

Manas AI Inc., a startup using artificial intelligence to discover new medicines, launched today with $24.6 million in initial funding.
The investment was led by General Catalyst and Manas AI co-founder Reid Hoffman. Greylock Partners, where Hoffman was a general partner until 2023, participated as well.
The billionaire LinkedIn co-founder launched Manas AI with Siddhartha Mukherjee (pictured), a prominent cancer researcher and Pulitzer Prize-winning author. The duo plans to harness AI to develop new medicines faster than would be possible using current methods. Manas AI will initially focus its efforts on developing new cancer drugs, with plans to treat autoimmune disease and rare conditions down the road.
Many medicines are based on small molecules, organic compounds with a diameter of about one nanometer. Before an organic molecule can neutralize a disease-causing protein, it has to attach to that protein through a process known as binding. This binding process will be a major focus of Manas AI’s research efforts.
One way scientists develop new drugs is by looking for a small molecule that can successfully bind to a disease-causing protein. They carry out their search using a method known as docking, which involves simulating many different ways of binding small molecules to a harmful protein.
During docking simulations, scientists must identify a section of the harmful protein’s surface area to which the small molecule can bind. Moreover, they have to find the orientation that the small molecule should adopt before the binding process. The molecule’s orientation can significantly affect the effectiveness of the treatment.
The task is challenging partly because there are many theoretical ways to bind a small molecule to a disease-causing protein, but only a few work in practice. Additionally, both the molecule and the protein are flexible, which means they can change their configuration while the binding process unfolds. Those factors make docking simulations both highly complicated and hardware-intensive.
Manas AI hopes to use AI models to speed up docking simulations by a factor of 100. In parallel, the company will launch a research initiative called Project Cosmos. The latter program’s goal is to map out the fundamental rules of molecule binding and harness them to speed up drug development.
Manas AI is partnering with Microsoft Corp. to support its research efforts. Its scientists will run their AI models on the technology giant’s Azure cloud platform.
The company plans to create what it describes as a full-stack therapeutic development pipeline. This means it will not only identify drug candidates, or small molecules that can potentially be turned into medicines, but also aim to manage the subsequent steps involved in developing new treatments. This includes the task of running clinical trials.
“Through the power of AI and our world-class team, we believe we can drastically reduce the time and cost it takes to bring game-changing new treatments to patients,” said Mukherjee.
Manas AI will use its seed funding to finance its AI development efforts, research drug candidates and carry out clinical trials.
THANK YOU