among its major investors. It is organized as a public benefit corporation, which lets it weigh its mission alongside profit.