Achira, a biotech startup with Greek roots, has achieved a remarkable milestone, securing $33 million in funding less than two years after its founding. The company, established in 2023 by Theofanis Karaletsos, John Chodera, and Zavain Dar, gained the confidence of Nvidia, with its venture arm NVentures participating in the investment round alongside Dimension (lead investor), Amplify Partners, and Compound.

Achira is pioneering AI-powered drug discovery, integrating artificial intelligence, machine learning, physics-based simulations, and quantum chemistry to streamline the pharmaceutical research process. Instead of relying on time-consuming experimental methods, the company employs high-precision computational models to accelerate drug development.

The co-founders emphasize that their goal is to shift drug discovery toward a computational-first approach, where 90% of the process relies on simulations and only 10% on laboratory experiments. The company expects to unveil its first AI models in 2025.

"Achira’s models will match the accuracy of experimental methods while dramatically improving efficiency in utilizing experimental data. This will help transform drug discovery into an engineering-driven process," said Theofanis Karaletsos, co-founder of Achira, in a LinkedIn post.

The new funding will support Achira in expanding its team and refining its AI-driven simulation models.

"We are a team of experts in machine learning, quantum chemistry, engineering, and computational drug discovery, united by a shared ambition to create an entirely new approach to discovering therapeutic compounds," the team states on the company’s website.

Achira’s high-precision computational models leverage AI and physics-based simulations to accelerate drug discovery. More specifically, the company develops molecular dynamics simulation models, incorporating neural networks and fundamental physics principles to generate synthetic data that enhances the discovery process.

According to its founders, Achira was built on the belief that current biomolecular simulation methods are technologically outdated, unable to fully utilize the billion-fold increase in computational power seen in recent decades. Meanwhile, machine learning models alone face limitations due to insufficient data and the high costs associated with experimental drug discovery.

"We are creating a new generation of simulation models that integrate geometric deep learning, physics, quantum chemistry, and statistical mechanics. By transforming these into advanced probabilistic and generative models, we offer unprecedented efficiency across a wide range of biomolecular applications," the company states.