Anthropic

In early 2021, seven researchers walked out of OpenAI. They were worried about the direction of artificial intelligence and wanted to build something different. Led by siblings Dario and Daniela Amodei, along with Jared Kaplan, Jack Clark, Chris Olah, and others, they founded Anthropic on January 26, 2021. The company raised $124 million in its first funding round that May.

Anthropic spent most of 2022 training its first large language model, Claude, but held off releasing it. The team wanted more time for safety testing, concerned about sparking a dangerous race in AI development. Claude finally launched in March 2023. The company also developed Constitutional AI, a framework that guides model behavior through a set of principles rather than relying solely on human feedback. This approach became a signature of the company's identity.

By 2026, Anthropic had grown to 2,500 employees and over 300,000 business customers. Claude became available on all three major cloud platforms: Amazon Web Services, Google Cloud, and Microsoft Azure. The company's revenue trajectory stunned analysts. From roughly $1 billion annual run-rate in early 2025, it climbed past $5 billion by August, with projections hitting $26 billion for 2026. In May 2026, Anthropic raised $65 billion in Series H funding led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, valuing the company at $965 billion. That made it the most valuable pure-play AI company in the world.

Anthropic operates as a Public Benefit Corporation, legally required to balance profit with social responsibility. The company's headquarters sit at 500 Howard Street in San Francisco. As of late 2025, more than 78 percent of Claude's usage came from outside the United States. Anthropic announced plans to triple its international workforce, recruiting across India, Japan, Korea, Australia, and Europe. The company signed agreements with Amazon for up to five gigawatts of new compute capacity, and with Google and Broadcom for five gigawatts of next-generation TPU capacity. [edit]