Anthropic, PBC is a San Francisco-based AI company founded in 2021 by former OpenAI executives, including siblings Dario and Daniela Amodei, to focus on AI safety, and is valued at $965 billion with a planned IPO in fall 2026. Its flagship product is the Claude series of large language models, offered in tiers like Haiku, Sonnet, Opus, and Fable, with its most powerful model, Claude Mythos, restricted to partnered US organizations for cybersecurity and life sciences. The company sources training data through Project Panama, a controversial effort to scan books, which led to a $1.5 billion copyright settlement with authors in 2025. Anthropic partners with Palantir to provide Claude to US federal agencies, including the Department of Defense, and the model was used in the 2026 Iran war and the US intervention in Venezuela. In February 2026, a dispute arose when the DoD demanded unrestricted use, but Anthropic refused to drop restrictions on mass domestic surveillance and fully autonomous weapons, leading the Trump administration to phase out its products—a decision blocked by a federal judge citing "First Amendment retaliation." Anthropic has also accused Chinese competitors like DeepSeek and Alibaba Cloud of distilling its models, and employs notable researchers such as Andrej Karpathy and John Jumper.