🔬Searching the Space of All Possible Materials — Prof. Max Welling, CuspAI

2026-02-26 · Source: Latent.Space - Www.latent.space · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Emerging Technologies & Innovation · Depth: Advanced, extended

Summary

Max Welling, a prominent figure in deep learning known for variational autoencoders and graph neural networks, discusses his career evolution and the motivation behind his startup, CuspAI. He highlights a shift from purely theoretical physics interests to problems with significant real-world impact, particularly climate change. Welling views nature as a "physics processing unit" (PPU), the fastest known computer, which, when combined with digital computation, can accelerate material discovery. CuspAI, which has raised $130 million and grown to 40 people, aims to automate the material development process, focusing on applications like carbon capture, advanced batteries, and sustainable plastics. The company's platform integrates generative models, multi-scale digital twins, and high-throughput experimentation, with a vision to empower chemists and material scientists through increasingly automated tools.

Key takeaway

For AI engineers and scientists seeking to make a tangible impact, consider specializing in AI for science. This emerging discipline offers opportunities to address critical global challenges like climate change through material innovation. Your expertise in developing and applying AI models can significantly accelerate the discovery of new materials for batteries, solar panels, and carbon capture, moving beyond software-only solutions to fundamental physical problems. Focus on building robust, modular tools that empower domain experts, rather than aiming for full automation immediately.

Key insights

AI for science is an exploding field, combining deep scientific problems with high-impact technological solutions.

Principles

• Physics provides fundamental mathematical tools for AI.
• Symmetry infusion reduces data requirements for neural networks.
• Material innovation underpins nearly all technological progress.

Method

CuspAI's platform uses generative models for candidate generation, multi-scale digital twins for evaluation, and integrates high-throughput experimentation with agentic orchestration to accelerate material discovery.

In practice

• Apply AI tools to material science for energy transition.
• Utilize equivariance to reduce data needs in neural networks.
• Automate experimental workflows with AI agents.

Topics

• AI for Material Science
• Equivariant Neural Networks
• Generative AI
• Stochastic Thermodynamics
• Automated Material Discovery

Best for: AI Scientist, Investor, Entrepreneur, AI Engineer, Research Scientist, Domain Expert

Related on AIssential

• AI materials discovery now needs to move into the real world — AI-driven autonomous labs aim to revolutionize materials discovery by accelerating synthesis and testing, despite current limitations.
• Accelerating science with AI and simulations — AI is at a second inflection point, integrating language and multi-modal reasoning to accelerate scientific discovery.
• Bridging AI & Science: The Impact of Machine Learning on Material Innovation with Joe Spisak of Meta — AI, especially machine learning, is fundamentally reshaping material innovation and scientific discovery.
• AI for Atoms: How Periodic Labs is Revolutionizing Materials Engineering with Co-Founder Liam Fedus — AI's next frontier involves connecting intelligent systems to the physical world for scientific and material discovery.
• AI breakthrough could replace rare earth magnets in electric vehicles — AI can dramatically accelerate the discovery of new high-temperature magnetic materials, reducing rare-earth dependence.
• 🔬Why There Is No "AlphaFold for Materials" — AI for Materials Discovery with Heather Kulik — Deep integration of domain expertise with AI and critical assessment of model outputs are crucial for success in AI for science.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Latent.Space - Www.latent.space.