When Was Anthropic Founded? 2021 Origins and Claude Timeline

Anthropic was founded in 2021 by siblings Dario Amodei and Daniela Amodei with other former OpenAI colleagues. Its May 28, 2021 funding announcement identifies Dario as CEO and Daniela as president and sets out the new company's research agenda. Britannica's company history corroborates the founding year and OpenAI origins.

Fact-checked October 21, 2026. The founding year, first funding announcement, and Claude's launch dates answer different questions. Here is the distinction at a glance:

How did Anthropic start?

Anthropic emerged from a group of former OpenAI researchers and leaders. Dario and Daniela Amodei were co-founders, not the entire founding team. Britannica also identifies researchers Jared Kaplan and Chris Olah among the co-founders.

The contemporaneous May 2021 announcement is useful because it explains what the team planned to do at the time: research large-scale AI systems that were steerable, interpretable, and robust. It highlights reliability, tools for understanding models, and integrating human feedback into development and deployment. It also links the team's earlier work on GPT-3, interpretability, scaling laws, and learning from human preferences.

This supports an account of a new research company with experienced AI researchers. It does not establish a single private disagreement or personal motive for every founder's departure from OpenAI.

Is May 28 the exact founding date?

Not on the evidence cited here. May 28, 2021 is the publication date of Anthropic's Series A announcement. A financing announcement is not a certificate of incorporation. The defensible short answer to “when was Anthropic founded?” is 2021, rather than an exact legal founding day inferred from that announcement.

Why was Anthropic founded?

Anthropic's stated goal was to advance AI capabilities and reliability while studying how to deploy the systems beneficially. In its later explanation of its core safety views, the company says it was founded because AI could have transformative impacts, with significant uncertainty about whether those impacts would go well. That is Anthropic's account of its mission, not an independently verified explanation of each founder's intentions.

Its company page describes Anthropic as a public benefit corporation aiming to build reliable, interpretable, and steerable AI. This corporate structure is distinct from being a nonprofit. Its research and governance work includes:

• Constitutional AI: the 2022 research paper describes using written principles, model-generated critiques and revisions, and reinforcement learning from AI feedback to shape assistant behavior. Human choices still determine the principles and training process.
• Interpretability: research into how neural networks work internally, rather than relying only on their visible answers.
• Governance: the Long-Term Benefit Trust was designed to connect corporate oversight with Anthropic's long-term public-benefit mission.

These are methods and commitments, not proof of superior safety or a guarantee that Claude cannot be jailbroken or produce harmful or incorrect answers.

Anthropic history: a source-backed timeline

This is a selective history of the company, its research, and Claude's access milestones, not a complete release log or a list of models available today.

When did Claude become publicly available?

There are two important 2023 dates, depending on what “launch” means:

1. March 14: introduction and access requests. Anthropic announced Claude after a closed alpha with partners, describing chat and API access and inviting requests for access. Partner integrations were one route to using it.
2. July 11: public web beta. The Claude 2 announcement explicitly said anyone in the US and UK could start using the beta at claude.ai. This was not worldwide availability on day one.

Neither date is Anthropic's founding date. Likewise, Claude Code's February 2025 research preview and May 2025 general availability describe a later coding product, not the original Claude assistant's debut.

Anthropic and Claude today

Anthropic is the company; Claude is its assistant and model family. The company's current public description still combines AI research with products and a public-benefit mission. Historical names in the timeline should not be treated as current model recommendations. For present-day choices, use the Claude guides and model directory rather than a fixed roster in a history article.

This guide is published by magicdoor.ai, a separate service, not Anthropic. If you want to explore Claude here, Explore Claude in magicdoor.ai. Signed-out readers are taken to sign-in with that destination preserved; after signing in, check the model picker for current availability. Opening the chat does not generate an answer.

Sources and scope

Official Anthropic sources linked above were checked October 21, 2026, including the May 2021 funding announcement, Constitutional AI paper, Claude launch announcements, company page, core safety views, and Long-Term Benefit Trust explanation. Britannica's history was checked on the same date to corroborate the founding year and former OpenAI team. Announcement dates describe the events they actually announce, not inferred incorporation dates. The timeline intentionally ends with selected 2025 milestones and makes no claim to cover every subsequent release.

FAQ

When was Anthropic founded?

Anthropic was founded in 2021. Its May 28, 2021 Series A announcement is a dated early company milestone, not evidence that it was incorporated on that day.

Who founded Anthropic?

Anthropic was co-founded by siblings Dario and Daniela Amodei with other former OpenAI colleagues. Its May 2021 announcement identified Dario as CEO and Daniela as president.

How did Anthropic start?

Anthropic began in 2021 with a team of former OpenAI researchers and leaders. Its early research agenda focused on making large-scale AI systems more reliable, interpretable, and steerable, including through human feedback.

When did Anthropic launch Claude?

Anthropic introduced Claude on March 14, 2023 after a closed alpha with partners and offered access by request. On July 11, 2023, Claude 2 launched with a public-facing web beta in the US and UK. These were different stages of access, not the same launch.

Are Anthropic and Claude the same thing?

No. Anthropic is the company founded in 2021. Claude is its AI assistant and model family, introduced in 2023. Claude Code is a coding tool that Anthropic introduced as a research preview in February 2025.