Who we are
Founders
- John Chodera (President / CEO)
- Theofanis Karaletsos
- Zavain Dar (Board)
Investors
Team
- Robin Betz
- Simon Boothroyd
- Gavin Crooks
- Elizabeth (Liz) Decolvenaere
- Benjamin DiFrancesco
- Alex Klingel
- Cody Liu
- Brian Lynch
- Alexandra (Lexie) McIsaac
- Nic Miller
- Ricardo Piro-Rael
- Amy-elizabeth Provost
- Joshua Rackers
- Andrea Rizzi
- Christopher Ryan
- Ana Silveira
- Daniel Smith (CTO)
- Eva-Maria T. Krauel
- Yuanqing Wang
- Marcus Wieder
Advisors
Latest News
2026-06-23 : [BIO2026] Achira Works With NVIDIA to Advance Atomistic World Models for Agentic Discovery
Announcing Achira
Update: Achira has secured $33M from Dimension, Amplify, NVIDIA, and Compound.
Achira is building atomistic world models to power the future.
We are a team of machine learners, quantum chemists, engineers, computer scientists, and entrepreneurs united by the ambition to create a fundamentally new way of modeling and simulating the world at the atomistic scale.
Achira is building the next frontier of world models for physical AI, synthesizing physics and AI to build general-purpose predictive models of how atoms and forces behave, interact, compose, and reconfigure under physical laws, in full motion, with thermal behaviors and environments explicitly simulated. Much of discovery in the physical world is naturally an out-of distribution task, exploring the fringes of where we have data and support in prior examples. Achira is founded on the belief that the only viable path to truly robust out-of-sample prediction for the physical world is through learning the underlying physical laws that govern the interactions of matter at the atomistic scale.
Achieving this turns intractable extrapolation problems for data into tractable interpolation problems, imbuing models with the data efficiency necessary to be useful in practical, real-world scenarios bounded by experimental resource and time constraints.
Current pure ML-approaches to atomistic modeling (e.g., in biology) have relied on learning predictive models and are limited by the paucity and scalability of high fidelity experimental data, traditional physical simulation approaches like molecular dynamics are technologically limited and ill-suited to harness the hundred-billion-fold increase in computing power that followed their development.
We are pioneering a third path.
Achira is creating a new class of atomistic world models that blend geometric deep learning, generative AI, physics, quantum chemistry, and statistical mechanics into advanced foundation simulation models and generative models. Based on the generalization power of their learned physics, these models will deliver unprecedented performance across a diverse set of applications, starting with drug discovery, with clear sightlines to transform other disciplines beyond. Our ultimate goal is learning the laws and systems that will allow humanity to make matter programmable starting with its fundamental units.
. . .
If you are a scientist or engineer with an uncompromising drive to build the foundational technological substrate for programming matter and interested in joining the team, or a drug discovery scientist interested in partnering with us to access the next generation of atomistic world models, we encourage you to contact us.
Follow us on Bluesky at @achira-ai.bsky.social.
