Quantum in Healthcare: How Cleveland Clinic Is Scaling Molecular Simulation Beyond Classical Limits

Quantum computing has spent years being described as nearly useful. Cleveland Clinic is past that point. In a collaboration with IBM and RIKEN, the institution's team completed the first large-scale quantum simulation of a protein-ligand complex in explicit water, scaling calculations across 10,000 to 13,000 atoms. That result does more than validate quantum's viability for biomedical research. It resets the timeline for when quantum-enhanced drug discovery becomes standard in pharmaceutical development.

At IBM Think 2026, Daniel Newman sat down with Dr. Kenneth Merz, Principal Investigator at Cleveland Clinic Research, to unpack what this milestone means for molecular modeling, free energy calculations, and the long road from bench science to clinical impact.

The conversation covers the embedded wave-function framework Merz's team developed to run atom-by-atom quantum calculations on full protein structures, how quantum hardware enabled probabilistic sampling that classical systems cannot replicate, and how fault-tolerant systems arriving around 2029 will shift what precision is achievable at the binding-site level. Merz also addresses the near-term case: quantum is already improving lead-optimization workflows by producing highly accurate binding-free energy data that pharmaceutical companies use today.

Key Takeaways:

🔹 Large-scale molecular simulation is no longer theoretical. Cleveland Clinic, IBM, and RIKEN calculated the quantum electronic structure of a 10,000-to-13,000-atom protein-ligand complex in explicit water, a first-of-its-kind result that confirms quantum can scale to biologically relevant systems.

🔹 The embedded wave function framework is hardware-agnostic. By breaking molecular structures into individual atoms and stitching calculations together, the approach runs on any quantum device capable of providing electronic structure samples and accepts AI acceleration at individual pipeline steps.

🔹 Quantum's role in drug discovery includes speed and accuracy. Free energy calculations that predict how tightly a drug binds to a receptor are already in use across the pharma industry. Quantum improves those results by accessing electronic structure data that classical force fields approximate rather than compute.

🔹 Hybrid quantum-classical architecture is the near-term operating model. Quantum hardware handles probabilistic sampling of complex electronic structure spaces. Classical systems post-process results. Fault-tolerant devices expected around 2029 will enable quantum to perform high-accuracy active-site calculations, while noisy devices will cover the surrounding molecular environment.

🔹 Lead optimization stands to compress drug development timelines. Merz describes a scenario where quantum-informed calculations reduce the number of compounds synthesized per program while increasing the hit rate, potentially shifting a 1,000-compound screen to 100 compounds with five times the useful output.

Five years from now, Merz sees the Cleveland Clinic producing millions of high-accuracy quantum mechanical calculations to train machine learning potentials, feeding a data layer that could reshape how computational chemistry informs every stage of drug development.

IMPORTANT: Note to Uploader: VIDEO EMBED – added by uploader.

Disclaimer: Six Five Media makes no guarantee, warranty, or representation as to the accuracy or completeness of the information presented. This content is produced independently and is not sponsored, endorsed, or reviewed for accuracy by any company mentioned. All trademarks and company names referenced are the property of their respective owners. This is not investment advice.

DANIEL NEWMAN:
Hi, everyone. The Six Five is on the road and we are here with another episode of Quantum is Real. Very excited to have this conversation. Today we have Cleveland Clinic's Kenny Merz. Should I say Dr. Kenny Merz? 

KENNETH MERZ: 

No, Kenny's fine. 

DANIEL NEWMAN:

All right. He's not going to have me say that. Kenny Merz is here with me, and very excited to talk to him about how Cleveland Clinic is taking quantum technology from the experimental phase, and they are making it real in the work that they are doing. This is such an exciting topic to me. So excited to have you all here. Kenny, even more excited for you to be with me. I'm excited to be here, too. Can't wait to talk about this. So you heard the setup, right? You're taking this technology, and quantum is infamously five years away from being five years away. Yeah, and the Gantt. Yeah, and my thesis is that's just not the case anymore. It seems that Cleveland Clinic is aligned. I mean, you guys are moving quickly from that experimental phase to that bringing quantum to the work you're doing. Talk a little bit about what that journey's been like.

KENNETH MERZ: 

Yeah, so it goes back some years, right? Because I was working in this area called linear scaling or fragment-based methods. So the idea is you take a big problem, you break it into little bits and then reassemble, right? And this was done with quantum chemistry, right? And it was very, very effective in the classical realm. And I did some research over the years, and I sort of put it aside. And then in 2010-ish or so, you start hearing like hydrogen atom, done with quantum or lithium hydride. Wow. And I realized, wait a minute, if they can only do that big, this linear scaling stuff, if we could get it down to the atom, we could scale it out, right? And Cleveland Clinic came, asked me out for an interview. I visited with them. They're just getting the devices partially assembled in the cafeteria. And I was like, yeah, I'm for it. Let's do it. I think I'll be able to scale up my systems by using these kinds of strategies. So I was kind of a noob, right, in quantum computing. I was an expert in quantum mechanics, and I wanted to bring in some of these concepts. But initially, you know, people hadn't even done something like the water dimer. So two water molecules, hydrogen bond, you know, this is really representative of like a protein-drug interaction. You need hydrogen bonds, right? And then also methane, you know, if you pour oil on water, it kind of gloms up. It's these, you know, hydrophobic interactions. But nobody had done, like, even the simplest, like methane dimers. So we did this research. We published it. And the other thing is, what's important in biology? You know, what's something that's essential off your mind? We're in water. Everything's in aqueous solution, right? And so we did… I didn't know this was going to be a quiz. Yeah, I'm going to have to quiz this guy, you know?

DANIEL NEWMAN: 

Definitely. I've got to keep going. It felt like I was back in college. Yeah. My teacher called on me when I was napping. 

KENNETH MERZ: I

Yeah, yeah. You did good. You did good, wasn't going to let you hang.

DANIEL NEWMAN: 

I was sitting there thinking a minute, what do I want to talk about?

KENNETH MERZ: 

Water, right? And so we published a paper where we looked at small molecules in an aqueous solution. During that period, we discovered this method called embedded wave function theory. But the tech isn't really particularly important. What's important is that this allowed us to take a molecule, say it has like benzene. It's got six carbons, six hydrogens. We could take each individual atom do each calculation independently and then stitch it all together. So from the quantum device perspective, we're just putting in one carbon and calculating the wave function on just one atom. So this is something they were doing years ago, like lithium hydride. So we could do atom by atom and then we stitch it together and we get the result. And the result matches the full calculations that were done classically, right? So, we did it on, like, cyclohexane, some small molecules internally for testing, and then we thought, well, let's go for a protein, right? And the smallest protein I know that folds is called trip cage. 300 atoms, right? And we basically had a colleague who had simulated this, and he had the ground state, so the folded state, and the unfolded state, which is higher in energy, right? And we went ahead in the gas phase, unfortunately. We did the calculation on this trip cage with 300 atoms, and it worked amazingly, right? And we got very good results, you know, matched classical methods reasonably well. And then we were sitting around like, well, what's next, right? And we did this pretty quickly. It would have been a struggle to do it all at Cleveland Clinic and IBM. So we needed a partner, right? And the partner turned out to be Reken. So I had visited Reken many times over the years. I have some good friends there. And reached out to them through IBM. And they were like, yeah, sure, we want to work with you. And then it became, what do we do next thing. And we didn't really know how big we could scale, which is interesting. And I'll come back to that in a second. So we decided to pick two compounds, benzamidine trypsin. So the reason we picked that is because it's these charged interactions that cause the binding. A molecule comes in because this is positive and this is negative, and it interacts through hydrogen bonds, right? Then the other one was T4 lysozyme binding butylbenzene. It's purely hydrophobic. It was a design pocket by a colleague of mine at UCSF. Brian Shrinkhead is his name. And so we felt this represented the hydrophobic interactions for drugs and hydrobond type interactions. So you think about, you know, the common drug Lipitor. Lipitor's got hydrophobic pieces that interact with the receptor, and it's got hydrophilic, hydrogen bonds. So this, you know, we sort of split the two, right? And that was sort of consciously done. And we wanted to do it in explicit water. So we basically put these molecules in a sphere of water. And we did the very same thing. We took these 10,000, 13,000 atoms, each individual single one, we run it on the quantum device, and then solve for the electronic structure. And it worked beautifully. There's some technical things, the binding energy, because we could compute the binding of the ligand for the receptor, was a little more positive than it should be. And there's some technical reasons for that. But that is how we got to where we are. That's sort of the long story. And, you know, people always say that quantum computing can't scale. Well, guess what? You know, here we are. We can scale. You just did it. You just did it.

DANIEL NEWMAN: 

Which, by the way, you had some news here at IBM Think. You kind of alluded to the partnership, but talk a little bit about where it's at right now and what you were able to announce here today.

KENNETH MERZ: 

Yeah, so we announced this very large scale calculations, the 10 to 13,000 atoms. It was a collaboration between Cleveland Clinic and IBM and Reken. A really wonderful collaboration. It's just a dream to work with IBM. It's a dream to work with Reken. All these talented people, plus my team. Also, Reken provided basically two full days of their state-of-the-art computer just dedicated to this. And IBM Kobe and IBM Cleveland provided hundreds of hours of quantum computer time to do what was necessary on the quantum device. So that's what we're reporting today, the first large-scale scaling of a protein molecule in explicit water.

DANIEL NEWMAN: 

Yeah, it was great at the keynote.

KENNETH MERZ: 

The slide, yeah.

DANIEL NEWMAN: 

Yeah, they were talking a little bit about that and just how much moving through proteins and how much this can become part of drug discovery, which has been one of the big promises of quantum is obviously we've been trying for a long time with HPC to do this sort of, but it's very deterministic in the number of, of variables. Quantum is particularly useful at solving these real-world, in-the-world, physical problems. And that's what there's a lot of, like I said, problems. And of course, it's going to be quantum and, you know, and that was something they talked a lot about. It's quantum and HPC.

KENNETH MERZ: 

Yeah, that's right.

DANIEL NEWMAN: 

It's quantum-centric supercomputing. Exactly. I think that's exactly how Arvind said it, too.

KENNETH MERZ: 

Yes. You guys aligned on that outside? Yes. Okay, great. Basically, we're using the Quan device for what it's incredibly good for, and that's sampling these very complicated electronic structure spaces. And then the Classical is very good at post-processing. What do they say when there's more unknown?

DANIEL NEWMAN: 

Quantum can be pretty… Yeah, that's right.

KENNETH MERZ: 

You know, for example, if you take something like propane, just everyone knows a protein, right? It's gas used, right? It's three carbon atoms. So to actually compute that totally would take 1.2 million of these particular objects to describe how the electrons are distributed. But the problem is, is only maybe 10,000 are useful. And what the quantum device does is helps you sift through that, and it's probabilistic, and so it's only pulling off the things that are most relevant. So the idea is that eventually it would just pull cleanly off only the most necessary.

DANIEL NEWMAN: 

Yeah, absolutely. So as you think about quantum, right, there's the hardware, there's the algorithm, there's the quantum-centric classical implementations. What is out there that's sort of caught your attention that you're most excited about in terms of these recent advancements?

KENNETH MERZ: 

I mean everything. Where do I start? So basically I think maybe the way to say this is the framework is what really excites me the most. Why do I care about the framework? So the framework, I can basically plug anything I want into this framework. So let's, for the sake of argument, say there's five individual steps. So I can change each one of these. I can bring in AI to maybe accelerate this step, right? And it's also hardware agnostic. So the quantum part, if any other vendor wants to do this, it's basically agnostic. They would just need to provide the samples from the device. along the same workflow. So the workflow provides like a framework where we can pursue not only drug discovery, like protein ligand complexes, we can look at batteries, we can look at polymeric materials, we can look at any number of things, photovoltaics, I can't pronounce it, photovoltaics, whatever. But you know, so you can look at any range of compound types. It's not just for this particular application. So if you really put me, you know, you got to pick one thing, that's what I would pick, I think. Because it allows me to just innovate, you know, all over the place off this framework.

DANIEL NEWMAN: 

And you've talked about how some of the quantum breakthroughs have enabled molecular modeling. And I sort of alluded to drug discovery, which is building on that. But how do you sort of see all this progressing into modeling, discovery, personalized medicine? I mean, there's so much to be excited about here.

KENNETH MERZ: 

Yeah. So basically, this is showing we can scale. We have also published, and that's not today's announcement, but we've looked at coupling quantum with what are called free energy calculations. So these allow us to predict how the free energy change when you bind a drug to a receptor. So this is something I'm getting pretty excited about now because we have this new fast framework that now we can do even more work in that area. So we have published two archives on that. One will be coming out in a journal soon. So that's where I think this is lead optimization. So you think about, again, take Lipitor. Say we want to get to Lipitor, but we have to change a hydrogen to a methyl, right? And we can computationally change that hydrogen to methyl. And I can tell you by Change that to methyl the binding free energy increases and makes it a better drug, right? And so pharma companies use these free energy methods all the time. I mean these thousands of these are being done at every company. So now what we the reason we can bring the quantum in the quantum would then improve the accuracy. I am over the classical approaches, right? And so that's, I think, one area where I'm really excited about. But you can also imagine using this to model metabolism or toxicity endpoints using these same kind of strategies.

DANIEL NEWMAN: 

So if you wanted to boil it down, though, for everyone out there that sort of does say, you know, you guys are researching, you're out in the sciences, you're speculating on what quantum might help you do. Right. What is quantum truly, like, if you kind of want to simply explain it, like, what is it helping you do today?

KENNETH MERZ: 

Yeah, that's a really great question. So I would say, you know, the thing that people don't appreciate, this guy Dirac in the 20s, he said basically quantum mechanics, It solves all of chemistry and most of biology, right? But the equations are too complicated. So this is like pre-Turing, you know, pre-information science, right? So he didn't know all this stuff was coming, and he was exactly right. But what's happened now is we've now been able to start solving these equations. So quantum mechanics is the bedrock of computational chemistry biology. So if you take quantum mechanics out, we're just, you know, a random field, right? It's just, it's like AI, right? It's dependent on the data set. But I can do these ultra high-level calculations, and I can get data that's as good as an experiment, that then can be used to inform machine learning. So that's why it's so important that we achieve this kind of high accuracy using quantum mechanics, because it's just fundamental to the whole field. And it informs everything.

DANIEL NEWMAN: 

Because you have more logical, fault-tolerant qubits, the opportunities to the point of what's going to become possible.

KENNETH MERZ: 

Yeah, but it's an interesting question, right? Sarah Sheldon and I were just sort of talking about this. So again, within this framework, we could imagine a scenario where, say, the active sites over the jug is bound. could be done with a fault tolerant device. We could do the rest of the molecule using say a force field or with like a noisy machine, you know, because the information here doesn't need to be as high of accuracy as the region around the drug binding site, right? So I think it's really interesting how it's gonna shake out in the coming years because we will have such a device in 2029 or so. And we'll be able to start doing some of these experiments and seeing what works the best, right? And, you know, next week, somebody could have an announcement that, you know, it makes this irrelevant anymore, right? So that's how fast it feels moving. And I can give you like one story on my side. We were trying variational quantum eigensolvers, so an optimization method. It wasn't working. The little hair I had, I was pulling out. But then on Friday, IBM gave us a pre-taste of this subspace quantum diagonalization. So Friday night over the weekend, I was talking to my guys, and we just dropped VQE and went 100%. So literally just changed the whole research of the group based on one breakthrough. So there could be another one down the line that I don't know about.

DANIEL NEWMAN: 

I mean, nobody knows, right? It's also just, we're in such exponential times, where between Friday and Saturday, we're seeing it with AI models, and just the amount of compute that's being required to be deployed. In this whole boom, the entire world, the whole data center boom, the whole world underestimated it at every step of the way. They say we're just not particularly good at thinking exponentially. We tend to think very linearly, so we look at things. That's what folks like you are in the lab doing, right?

KENNETH MERZ: 

Yeah, I just had this goal when I interviewed. Maybe they laughed at me, I don't even know, because I said I wanted to do quantum computing, a protein, and here we are. We did it.

DANIEL NEWMAN: 

So if we go five years into the future, what does the success of a quantum program at Cleveland Clinic look like through your lens?

KENNETH MERZ: 

Yeah, so it's going to be more on the healthcare life science side, but I think also you need to just improve the methods, say in the gas phase, things like that. So to me it would be maybe ultra high accuracy free energy calculations where that would then directly plug into drug discovery. at every company in the world, right? And so that would be, I think, a big breakthrough. Another one would be data sets, right? So that can be used in machine learning. So we could create millions of highly accurate calculations that then could be used to inform machine learning potentials, right? It would be another big win, I think, again, in the life sciences area. So this could be like, take every known protein, ligand complex, you know, calculate millions and millions of interactions of it to the receptor at the best possible accuracy possible.

DANIEL NEWMAN: 

It feels like the speed of development, you know, we sort of have inflected in some ways and you're seeing, you know, from idea to FDA approval, for instance, shortening in some cases, but it does seem so opportunistic. There's so many diseases and different It can inform everything. So, I mean, the drug discovery process, right, there's sort of like

KENNETH MERZ: 

Lead discovery, lead optimization. So I think where this would incredibly benefit is lead optimization. So instead of making 1,000 compounds to get 10 that are promising, maybe they only make 100, but get 50 that are promising. So that's the kind of return. And then it could be preclinical where you could determine if something would first pass, degrade through metabolism, or turn you purple because it has a toxic side effect, right? So that's where, and then, but the clinical trial side is where you spend time with patients, dosing them. So I'm not sure. How do you speed that up? Yeah, from what I'm doing, right? But it could be that, for example, Serpol talked about this. like using these kinds of things to patient select, select optimal patient pools, right? And that can be done with quantum computing, but it's more of an optimization, not a quantum chemistry problem.

DANIEL NEWMAN: 

Yeah, but it is really interesting how we can, you know, how we can go faster, but safely. It gets me thinking, too, about the quantum-centric, where quantum advantage comes into play, where even just being able to more quickly, like simulation, like how can you more, you know, simulate biology? To your point, if you could actually have, you know, like what we're doing in many industries, like engineering, where you can simulate the bridge, you can simulate the buildings, build and understand the structure. Of course, you got to be, humans are different. Can't just say, oh, it worked in the computer. Let's put it out. But it is super interesting. Look, Kenny was super excited to talk to you. This is really interesting. Congratulations on all of the progress that you're making. And I do look forward to continuing to follow the journey over at Cleveland Clinic. Sounds great.

KENNETH MERZ: 

Enjoyed it. Thank you so much.

DANIEL NEWMAN: 

And thank you, everybody, so much for being part of this 6-5. We are on the road here at IBM Think 2026. That was a great conversation with Cleveland Clinic. So much going on. But the verdict is quantum is now. Quantum is real. And this is one instance, one example of where that is the case. Stay with us for so much more. We will continue to cover this topic and a lot of other things on the 6-5. See you all soon.

‍