P
Back to articles
News4 min read

Accelerating discovery with the AI for Math Initiative

Google DeepMind and Google.org launch the AI for Math Initiative with five top research institutions to accelerate mathematical discovery using AI systems including AlphaProof, AlphaEvolve, and Gemini Deep Think.

Source: Google DeepMind Blog
Accelerating discovery with the AI for Math Initiative

Google DeepMind Launches Pioneering AI for Math Initiative

Google DeepMind, in partnership with Google.org, has announced the launch of the AI for Math Initiative, a groundbreaking program designed to harness the power of artificial intelligence for mathematical research. The initiative brings together five of the world's most prestigious research institutions to explore how AI can accelerate discovery across the mathematical sciences.

"Mathematics is the foundational language of the universe, providing the tools to describe everything from the laws of physics to the intricacies of biology and the logic of computer science," wrote Pushmeet Kohli, VP of Science and Strategic Initiatives at Google DeepMind, and Eugénie Rives, Senior Director of GenAI Strategy at Google DeepMind, in a joint announcement. "For centuries, its frontiers have been expanded by human ingenuity alone. At Google DeepMind, we believe AI can serve as a powerful tool to collaborate with mathematicians, augmenting creativity and accelerating discovery."

World-Class Research Partners

The inaugural partner institutions of the AI for Math Initiative are:

artificial intelligence mathematics
Photo by Growtika on Unsplash
  • Imperial College London — A global leader in mathematics and AI research
  • Institute for Advanced Study — Home to legendary mathematicians including Albert Einstein and Kurt Gödel
  • Institut des Hautes Études Scientifiques (IHES) — The renowned French research institute known for breakthroughs in pure mathematics
  • Simons Institute for the Theory of Computing (UC Berkeley) — A hub for theoretical computer science and mathematics
  • Tata Institute of Fundamental Research (TIFR) — India's premier center for research in mathematics and the sciences

These partners will collaborate on three core objectives: identifying mathematical problems that are ripe for AI-driven insights, building the infrastructure and tools needed to power these advances, and ultimately accelerating the overall pace of mathematical discovery.

Cutting-Edge AI Technologies

DeepMind AI research
Photo by Google DeepMind on Unsplash

Google's support for the initiative includes funding from Google.org and access to DeepMind's most advanced AI systems. Researchers will have access to Gemini Deep Think, an enhanced reasoning mode of Google's flagship AI model; AlphaEvolve, an agent designed for algorithmic discovery; and AlphaProof, a formal proof completion system that assists with rigorous mathematical verification.

The initiative creates what DeepMind describes as a "powerful feedback loop" between fundamental research and applied AI, paving the way for deeper partnerships between the AI and mathematics communities.

A Remarkable Track Record

The launch comes on the heels of extraordinary progress in AI reasoning capabilities. In 2024, DeepMind's AlphaGeometry and AlphaProof systems achieved a silver-medal standard at the International Mathematical Olympiad (IMO), one of the world's most challenging mathematics competitions for pre-university students.

More recently, DeepMind's latest Gemini model equipped with Deep Think reached a gold-medal level performance at the 2025 IMO, perfectly solving five of the six problems and scoring 35 points — a milestone that demonstrates the rapid acceleration of AI's mathematical reasoning abilities.

Perhaps most strikingly, DeepMind's AlphaEvolve system has been applied to over 50 open problems in mathematical analysis, geometry, combinatorics, and number theory, improving the previously best-known solutions in 20% of them. In a landmark achievement, AlphaEvolve discovered a new, more efficient method for matrix multiplication — a core calculation underpinning nearly all of computing. For the specific problem of multiplying 4×4 matrices, AlphaEvolve found an algorithm using just 48 scalar multiplications, breaking the 50-year-old record set by Volker Strassen's algorithm in 1969.

In computer science, AlphaEvolve has also helped researchers discover new mathematical structures that reveal certain complex problems are even harder for computers to solve than previously understood, providing a clearer picture of computational limits and guiding future research directions.

What This Means for the Future

"We are only at the beginning of understanding everything AI can do, and how it can help us think about the deepest questions in science," the DeepMind team noted. By combining the profound intuition of world-leading mathematicians with the novel capabilities of AI, the initiative aims to open new pathways of research, advancing human knowledge and unlocking breakthroughs across scientific disciplines.

The AI for Math Initiative represents a significant step toward a future where AI and human mathematicians work side by side, with AI handling computational heavy-lifting and pattern recognition while humans provide creative intuition and problem framing. As AI reasoning capabilities continue to advance at a rapid pace, the potential for transformative discoveries in pure and applied mathematics has never been greater.

Related Articles