Thomas Ehmer – Merck KGaA

Thomas Ehmer from Merck is interviewed by Yuval Boger. Thomas discusses the rapid advancements in quantum computing architectures, recent breakthroughs in logical qubits, and quantum communication protocols. He describes potential quantum applications in logistics optimization and material sciences, highlighting collaborative efforts within initiatives like QUTAC. Thomas contemplates the necessity of in-house quantum computing resources, reflects on Germany’s national investment in quantum technology, and much more.

Full transcript

Yuval Boger: Hello Thomas and thank you for joining me today.

Thomas Ehmer: Hi Yuval, pleasure.

Yuval: So who are you and what do you do?

Thomas: That’s a good question Yuval and because I know the podcast, I prepared a bit. I’m a bit of a superposition and it depends really on whom you ask. So for my band, I’m the keyboard player, for my kids, I am the dad. But I think for this podcast, I’ll tell you what I’m doing in my business. I’m one of the co-founders in the Merck group, Merck KGaA. We are not Merck & Co. so there are two different independent companies. We have to tell this every time. So if you listen from the US, we operate under EMD Serono. Rest of world, we are Merck KGaA. And some eight years ago, I was looking for novel technologies. I’m still looking for novel technologies and we did found a quantum task force with the objective to check if it makes sense that our company takes care of quantum technologies. It does. Yeah, and then I’m a physicist by training. I look mainly at all applications that are novel to our company and should make sense. I’m a bit of a troublemaker because I bring technologies in-house that we don’t know yet how to use or we want to use but don’t have budget and so on. So that’s what I do for a living. Eight years,

Yuval: wow.

Thomas: Yes.

Yuval: Really one of the pioneers. So let’s go back eight months. What have you learned in the last eight months about quantum and quantum readiness that maybe you didn’t know before that?

Thomas: Eight months is February. What did I learn in the last eight months? So I think what is confirming is that we have a very good emerging quantum computing community. I very much like all these conferences and it’s always good to see that we, independent on whom you ask, it seems that we are a community who wants to work together. Yeah, and I think this is, as some might say, we approach quantum winter. We don’t know. But I think it’s good to see that there are still some brave people who carry the flag and work together. I see also that a lot of novel quantum computing architectures make it close to the market or are close to the market. So it’s really good to see. And I mean, personally, I was keen to see the recent announcement from Quantinuum, for example, where they have, for example, logical qubits apparently entangled. Eight years ago, we didn’t even know how to build a single qubit. Now we have logical qubits, at least in some claim that they have, and I trust them. So this is interesting to see. And also the quantum communication part, people are really getting keen on all these quantum key distribution, communication protocols and so on. So that’s, yeah. There’s not a single thing. I think it’s just really exploding. That’s good to see how far we move.

Yuval: In your company, what do you see as the key application for quantum? Is it drug discovery or is it more things like optimization, supply chain and so on?

Thomas: It depends on whom you ask. So we are super interested in quantum applications because once they come, if ever they come, they might be transformative. There’s so much hope and hype on drug discovery. Personally, I’m not really a big fan or believer that computational drug discovery is outperforming wet drug discovery. So that’s my personal take. I think it’s a partnership and everybody’s betting on artificial intelligence. So drug discovery is a very complex thing and computational methods. Quantum computing is an additional computational method. So I think it’s a long way. There’s a very good paper out from the colleagues from Boehringer who did investigate if quantum computing ever will work. So it looks like we need new algorithms. So that’s my take. And the other problems, logistics problems, this is how it started for me actually to dive into quantum computing. So some years back, the D-Wave were the first ones on the market with their traveling salesperson optimization. I still think it’s a miracle that we have these annealing things. And apparently, so these ultra wicked complex problems might benefit also from quantum computing. I’m a bit undecided. I mean, we see the three classes, simulation, optimization, and machine learning. And yeah, for our company, none of them currently is relevant from an application perspective. But if they will come to maturity, likely the material sciences is also of important.

Yuval: It’s very popular these days to talk about chat GPT moment, a moment where AI moved from science fiction to we got to do it now. What do you think the quantum chat GPT or the quantum GPT moment is going to look like?

Thomas: The quantum that I have no clue. So I think the GPT moment was when everybody and their sister and brother were exposed to a functionality they could never imagine, which is mimicking human intelligence. Actually, it does not, but it’s mimicking and it’s very convincing. So I think we will see how this develops compared to quantum. I have actually no clue. I could imagine that eventually something like this communication protocols might have an impact if everybody could just plug in whatever app to have a quantum secured point to point communication without any agency being able to listen in without me knowing they listen in. So this might be eventually one of these, which has mass impact because I think quantum computing per se is rather an expert-ish niche co-processor thing. So I’m not sure how to. It’s a good question. I have to ask my guys what they think.

Yuval: You mentioned community earlier. And I know you’re part of a community of several, I think, German companies that work together to develop quantum. Could you tell me a little bit about that? And also related to that, are you worried about other pharma companies, about collaborating with other pharma companies to do quantum or right now, a rising tide lifts all boats?

Thomas: Yes. So the yes is to the last question. So the company you mentioned at the beginning, so it’s an association called QUTAC, stands for German Quantum Technology Application Consortium. We have a very cool homepage. And it’s an association of the 13 big German companies. I don’t want to mention them all because otherwise the podcast would be over. What we are doing, we organize ourselves in several streams and with the focus to identify quantum technology use cases that we could work on either as developers or co-developers or application user guidance for the developers. Independent from our own, let’s say business. For example, we work together with automotive or pharma and reinsurers, logistics companies on, let’s say, logistics problems. Everybody has their problems and we test out different approaches on knapsack, etc. So that’s one thing. And the other thing is we share the burden to evaluate vendors, potential service providers. And by being attractive enough, if as a partner you manage to convince QUTAC, you not only have one customer, but you would have 13 customers. So that’s the main point. And also I think it’s very helpful. All of us have relatively small quantum teams. So these are big worldwide corporations. I mean, BMW, SAP, B├╝hringer, etc. And all have relatively small quantum teams. And it’s very comfortable for us as a team to not be alone so that we have a bit of a peers and we can share interests. Regarding the pharma guys, I mean, it started I think at the Q2B several years back. There was an initiative that we all say everybody claims we can speed up drug discovery or drug development or we can bring better drugs to the world by quantum computing. Who knows how this works? Nobody did. So it’s also very open, very collaborative. And I think it’s rather a friend-ish approach, bosonic approach that we work together instead of competition because currently it’s really pre-competitive. There’s also the Pistoia Alliance who is setting up one pharma consortium thing. There are several pharma consortiums emerging. So that’s quite open. And the third part I forgot. Now, I think you answered everything here.

Yuval: When you think about QUTAC, what is the issue that you’re facing right now? What is the key thing that you’re looking for a solution?

Thomas: So we have different streams. So one stream we started focusing on the logistics and the material sciences streams. It looks like currently we have reshuffled a bit. So one part of us is looking at generative components. So we’re working on quantum generative network, so QGANs, to check in how far they perform better than classical GANs. We have a paper out. You can make sense yourself if it’s worthwhile or not. I mean, it’s early days, but this is where we jointly work on because everybody is interested, of course, in these generative networks. The other thing is in the logistics part, we work on knapsack problems to understand can quantum do better knapsack things. And there’s another stream which is composed of different people. They work mainly on material science topics, so quantum chemistry using quantum computers. I don’t have too much insight in that, but it’s always shared problems that we say, how can you better do ground energy finding, whatever. What is quantum Monte Carlo? Where does it make sense? Where does it not make sense? And the third stream I mentioned is really looking at infrastructure from different perspectives. So we are also writing there some papers with the target audience of being other quantum computing users to guide them if it makes sense to go on neutral atoms. I mean, you know, neutral atoms or we’re just working on an ion trap paper. And the other part is also political communication. And in principle, also, we wanted to do a bit of education, so help grow talent for our companies. But that’s not so far progressed as we would have wanted just due to the lack of resource we have.

Yuval: You mentioned the comparison between computational chemistry or using quantum computing for computational chemistry versus wet labs. In a company like Merck, how much communication is there between the people who are actually working on developing drugs and people like yourself who are dealing with future technologies?

Thomas: That’s a good question. So I think there is enough communication. There is even a common ground. There are just not enough solutions. I mean, we have recently discussed the potential to do, I don’t know, what was this, the idea to use quantum computer annealing approaches to find binding pockets in complex targets. And when I talk to our researchers who do the wet stuff or other computational stuff, they said it sounds interesting. On the other hand, we already have 15 methods or 20 methods and I have 15 people, so I just don’t have capacity to digest this. So I think there is an appetite that it’s really tough to understand this added value given that you have new ways of thinking. So quantum algorithm is using quantum features, nobody really understands what these are, even the quantum guys don’t. And then, I mean, how to translate your existing problems in a way that you can make additional value for using quantum components or quantum core processor. That’s not easy. And unfortunately, most have objectives that are not so much forward looking and have time to think what will be happening in 10, 5 years because in the end, you need to deliver the next target. And it’s just a bit of operational pressure. I mean, it’s a bit of a dance and mainly of interest, so people listening with interest, but it’s just currently it’s more playful with the intent to detect good things, but it’s not a threat currently if we would not look at it, I think.

Yuval: Do you see a need for a company like Merck to purchase a quantum computer anytime soon in the next few years or everything is on the cloud as far as you’re concerned?

Thomas: I mean, there are different answers to that. I think it depends on what you want to achieve, so if we want to contribute as a material supplier, for example, because we have the very best isotopes and the very best gas to do nitrogen vacancy, whatever stuff, well, so do we need a quantum computer for that? Maybe not. If we want to learn to program it, there are so many solutions out in the cloud. I doubt that we need one in-house. And then there’s, on the other hand, the approach that the American clinics currently are doing. So every clinic who thinks they have interest in novel technologies, they put a quantum computer on site, which is then I think rather a symbol because if every day you pass a quantum computer, eventually it triggers your way of thinking, what could I do with it? So I think for this sake, more as a symbol, it could make sense. And I mean, it looks good. The machines, depending on which machines you use, but I think definitely┬á from a productive use point, it doesn’t make sense this moment in time. I would not even know whom to hire to maintain and operate these things because they are very fragile and you need to calibrate all these. And then the question is also which technology do you want to have? So you would need to buy one of each flavors, modalities, how they’re called. I would not know what to choose. So therefore, I think it’s also not from a financial investment. Maybe it’s not a good value investment because the next generation makes the other one obsolete. So it’s not that it’s getting more and more from a collector’s viewpoint that you can collect. I have the very first, I don’t know, a QuEra machine, and then I put it somewhere. I have the very first IBM machine with 127 qubits. I think it doesn’t make sense.

Yuval: Germany is one of the countries that’s investing heavily in quantum on the national level. And do you feel some of the results of this investment or do you think it’s going to be years out?

Thomas: So personally, I feel some of it because I’m also profiting from a federal funded project, which I’m very grateful for because otherwise we would not do this project because it’s just too risky. And because we have this funding, I can spend part of my time to think on things I would not think for. And the interesting part is that, I think we heard this on several talks also on conferences that because you think on quantum algorithms, you detect sometimes solutions that are completely independent from quantum that help your business nevertheless. So that’s one aspect. Do I see other… I mean, Germany spends a lot of money. The UK does, France does. Others do too. I can’t really say that I have seen quick returns, but I see that there is a lot of attention and even a bit of understanding in the decision makers. It’s moving into a direction which is at least they want to understand what it is and it’s quite complex, but I think it’s more technology friendly these days than it was in the past. Unfortunately, we have the situation in the world. So I think everywhere funds are cut for fundamental research. But I mean, that’s I think independent from the aim that we want to spend much money on that. I mean, one thing on the other hand is what I find a bit disappointing is that we have a bit of this fragmentation in Germany. We have all the land and everybody wants to do their own valley. So we have a quantum valley in Munich, we have a Saxony valley. So in Germany, everybody is a bit of building their own center and everybody is in the center of their world. I’m not sure how this is in other countries. I think France is by definition different that everything goes to Paris and then you have to see what’s coming out there. But the original question is, yes, I think it makes sense. You slowly see. I think we’re a bit far behind. Others started earlier and have more mature startup scenes, but we’re catching up.

Yuval: You spoke earlier about the quantum winter and also mentioned that you’ve been doing this for eight years. Do you think there is a risk that a company like Merck says, you know what, we’ve been at this for a while. Let’s shelve it and come back in three years and maybe things will be different.

Thomas: I think there is in principle the risk because we are not doing this as a profession. Everybody’s doing this as their hobby. We have a very small team under our digital, just forget, but we have a central function and we have a digital unit. And somewhere in this area, we have a unit who is doing quantum and AI lab. And these are not many people. And they spend a very small part of the time doing what they do. And what I’m doing, thanks to my boss and my boss’s boss and everything, so I am doing this on top of my job. I have zero budget except the federal funding, so there is not much to cut. Because otherwise, it’s late in the evening, we have this call after business hours. So I think many people are committed despite not having money because it’s not like in other companies where you have a really a small team and hire a few students. For us, it’s relatively small. And I rather see that we are doing more than is visible to the top management. And when we did this task force, people were surprised how active we are. If you’re not in the field, of course, you don’t see it. But then when you say we should do something about quantum computing, you see that we’re relatively well networked. I think that’s… Does this answer the question?

Yuval: Absolutely. So hypothetically, if you could have dinner with one of the quantum greats, dead or alive, who would that person be?

Thomas: I’m prepared, so I’m not so… I don’t fake surprise. And it took me a long time to think. And I’m in a very lucky position that I already had lunch with some of my idols, they’re alive. Therefore, I go for the dead ones. And this is a tough thing. I think I would go for David Bohm or John Wheeler, because I really like the way they put more into quantum, which has to do with consciousness, research, self-perception, the miracles of the observer, what is the observer, so the bigger picture. Currently, I’m tending in my philosophical interpretation of the word a bit to the monistic view, which is maybe a bit over-exaggerated. But the miracle that we have something which is self-perception and we have something which is an external world. And how come? We don’t know. And there are so many different interpretations and I think Bohm did a very good… with his implicate order, he had a very good idea to think about that we need new ways of expressing ourselves, if I understood this correct. And I’ve seen some rare interviews back on YouTube with him and he seems to be a very super interesting person. But he’s dead, unfortunately, so no chance.

Yuval: Very well. Thomas, thank you so much for joining me today.

Thomas: Yeah, thanks for having me. It was a pleasure.