Join us at the NVIDIA Atomistic Simulation Summit, a two-day in-person event co-hosted by NVIDIA and the Max Planck Institute for Polymer Research, September 3–4.


Built for developers advancing chemistry and materials science modeling and simulation, the summit brings together leading researchers and practitioners to explore the latest progress, challenges, and opportunities in atomistic simulation.


Sessions cover domain-specific libraries for high-performance simulations and AI, GPU application porting best practices, mixed and reduced precision strategies, machine learning potentials and generative AI, dataset generation, batched simulation approaches for scaling workloads, and community-contributed technical talks.


Whether you're pushing the boundaries of molecular dynamics, developing next-generation force fields, or scaling simulation workloads, the NVIDIA Atomistic Simulation Summit 2026 is where the community comes together to advance computational chemistry and materials science.

Agenda
Speakers and timing are subject to change. 

Day 1

8:00 AM - 8:45 AM
Registration
 
8:45 AM - 9:00 AM
Welcome and Introductions

Piero Altoe’

9:00 AM - 9:20 AM
MLIPs: status and future directions

Gabor Csanyi

9:20 AM - 9:40 AM
Investigating the accuracy and performance of bulk and surface properties of inorganic materials calculated by the machine-learned interaction potential MACE

Johan Carlsson

9:40 AM - 10:40 AM
Foundation Machine-Learned Potentials for Thermodynamic Property Prediction: From Liquid Densities to Drugs and Materials

Tim Duignan, Orbital

Machine-Learning Potentials for Heterogeneous Catalyst Modeling: Opportunities and Challenges in Industrial R&D
Sandip De, BASF

10:40 AM - 11:00 AM
Coffee Break
 
11:00 AM - 12:30 PM
Navigating the Machine-Learned Potential Era: Performance Benchmarks and Advanced ML/MM Integration via OpenMM

Thomas Markland, Stanford

ALCHEMI: A GPU-Accelerated Ecosystem for Atomistic Modeling with MLIPs
Justin Smith, NVIDIA

Agentic Workflows for Atomistic Simulations
Fabio Falcioni, NVIDIA

12:30 PM - 1:30 PM
Lunch
 
1:30 PM - 3:30 PM
Integration of Skala ML Functional in CP2K

Thomas Kühne, HZDR/CASUS

Large-scale Quantum Chemistry with ORCA: NVIDIA cuEST and new SURFF minimal-parameter potential
Bernardo de Souza, FACCTs GmbH

cuEST: Accelerating Quantum Chemistry on GPUs
Robert Parrish, NVIDIA

Porting to the GPU, with agents
Piero Altoe, NVIDIA

Leveraging agents with NV libraries
Piero Altoe, NVIDIA

3:30 PM - 4:00 PM
Coffee Break
 
4:00 PM - 5:00 PM
Discussion on MLIP inference benchmarking

Meta FAIRChem + ALCHEMI

 

5:00 PM - 6:00 PM
Poster Session
 
7:00 PM - 10:00 PM
Social Dinner
 

Day 2

Parallel Sessions 1: Machine Learning                                                                                Parallel Sessions 2: Classical/quantum

9:00 AM - 10:15 AM
Parameterizing Rotation-Equivariant Functions of Atomic Environments

Chuck Witt, Harvard

Simulating Molecules in the Age of AI
Nicholas Gao, CuspAI



Toward Scientific Superintelligence: Autonomous Agent Frameworks for Accelerated Materials Discovery
Cameron Owen, Lila

Flexible Cutoff Learning: Optimizing Machine Learning Potentials After Training

Jan Hamaekers, Fraunhofer Institute for Algorithms and Scientific Computing SCAI

GPU-resident GROMACS: The performance wins and software engineering realities of NVSHMEM
Andrey Alekseenko, Szilárd Páll, KTH Royal Institute of Technology

Fast training of bespoke molecular mechanics force fields using machine learning potentials and the Open Force Field infrastructure
Daniel Cole, Newcastle

10:15 AM - 10:45 AM
Coffee Break
 
Coffee Break
 
10:45 AM - 12:25 PM
From electron densities to a tomographic materials representation

Tejs Vegge, Technical University 

 

Achieving quantum accuracy: from imperfect solids to devices
Michele Simoncelli, Columbia


metatensor and metatomic: Foundational libraries for interoperable atomistic machine learning
Guillaume Fraux, EPFL

Skala: A Deep Learning Exchange-Correlation Functional with Hybrid Accuracy at Semi-Local Cost
Gregor Simm, MSFT

Alchemical Free Energy at Scale: Open-Source Tools for Large Binding Affinity Campaigns

Hannah Baumann, Open Molecular Software Foundation / Open Free Energy


AmberStudio: A Framework for High-throughput GPU-accelerated AI-enhanced Drug Discovery
Ryan Snyder, Rutgers

Improving GPU Performance for NAMD Molecular Dynamics
David Hardy, UIUC


Party Like it's 1999: How Molecular Simulations Can Turn a New Page
David Cerutti, Psithera

12:25 PM - 1:30 PM
Lunch
 
Lunch
 
1:30 PM - 2:45 PM
AI agents and Autonomous Labs

Christopher Sutton, University of Toronto

ML-PEG: Machine Learning Performance and Extrapolation Guide
Alin Elena, Scientific Computing Department, Science and Technology Facilities Council, UK

Emulating protein equilibrium ensembles at scale
Lixin Sun, MSFT

Accelerating QM and QM/MM simulations in QUICK and Amber with cuEST

Andreas Goetz, UCSD

Accurate Reference Data for Periodic Systems from Coupled Cluster
Andreas Irmler, TU Wien

Advancing Quantum ESPRESSO for Enhanced Accuracy and Scalability in Atomistic Simulations on NVIDIA GPU-Accelerated Architectures
Francesco Andreucci, SISSA

2:45 PM - 3:15 PM
Coffee Break
 
Coffee Break
 
3:15 PM - 4:30 PM
Extending the successes of atomistic MLIPs up a scale or two

Rose Cersonsky, UW Madison


A Quantum Foundation Model for Accurate Atomistic Simulations in Drug Design
Jean-Philip Piquemal, Sorbonne

 

The MLCG Ecosystem: Scalable Machine-Learned Coarse-Graining for Biomolecular Simulation
Jacopo Venturin, Clementi Lab

Directives-Based GPU Programming: the Portable Performance Path for Plane-Wave Pseudopotential Programs?

Phil Hasnip, University of York

QuantumATK: Towards a GPU-Centric Platform for Atomistic Simulation from Machine-Learned Potentials to Post-DFT Methods
Gabriele Penazzi, Synopsys

EmuGEMM: High-Precision GEMM on Low-Precision Tensor Cores, with a Case Study in Quantum Transport
Denghui Lu, Luisier

4:30 PM - 4:45 PM
Transition
 
4:45 PM - 5:15 PM
Closing Plenary
 

Event Details

Please register below to secure your spot. Please note that registration is subject to approval due to limited space, and travel and accommodation are not included. You will be notified of your registration status once your application has been reviewed.

September 3-4, 2026 / Max Planck Institute for Polymer Research, Mainz, Germany

Venue

Ackermannweg 10
Gonsenheim
Mainz RP 55128
Germany