01
Simulation
Ground scientific questions in molecular physics.
Atoms for Humanity
Building a sustainable molecular future.
Modern civilization is built from molecules. We are developing the scientific capabilities needed to discover and design chemistry for human and environmental health.
- MOLECULAR SCIENCE
- OPEN INFRASTRUCTURE
- PUBLIC BENEFIT
01 — THE CHALLENGE
The chemistry of modern life needs to change.
Chemistry underlies our materials, medicines, manufacturing, energy systems, agriculture, and technologies. But many of the molecular systems inherited by modern civilization were optimized for performance and cost without fully accounting for environmental persistence, toxicity, biological compatibility, or long-term sustainability.
We believe many of these problems can ultimately be addressed at their source: by discovering and designing better chemistry.
02 — CHEMICAL SPACE
Finding better chemistry means searching an enormous space.
The spaces of possible molecules, reactions, catalysts, materials, and chemical processes are vast. Traditional experimental discovery can explore only a small fraction of them.
Increasing humanity's ability to navigate this space is one of the fundamental requirements for building a more sustainable molecular future.
03 — WHY NOW
A new mode of molecular science is becoming possible.
Molecular simulation provides a way to interrogate physical chemistry computationally. Modern computing provides enormous scale. Artificial intelligence increasingly provides the ability to automate scientific reasoning and exploration. Connecting these capabilities creates the possibility of computational experiments operating at scales that were previously impractical.
02
Compute
Run large numbers of calculations in parallel.
03
AI
Automate hypothesis generation, analysis, and scientific iteration.
04
Infrastructure
Connect these capabilities into a coherent scientific system.
04 — THE DIGITAL LABORATORY
A digital chemical laboratory for humans and AI.
Atoms for Humanity is building open infrastructure that allows molecular simulations to be constructed, executed, analyzed, and reproduced at massive scale.
Human researchers can use this environment to investigate chemical systems systematically. AI systems can use the same scientific tools to formulate hypotheses, run computational experiments, interpret results, and decide what to investigate next.
Give increasingly capable AI access to a massively scalable digital chemical laboratory, and it gains something essential: the ability to test ideas against physical models rather than merely propose them.
- 01Hypothesize
- 02Simulate
- 03Analyze
- 04Learn
AT SCALEMany computational experiments in parallel
EXPERIMENT 001
EXPERIMENT 002
EXPERIMENT 003
EXPERIMENT N
05 — OPEN INFRASTRUCTURE
Infrastructure for molecular simulation at scale.
AFH develops a connected open-source stack spanning scientific data, program interoperability, distributed computation, and cloud access.
01
INTEROPERABLE FOUNDATIONS
qc*
A family of interoperable tools providing common scientific data structures, program interfaces, physical reference data, parsing, and molecular informatics.
02
DISTRIBUTED COMPUTE
BigChem
A distributed application for running quantum chemistry calculations at scale across clusters of computers or cloud infrastructure.
03
SECURE CLOUD ACCESS
ChemCloud
A web application and Python client for exposing BigChem computational infrastructure securely over the internet.
06 — RESEARCH
Building the tools—and using them to do science.
Infrastructure matters because of the science it makes possible. AFH develops new computational capabilities alongside an active research program designed to test them against difficult molecular problems.
01
High-throughput simulation
High-throughput simulation
Molecular simulation is traditionally organized around individual calculations and bespoke workflows. AFH is building systems that allow large computational experiments to be defined, distributed, executed, analyzed, and reproduced systematically. This capability forms the foundation for everything else we do.
02
Reaction-network discovery
Reaction-network discovery
Chemical systems can contain vast networks of possible reactions and intermediates. AFH is developing computational approaches for discovering and navigating these networks automatically. Scalable reaction discovery can help transform chemical exploration from a process constrained by individual hypotheses into a more systematic search over possible chemistry.
03
Catalyst design
Catalyst design
Catalysts determine which reactions become practical, selective, and efficient. AFH uses molecular simulation and computational exploration to investigate catalyst behavior and develop approaches to more systematic catalyst design.
04
Sustainable polymers and materials
Sustainable polymers and materials
Materials and polymers shape nearly every part of modern life. We are interested in molecular systems that can deliver useful performance while improving compatibility with human and environmental health.
07 — OPEN SCIENCE
Scientific capabilities should compound.
We believe the ability to simulate matter at scale should become a scientific public good.
AFH develops open software, workflows, data, protocols, and scientific results so that researchers beyond our organization can use, extend, and improve what we build.
- Open-source scientific software
- Shared computational infrastructure
- ChemCloud access for external researchers
- Open scientific workflows and data
- AFH Software Fellows
- Scientific collaboration
AFH also provides shared compute and ChemCloud access to external researchers and research groups.
08 — THE INSTITUTION
Building for the long term.
Atoms for Humanity is a 501(c)(3) nonprofit research organization founded to develop scientific capabilities that can serve the public over long timescales.
AFH's current research and infrastructure program is supported by a $2.2 million grant from Astera Institute.
About the organization