Schrodinger, Inc. Common Stock Morgan Stanley 24th Annual Global Healthcare Conference

Thanks for having us, Craig.

Sure. On the scientific side, right?

Right. On the platform side.

Yeah. Yeah. Well, as you said, we've been at this for 36 years, actually, since 1990, advancing a computational platform for design of molecules, for designing molecules and predicting the properties of molecules.

One of the biggest advances is developing methods that allow us to actually accurately simulate what molecules are doing in the incredibly complex environment they're in. Molecules, drug molecules, for example, exist in water. Water turns out to be pretty sophisticated, and the interactions of water with small molecules is complex, and then, of course, when small molecules bind to proteins, that's a very complex and highly dynamic process.

We've developed all sorts of technologies for understanding the structures of proteins, understanding the structure of proteins in water, understanding how molecules bind to proteins and what happens, and then the result of that is the ability to predict some incredibly complex properties, such as how tightly a molecule binds to a protein, how soluble a molecule is, how permeable it is. These are incredibly important properties for developing drugs. A drug molecule has to be potent and selective and soluble and permeable and so on. There's a huge amount of technology that underlies the ability to use first principles physics to predict accurately the properties of molecules and essentially emulate experiment. It would probably not be a good use of time to get into the underlying physics that results in it, but I hope that gives you a general sense.

Now, since then, once you have this core engine, this physics engine, how do you amplify it? How do you get that kind of technology out to customers? One is you have to be able to manage huge amounts of data. I mean, running these physics calculations on massive scales, approaching trillions of molecules, you need a very robust enterprise system for storing all of this data, being able to analyze it. We've done some really nice work on the engineering side to manage all of that data. The other thing is these physics-based calculations are computationally expensive. But I said earlier that you have to be able to explore maybe trillions of molecules, certainly on the order of hundreds of billions of trillions of molecules. You can't do that using physics alone.

We've put a significant effort into developing machine learning and AI methods that are trained on physics, that amplify the physics to a scale that allows us to explore huge amounts of chemical space. That's another very important aspect of the platform. I think that's, we're touching on sort of the main And- sort of breakthroughs, and there's a lot more to do.

Go ahead, Pat. As I'd say, and then, of course, you've talked about where we are today and the next step.

What we pride ourselves on is thinking what a computer is going to be able to do in five years, and how do we start working on it now. When we started working on our first, essentially, agent three or four years ago, it didn't work, and we kept building it and improving it, and finally, as LLMs and agents reached the space they did, we built Bunsen, which is our new agent that can run all these workflows at a scale that was previously unfathomable. We got ahead of that.

It took years of getting ready so that when the technology got there, we could do it, which we'd done with GPUs in the past with physics, the cloud at scale, we did with running physics at a larger scale. We really look forward to the next generation of technology with science.