Geneva, Switzerland – September 9, 2026
OpenAI has claimed a historic breakthrough: an internal artificial‑intelligence system produced a formal proof that resolves the Navier‑Stokes existence and smoothness problem, one of the seven Millennium Prize Problems that have eluded mathematicians for nearly a century.
The Navier‑Stokes equations govern the motion of fluids—from air over airplane wings to blood flow in capillaries. The central question asks whether smooth solutions can develop singularities, or “blow‑ups,” in finite time. A positive answer would confirm that under certain conditions fluid flow can become infinitely turbulent, while a negative answer would guarantee smooth behavior.
Context & Significance
First posed in 1900, the problem has been a cornerstone of mathematical physics, with a $1 million prize attached by the Clay Mathematics Institute. A solution would have profound implications for weather forecasting, aerospace engineering, and climate modeling, where turbulence remains a major source of uncertainty.
Research Scale & Design
OpenAI’s internal AI, a specialized large‑scale model distinct from its public GPT‑6 Astra, was tasked with exploring the problem over a 72‑hour sprint. Approximately 10,000 AI agents generated 2.7 million messages, consuming 130 billion output tokens. The resulting proof was formally verified using Lean, a proof‑assistant that checks each logical step. Verification took an additional 17 hours.
"The integration of massive parallel reasoning with formal verification opens a new paradigm for tackling deep mathematical challenges," said Dr. Kate O’Brien, OpenAI’s Director of Research.
Key Findings
- The proof demonstrates that a singularity can form in finite time under realistic energy‑conserving conditions, resolving the existence and smoothness question in the affirmative.
- The argument relies on a constructed vortex that spirals inward while being stretched, leading to unbounded velocity gradients while preserving finite kinetic energy.
- The proof is fully formalized in Lean, providing a machine‑checkable certificate that can be independently audited by the mathematical community.
Implications & Challenges
If the proof withstands peer review, it would be the first AI‑generated solution to a Millennium Prize problem, heralding a new era where machine intelligence collaborates on the most abstract scientific questions. It could accelerate research in fluid dynamics, enabling more accurate simulations of turbulence in climate models and aircraft design.
However, the result is not yet accepted by the broader mathematical community. OpenAI has released the Lean certificate for scrutiny but has opted not to claim the prize at this stage. Mathematicians such as Tristan Buckmaster (NYU) and Levent Alpöge (Anthropic) have called for independent verification, noting that the proof’s novelty warrants thorough examination.
"We must be cautious. A proof, no matter how elegant, needs to be examined line by line by experts before it becomes part of the canon," Buckmaster said.
Sources
- OpenAI press release, September 9 2026,
- Inside AI article, September 9 2026,
- Clay Mathematics Institute, Millennium Prize Problems,
FIRAT Editorial Board
Institutional Research Desk · Foresight Institute of Research and Translation
The collective editorial and research translation board of FIRAT, synthesising peer-reviewed evidence, policy briefs, and division milestones across our seven foundational research pillars.
