Dario Amodei

Dario Amodei is an American physicist, AI researcher, and technology executive, born in 1983 in San Francisco. He is the co-founder and chief executive officer of Anthropic, the public-benefit corporation that develops the Claude family of large-language models, and was previously vice president of research at OpenAI, where he led the GPT-3 effort. As of May 2026, he leads Anthropic following its February 2026 Series G at a $380 billion post-money valuation, with the April 2026 release of Claude Opus 4.7 as the company's current flagship model.

At a glance

• Education: Bachelor of Science in physics, Stanford University (transferred from Caltech); PhD in physics with a biophysics focus, Princeton University (2011), advised by William Bialek and Michael Berry; postdoctoral fellow, Stanford University School of Medicine.
• Current role: Chief Executive Officer and co-founder of Anthropic, since 2021.
• Key contributions: lead of the GPT-3 effort at OpenAI; co-founder of Anthropic (April 2021); architect of Constitutional AI as Anthropic's published alignment methodology; author of "Machines of Loving Grace" (October 2024); the Responsible Scaling Policy framework for capability releases.
• X / Twitter: @DarioAmodei
• LinkedIn: Dario Amodei
• Personal site: darioamodei.com

Origins

Amodei was born in San Francisco in 1983 to an Italian father and an American mother. His sister Daniela Amodei, four years younger, would later join him as a co-founder of Anthropic. He attended California Institute of Technology as an undergraduate before transferring to Stanford University, where he completed a Bachelor of Science in physics.

He moved east for graduate school, completing a PhD in physics with a biophysics focus at Princeton University in 2011 under advisors William Bialek and Michael Berry. The doctoral thesis, "Network-Scale Electrophysiology: Measuring and Understanding the Collective Behavior of Neural Circuits," concerned the collective behavior of neural ensembles and won the Hertz Foundation Doctoral Thesis Prize in 2011 and 2012. He continued in computational neuroscience as a postdoctoral fellow at the Stanford University School of Medicine after completing the doctorate.

Career

Amodei's professional path moved from computational neuroscience into industrial AI research over the 2014 to 2016 period. He joined Baidu in November 2014 as a research scientist on Andrew Ng's Silicon Valley team, working on Deep Speech and large-scale speech recognition systems. He spent roughly a year at Baidu before joining Google Brain in late 2015 as a research scientist.

In 2016 he joined OpenAI, then a non-profit AI research laboratory, where he held research-scientist and team-lead roles before being promoted to vice president of research. He led the GPT-3 effort, the 175-billion-parameter language model whose paper "Language Models are Few-Shot Learners" appeared in 2020 with Tom Brown as lead author. The work established large-scale autoregressive pre-training as the dominant paradigm in language modeling and catalyzed industry adoption of the scaling-laws thesis articulated by Jared Kaplan and OpenAI collaborators in early 2020.

Amodei left OpenAI at the end of 2020 with his sister Daniela and a group of senior researchers, citing a need for an organization in which AI safety research could be pursued under different commercial and governance constraints. The group formally incorporated Anthropic as a Delaware Public Benefit Corporation in early 2021, with Amodei as chief executive and Daniela as president. The seven public co-founders are Dario Amodei, Daniela Amodei, Tom Brown, Sam McCandlish, Jared Kaplan, Jack Clark, and Chris Olah. The first public Claude model launched in March 2023.

Amodei has remained chief executive of Anthropic continuously since the founding. The company has shipped successive Claude generations through Opus 4.7 in April 2026, raised approximately $73 billion in cumulative funding through the February 2026 Series G at a $380 billion post-money valuation, and become a frequent reference point in US AI policy discussions. Amodei testified before the Senate Judiciary Committee's Subcommittee on Privacy, Technology, and the Law in July 2023 and has been recalled to Congress on multiple occasions since.

Affiliations

• Baidu: Research Scientist, 2014-11 to 2015
• Google Brain: Research Scientist, 2015 to 2016
• OpenAI: Research Scientist and team lead, 2016 to ~2019
• OpenAI: Vice President of Research, ~2019 to 2020-12
• Anthropic: Co-founder and Chief Executive Officer, 2021 to present

Notable contributions

Amodei's body of work spans large-scale model training, alignment methodology, and public AI policy.

• GPT-3 effort at OpenAI (2019 to 2020). Led the team that scaled the GPT-3 series, the 175-billion-parameter language model whose results established few-shot prompting and the scaling-laws thesis as the dominant industry direction. Tom Brown, who later co-founded Anthropic with Amodei, was lead author on the GPT-3 paper.
• Constitutional AI (December 2022). Anthropic's published alignment methodology in which a model is trained to follow a written constitution of principles via self-critique and revision rather than human-feedback labeling alone. Constitutional AI is the basis for the alignment posture in shipped Claude models.
• Claude family (March 2023 to present). The Opus, Sonnet, and Haiku tiers of Anthropic's flagship model line. As of May 2026, Claude Opus 4.7 holds the leading position on SWE-bench Verified at 74.0 against GPT-5.5 at 68.5, the only major coding benchmark where Anthropic leads.
• Responsible Scaling Policy (September 2023, revised through 2025). Anthropic's framework defining capability thresholds at which additional safety measures are required. Amodei has been the primary public voice for the policy and the broader scaling-with-safety position.
• "Machines of Loving Grace" (October 2024). A 14,000-word essay laying out a vision for the upside of advanced AI across biology, neuroscience, economic development, governance, and meaningful work. The essay is widely cited as the canonical statement of Amodei's positive vision and is treated as an editorial counterpoint to the safety-focused statements that dominate his shorter public commentary.
• "The Urgency of Interpretability" (April 2025). A follow-up essay arguing that mechanistic interpretability research must accelerate if it is to mature in time to inform safety decisions for the next generation of frontier systems.
• US Senate testimony (July 2023). Amodei testified before the Senate Judiciary Subcommittee at "Oversight of A.I.: Principles for Regulation," warning of medium-term bioweapon-misuse risks and recommending testing, auditing, and product-recall mechanisms.
• Frontier Model Forum (July 2023). Anthropic was a founding member alongside Google, Microsoft, and OpenAI of the industry coordination body for frontier AI safety. Amodei was a public face of the announcement and Anthropic continues as a Forum participant.

Investments and boards

Amodei's footprint in this section is concentrated in his operating role at Anthropic rather than a personal angel-investing portfolio. The contrast with Sam Altman, whose disclosed personal portfolio at OpenAI spans hundreds of companies, is structural rather than incidental: Amodei is a researcher-operator without a parallel investing program.

• Anthropic (AI): Co-founder and Chief Executive Officer, 2021 to present. Public Benefit Corporation incorporated in Delaware. Approximately $73 billion cumulative funding through the February 2026 Series G at a $380 billion post-money valuation; reports of investor offers at an $800 billion valuation in April 2026 that the company has so far declined.

No public personal angel-investor activity on record in AI, semiconductors, datacenters, software, or energy as of May 2026. Anthropic's prior corporate-board memberships in industry coordination bodies, principally the Frontier Model Forum, sit alongside Anthropic's commercial roles rather than in Amodei's personal capacity.

Network

Amodei's strongest professional relationship is with his sister Daniela Amodei, Anthropic's president and co-founder, who handles the company's operating, partnerships, and external functions in parallel with his research and public-policy work. The pair worked together at OpenAI and have co-led Anthropic since the founding.

The five other public co-founders of Anthropic, all from OpenAI, form the long-running technical and policy core of his network: Tom Brown (training and infrastructure, GPT-3 lead author), Sam McCandlish (research leadership), Jared Kaplan (chief science officer; theoretical physicist by training and lead of the 2020 scaling-laws paper), Jack Clark (policy lead), and Chris Olah (interpretability research). Amodei's relationship with Sam Altman, his former OpenAI colleague, has been editorially material since the 2021 split. The two are now the principal head-to-head competitors at the frontier tier and are routinely paired in industry coverage. Andrew Ng, Amodei's manager at Baidu, was the early industrial-AI mentor of the pre-OpenAI period. Anthropic's funding base includes long-running relationships with Google, Amazon, and several institutional investors led by GIC and Coatue in the Series G.

Position in the field

As of May 2026, Amodei is one of three frontier-lab chief executives most often named in mainstream coverage alongside Sam Altman at OpenAI and Demis Hassabis at Google DeepMind. The three roles map onto distinct organizational archetypes: Altman runs the largest standalone consumer-AI organization at the highest private valuation; Hassabis runs the longest-tenured frontier research lab inside a public-company parent; Amodei runs the safety-positioned Public Benefit Corporation with an enterprise-first commercial channel. Anthropic's $380 billion February 2026 mark places it second to OpenAI's $852 billion March 2026 close on the private-company valuation axis.

Among the three frontier-lab chief executives, Amodei carries the longest formal research record into the role, with a Princeton physics PhD, a postdoctoral term at Stanford Medicine, and a publication history from the OpenAI period. The credential profile is closer to that of Yann LeCun at Meta AI and Hassabis at Google DeepMind than to operator-investor profiles like Altman. The public commentary balance is similarly distinctive: where Altman commentary skews toward AGI timelines and macroeconomic framing, and Hassabis commentary skews toward scientific applications, Amodei commentary skews toward concrete capability claims, alignment methodology, and policy specifics.

Outlook

Open questions over the next 6 to 18 months:

• Anthropic next funding step. Whether the April 2026 reported $800 billion offers convert into a closed round, hold steady, or move toward an IPO instead.
• Claude successor releases. Capability and pricing position of the next generation beyond the 4.x line, particularly against the GPT-5 family from OpenAI and the Gemini 3 series from Google DeepMind.
• Senate and congressional engagement. The frequency and substance of Amodei's appearances before US legislative bodies as AI legislation matures.
• Frontier Model Forum evolution. Whether the cross-industry coordination body acquires a more formal regulatory or standards role.
• Responsible Scaling Policy revisions. Future updates to the capability-threshold framework as Claude scales further and as competitor labs publish parallel policies.
• Public-policy posture on China. Continued export-control advocacy following Amodei's January 2025 essay on China and export controls.

Sources

• Dario Amodei. Wikipedia biographical entry; covers career, awards, publications, and the Anthropic founding.
• Dario Amodei, PhD. Hertz Foundation Fellow profile with the doctoral-thesis-prize citations.
• Dario Amodei. Personal site collecting essays, including "Machines of Loving Grace" and "The Urgency of Interpretability."
• Anthropic raises $30 billion at a $380 billion post-money valuation. Anthropic's February 2026 Series G announcement.
• Microsoft, Anthropic, Google, and OpenAI launch Frontier Model Forum. The July 2023 announcement of the cross-industry coordination body.
• Written Testimony of Dario Amodei, Co-Founder and CEO, Anthropic. The July 2023 Senate Judiciary Subcommittee statement on AI regulation.
• Anthropic's Responsible Scaling Policy. The September 2023 policy framework defining capability thresholds.
• Constitutional AI: Harmlessness from AI Feedback. The December 2022 Anthropic research paper documenting the Constitutional AI methodology.
• Language Models are Few-Shot Learners. The 2020 GPT-3 paper led by Tom Brown with Amodei among the senior authors.
• Network-Scale Electrophysiology: Measuring and Understanding the Collective Behavior of Neural Circuits. Amodei's 2011 Princeton doctoral thesis.
• Photo: File:Dario Amodei in 2023.jpg on Wikimedia Commons, CC-BY 2.0 Simon Walker / No 10 Downing Street (24 May 2023, London).