Daniel Levy, President and Co-Founder of Safe Superintelligence (SSI)

Daniel Levy is an artificial intelligence (AI) researcher and the president of Safe Superintelligence Inc. (SSI), which he co-founded in June 2024.

Levy is a computer scientist, distinct from the actor Daniel Levy who starred in the television series Schitt's Creek.

The GPT-4 Technical Report credits Levy as overall vision co-lead and optimization lead. He earned a PhD in computer science at Stanford University, completed in 2022 (per his advisor John Duchi's CV).

Who Daniel Levy Is

Daniel Levy has been president of Safe Superintelligence Inc. (SSI) since July 2025, and he co-founded SSI with Ilya Sutskever and Daniel Gross in June 2024.

Sutskever's July 3, 2025 note on the SSI updates page says: "I am now formally CEO of SSI, and Daniel Levy is President." The note adds: "The technical team continues to report to me." TechCrunch reported the change the same day.

• Current role: President and co-founder.
• Organization: Safe Superintelligence Inc. (SSI).
• Known for: Credited as overall vision co-lead and optimization lead in the GPT-4 Technical Report, and co-founding SSI.
• Education: Doctor of Philosophy (Ph.D.) in computer science from Stanford, completed in 2022 (per his advisor John Duchi's CV).
• In the news for: SSI's NVIDIA compute partnership, announced in late July 2026.

Education and Academic Training

Daniel Levy studied at Lycée Louis-le-Grand and École Polytechnique, then earned an MS and a computer science PhD at Stanford University.

According to his Stanford page, Levy earned a Bachelor of Science (B.S.) and a Master of Science (M.S.) at those two schools in 2014 and 2015. The Stanford page does not say which degree came in which year.

Levy then pursued graduate study at Stanford University, where he earned a Master of Science under the supervision of Stefano Ermon. His master's research centered on probabilistic models and reinforcement learning (RL). His Stanford page gives no years for this degree.

Levy earned a PhD in computer science from Stanford, completed in 2022 (per his advisor John Duchi's CV). His doctoral research focused on machine learning, mathematical optimization, and data privacy. John Duchi's curriculum vitae (CV) records Levy as a Stanford doctoral student from 2019 to 2022. Duchi's CV lists Levy's current role as President, Safe Super-Intelligence. Similarly, Duchi's Stanford group page lists Levy as a 2022 computer science doctoral graduate and co-founder of SSI.

Internships and Academic Service

Levy's Stanford page lists three industry research internships, at Facebook in 2016 and at Google in 2017 and 2020.

In 2016, Levy worked as an intern at Facebook Applied Machine Learning, according to his Stanford page. He completed an internship at Google Brain in 2017, working with Jascha Sohl-Dickstein and Matt Hoffman. In 2020, Levy served as a research intern at Google Research, working alongside Ananda Theertha Suresh, Satyen Kale, and Mehryar Mohri.

He served as a teaching assistant for CS229 Machine Learning under Andrew Ng and John Duchi in Fall 2016. In Winter 2021, Levy assisted John Duchi as a teaching assistant for EE364A Convex Optimization.

Beyond teaching, Levy reviewed papers for machine learning conferences. His peer review service includes the International Conference on Machine Learning (ICML) in 2019 and 2021, and the Conference on Neural Information Processing Systems (NeurIPS) in 2020. He reviewed papers for the International Conference on Learning Representations (ICLR) in 2019 and 2020, and the Association for the Advancement of Artificial Intelligence (AAAI) in 2020. Levy also reviewed for Advances in Approximate Bayesian Inference (AABI) in 2018 and the R2L Workshop at NeurIPS 2018.

OpenAI and the GPT-4 Technical Report

Two documents describe Levy's work at OpenAI: the GPT-4 Technical Report and his undated Stanford page.

His undated Stanford page states: "I currently lead the Optimization team at OpenAI." Neither that page nor the GPT-4 report gives start or end dates for his work at OpenAI.

OpenAI published the GPT-4 Technical Report in 2023. Its Vision section lists Levy among core contributors as "Overall vision co-lead, optimization lead." The same technical report lists Levy under vision contributions for optimization and architecture as well as training run babysitting. OpenAI also credited Levy under additional contributions for blog post and paper content.

On June 19, 2024, TechCrunch described Levy as an "ex-OpenAI engineer" in its report on SSI's launch.

Co-Founding SSI and Becoming President

On June 19, 2024, Ilya Sutskever, Daniel Gross, and Daniel Levy signed the public launch note. The note calls SSI "the world's first straight-shot SSI lab, with one goal and one product: a safe superintelligence." The same line appears on the SSI homepage as of October 2026. The launch note stated that SSI is an American company with corporate offices located in Palo Alto and Tel Aviv.

Former co-founder Daniel Gross left SSI on June 29, 2025.

SSI has not described Levy's duties as president beyond the title. For how SSI is set up, see Safe Superintelligence explained.

Financing Rounds and Strategic NVIDIA Compute Access

SSI itself has announced one funding round.

On September 4, 2024, SSI said on its updates page that it raised $1 billion from NFDG, a16z, Sequoia, DST Global and SV Angel.

On April 12, 2025, TechCrunch reported that SSI raised an additional $2 billion at a $32 billion valuation. TechCrunch attributed the report to the Financial Times, noting the round was reportedly led by Greenoaks. SSI did not comment on the report.

In late July 2026, SSI entered a long-term strategic partnership with NVIDIA, as announced on the SSI updates page. SSI said the deal will let it scale its compute by 10x. The financial terms were undisclosed by SSI and NVIDIA in their public announcements. Reuters, in a report carried by The Star, cited a person briefed on the deal as saying NVIDIA would make a $5 billion equity investment. Bloomberg reported the same figure.

A release relayed by StorageNewsletter noted SSI secured access to the NVIDIA Vera Rubin platform. The release describes SSI as "led by Ilya Sutskever and Daniel Levy." As of October 2026, neither the SSI homepage nor its updates page lists a product. TechCrunch wrote in April 2025 that the product "is presumably still in the works."

Published Research, 2017 to 2024

Levy's papers cover optimization, privacy, and probabilistic models, with venues including NeurIPS, ICLR, AAAI, UAI and EMNLP. His publications include collaborative studies with Stanford faculty members Stefano Ermon and John Duchi.

In 2017, Levy co-authored an ICLR paper on data noising as smoothing in neural network language models, with Jurafsky and Ng among its seven authors. At UAI 2017, Levy and Mussman published research on fast amortized inference with Stefano Ermon. In November 2017, Levy, Matt Hoffman, and Jascha Sohl-Dickstein submitted neural Hamiltonian Monte Carlo research to arXiv. The paper appeared at ICLR 2018.

At AAAI 2018, Levy and Ermon presented research on deterministic policy optimization combining pathwise and score function estimators. Levy also co-authored a UAI 2018 paper on Bayesian optimization and attribute adjustment with Ermon and three co-authors. His NeurIPS 2019 paper with Duchi, Necessary and Sufficient Geometries for Gradient Methods, was selected for an oral presentation.

In October 2020, Levy, Carmon, Duchi and Sidford submitted large-scale distributionally robust optimization methods to arXiv. The paper appeared at NeurIPS 2020.

At NeurIPS 2021, Levy co-authored Learning with User-Level Privacy, whose seven authors include Satyen Kale, Mehryar Mohri and Ananda Theertha Suresh, his hosts at Google Research in 2020. A second NeurIPS 2021 paper with Asi and Duchi examined function difficulty and growth conditions in private optimization. At EMNLP 2021, Levy co-authored research on distributionally robust multilingual machine translation. Duchi's CV lists a 2024 arXiv revision, Geometry, computation, and optimality in stochastic optimization, by Cheng, Levy and Duchi.

Net Worth and Public Record

The public record contains no reliably published net worth figure for Daniel Levy of SSI. Neither SSI nor Levy has disclosed his shareholding or pay.

For the wider field, see what superintelligence means and AI people to watch.

SSI posts its leadership and funding changes on its updates page.