HPE World Quantum Day | Tom Beck, Oak Ridge National Lab

Clips
More from HPE World Quantum Day

Tom Beck

Section Head for Science Engagement and Acting Group Leader for Quantum-HPC

Oak Ridge National Laboratory

Dave Vellante

Co-Founder & Co-CEO

SiliconANGLE Media, Inc.

play_circle_outline Frontier at Oak Ridge: Tom Beck, Science Engagement Head and Acting Quantum HPC Leader, on Open-Science Supercomputing

play_circle_outline HPC, AI and Quantum Convergence: GPUs, AI-Designed Circuits, Error Correction, and Quantum Accelerators

play_circle_outline National security concerns: post-quantum encryption and quantum networking implications.

play_circle_outline Funding/readiness dilemma: when to invest large-scale in quantum research and infrastructure.

play_circle_outline Vendor qubit trade-offs: superconducting (fast), ion traps (high fidelity), neutral atoms.

play_circle_outline Large-scale simulations on Frontier: 455 billion-atom water simulation equivalent to bacterium-sized cube.

play_circle_outline Quantum networking for scalable architectures versus monolithic million-qubit systems.

Info
Transcript

Tom Beck, Oak Ridge National Lab

Tom Beck

Section Head for Science Engagement and Acting Group Leader for Quantum-HPC Oak Ridge National Laboratory

Dave Vellante

Co-Founder & Co-CEO SiliconANGLE Media, Inc.

HOST

In this interview from HPE World Quantum Day 2026, Tom Beck, section head for science engagement and acting group leader for quantum-HPC at Oak Ridge National Labs, joins theCUBE's Dave Vellante to discuss the path toward practical quantum computing and how hybrid models integrating quantum, HPC and AI are being put to work on some of science's hardest problems. Beck, speaking in front of Frontier — the world's fastest supercomputer for open science — frames quantum not as a replacement for classical HPC but as a specialized accelerator. He explains how Oak R... Read more

explore Keep Exploring

What is the role of the section head for science engagement and acting group leader for Quantum HPC at the facility, and how is their time typically spent? add

How does quantum computing connect with high-performance computing (HPC) and AI, and how are these technologies expected to coexist and integrate? add

Is quantum computing becoming a near-term concern for businesses and national security, and how might it be applied (for example in cybersecurity, error correction progress, and quantum networking)? add

When should large-scale funding or an Exascale-Project–style initiative be launched to accelerate quantum computing, and should we replicate that model now? add

What is the process for developing, evaluating, and selecting a new large-scale computing technology (e.g., an Exascale supercomputer or a quantum computer)? add

What scientific fields will the Frontier supercomputer be used for, and how large of a molecular (e.g., water) simulation can it run? add

What are the main practical challenges to building and scaling large-scale quantum computers? add

bolt Powered by CUBE AI

Tom Beck, Oak Ridge National Lab

search

Dave Vellante

>> Welcome back to World Quantum Day, our celebration that's made possible by Hewlett Packard Enterprise. I'm Dave Vellante, and we're here at Oak Ridge National Labs, the iconic site, the original Manhattan Project. But now we're doing amazing ... well, we're observing the amazing work the scientists are doing here. We're in front of Frontier, which is the world's fastest computer for open science. I'm here with Tom Beck, who's the section head for science engagement and the acting group leader for Quantum HPC. Tom, thanks for spending some time with us. Really appreciate it.

Tom Beck

>> Very happy to be here.

Dave Vellante

>> So that's quite a title.

Tom Beck

>>

Dave Vellante

>> What is that all about? What's your role here at the facility and how do you spend your time?

Tom Beck

>> So I was in academics for a long career at the University of Cincinnati. And my field is computational chemistry. And I heard about this opening for a section head position in the computational facility. I had been a postdoc at Los Alamos National Labs, which really initiated large scale computing for science. And I saw this as a great opportunity. So I applied for this job, and one thing led to another, and I got to join the lab. So the purpose of our section is really to be the science outreach group for bringing in projects and partnering on projects to run on Frontier in very wide range of topics. So my background is chemistry and physics, but there's particle physics, fundamental quantum chromodynamics calculations, astrophysics, all the way up to engineering and biology.

Dave Vellante

>> Okay.

Tom Beck

>> So we partner with those, all the people coming in.

Dave Vellante

>> So this is one of the fastest, if not the fastest computer in the world. And of course things change very quickly.

Tom Beck

>> Yes.

Dave Vellante

>> And there's a lot of discussion today in the industry about the bridging from HPC, to AI, to quantum and how those all fit together. Can you explain what you're doing today with quantum computing? How does it connect with the work that you're doing in HPC?

Tom Beck

>> Well, quantum is a rapidly-growing capability. We see that as the next frontier in high performance computing, but it's likely to be, at least initially, a more specialized accelerator of certain computational workloads. GPUs have been adapted to take on really most forms of computing up to this point, and they are really the heart, blood of AI computation. Quantum computing, now we're trying to figure out how to link quantum computers to HPC so that we can download certain parts of the calculation onto HPC machines to accelerate those, but do the really hardest quantum part or something that can be accelerated by quantum on the quantum device. So it's really a game of transfer of information. How do you accelerate the flow of information between the two machines and do the hardest thing on the quantum device?

Dave Vellante

>> I mean, in the technology industry, especially as a media company, we like to think in absolutes, "Oh, this technology is going to replace that technology," but it seems that we're seeing this coexistence. You've got CPUs, you've got GPUs really-

Tom Beck

>> Right....

Dave Vellante

>> driving accelerated computing, you've got quantum. How do these technologies all play together?

Tom Beck

>> Well, I think we're working that out in terms of the quantum integration and AI, but there's all these interesting relationships that are developing. Like AI is being used now to design more efficient quantum circuits. The AI is done on a classical machine, but that can allow you to adapt the quantum circuits to run more efficiently on the quantum device. That's an example. So AI for quantum, but there may be quantum uses for AI as well to accelerate the convergence to the optimal neural network model for a certain process. So it's a chaotic space. Many things are happening and relationships are developing between all three of those entities.

Dave Vellante

>> And we've heard the GPUs can help with error correction and-

Tom Beck

>> Yes....

Dave Vellante

>> stabilizing quantum systems.

Tom Beck

>>

Dave Vellante

>> Can you explain that a little bit, how those

Tom Beck

>> Well, quantum ... chips, so to speak, or devices, or qubits have more errors than classical devices. And that's because of the quantum nature and that noise can influence these very delicate states. So you have to have a way to diagnose the problems that have happened, and then you have to figure out what to do based on that information. And that's actually a big computational problem in itself. At the current state, we think we may need hundreds of qubits, physical qubits, to have one logical qubit in a device. And that's a big barrier because it's not like that on a classical HPC issue.

Dave Vellante

>> There's more of a one-to-one relationship there.

Tom Beck

>> Yeah.

Dave Vellante

>> Okay. What excites you about quantum? Are there use cases, the early use cases that you all see as having great potential?

Tom Beck

>> I am a theoretical chemist and physicist and a big fan of Richard Feynman. I've done work related to path integrals that he developed. I think he was a visionary, as everybody knows, in the quantum computing space. But he really saw quantum computers as devices to model quantum nature. So I think that's really a core application at this time. I mean, national security, encryption, things like that, that's a big application as well. But my interest is more in material science and computing the properties of materials, evolving systems in time. If you have two particles colliding in fundamental physics, we really don't know how to evolve those states, those Schrodinger wave functions in time in an efficient way. It's an exponential scaling problem. So we can only look at really small systems. And a quantum device could give us the capability to scale that up to really interesting physics.

Dave Vellante

>> Which is not possible. It's hard to believe, but it's not possible with the classical supercomputers.

Tom Beck

>> Absolutely. We can only simulate on the order of 50 qubits realistically on a machine like Frontier. And after you get beyond that, you're basically trying to model two to the nth power states where n is the number of qubits. If n is 300 qubits, that's the number of atoms in the entire universe, 2 to the 300 power. So that tells you something about the available computational space that a quantum machine could give you, even though it's very hard to access all of that state space, because you have to build that up with gates and many qubits.

Dave Vellante

>> So the potential is enormous. When I speak to business executives, they're just drowning in AI right now, trying to figure out how to make agents not go rogue. In many ways, they get a little bit more stressed out when they talk about quantum. Now, of course, they see quantum as potentially a defense in cybersecurity. They're concerned about post quantum encryption and exfiltration of data today that's encrypted. And so they're focused on that, but it's hard for them to really apply the potential to their business. But quantum is coming maybe sooner than they think based on what you're saying.

Tom Beck

>> Well, I think there's major things to work out. When I came into this field ... I mean, I've known quantum mechanics for a long time, but I got into quantum computing only a few years ago. And I'm always open-minded about ideas, but fairly skeptical at the time. But over those few years, I've watched the development of the vendor industry and how they're developing new chip or qubit technologies. And error correction is making progress. So I really do feel that it is moving in a positive direction, and within a few years, we will be in a very different place. So I'm more optimistic than I was a few years ago. And in fact, going headfirst into working on this problem.

Dave Vellante

>> I mean, based on what's happening in the industry with LLMs and open clause, and other things, something new comes out every couple of months, so nothing should surprise us anymore, right?

Tom Beck

>> Right. You do not know when there's going to be some great idea on a new qubit that could allow for much more efficient error correction. And when that happens, I mean, things are going to change very fast.

Dave Vellante

>> Now national security is part of your mission here. How do you see leveraging quantum for that mission? How does it compliment what you're doing with things like Frontier?

Tom Beck

>> I mean, we are an open science lab, which is different from Sandia or Los Alamos, which are national security nuclear labs. We have nuclear energy going on here. So there is some difference there. We do have classified research going on at this lab, but it's a small part of our mission.

Dave Vellante

>> Not as much top secret.

Tom Beck

>> Not as much top secret. But clearly in the military space, in terms of information technology, that's a big thing. Quantum networking, how to transmit information securely is a major thing. We have a lot of research going on here in quantum networking as well. And I think if you want to build up a large quantum computer, there's different ways we can think about that. Trying to have one big qubit system with maybe a million qubits or having smaller units that are networked by quantum networks. So I think quantum networking is a major thing for the military as well.

Dave Vellante

>> And we've certainly seen networking and AI, and it became a bottleneck, which now is a fundamental component of every sort of AI factory and system. I know you got here when the room behind us was empty five years ago, so you saw Frontier being installed-

Tom Beck

>> From the ground up.

Dave Vellante

>> Right. And so when you think about the journey to Exascale, if you will, how do you compare it and contrast it to what we're going to see or are seeing with the journey to quantum?

Tom Beck

>> I think we are making steps in that direction. We had the whole Exascale Computing Project that was centered at Oak Ridge and all across the DOE complex where people developed many scales of software for applications and for fundamental libraries and things like that. And it had very large scale funding at that time for the ECP. So there's a lot of proposals now to reproduce that model to take quantum computing and drive it to the next level. But there is a question of readiness, like when do we invest the really big money to push the field forward? And you don't want to pick a winner before the technologies are mature enough to make a good decision. So we're at that stage.

Dave Vellante

>> A delicate balance. You don't want to be too late, but you don't want to over rotate to something that's not going to pay off the way that you hope.

Tom Beck

>> Yeah. It's a high stakes decision.

Dave Vellante

>> How does that work? Do organizations come to you with ideas and then you filter those ideas? Can you explain the whole process?

Tom Beck

>> Well, there's a whole arch to developing a new technology, a large scale, Exascale supercomputer or quantum computer. We interact with a lot of vendors across the whole space. We talk with them. We learn about their technologies. We look at computational readiness of the technology. And in quantum computing, you have different kinds of qubits. You have super conducting, and then you have ion traps, and you have neutral atoms as three examples. And they have very different physical properties, like superconducting qubits are much faster in their turnaround time. Ionic trap qubits tend to have higher fidelities, and then it becomes a trade-off. And then it becomes which application do we really want to run on this machine? If fidelity really matters, and you can scale up to a larger number of qubits on an ion trap machine, you might lean in that direction. But you have to assess what applications are we really pushing at our center, I think. Or you can get one of each machine if you have the resources for that and have different offloads for different jobs.

Dave Vellante

>> As the Brits like to say, "Horses for courses." Is there an HPC-AI quantum workflow? So like a hybrid workflow that you see emerging where the system is intelligent enough to dynamically route that work or is that just a pipe dream? Do you see it more separated and you put the right workload with the right system?

Tom Beck

>> We are really focusing on the convergence of those three now because we're betting that that really is going to be the future. I think there's a lot of hurdles to overcome in making that work. In the DOE Genesis Mission now, that's a big emphasis in DOE. And that's really one of the big pushes in that area, is to span those three spaces, make them work together. I mentioned earlier talking to you that we're starting this project on fusion energy, and the purpose is to set up a scheme that utilizes all three of those spaces to really solve a problem, not just a pipe dream, but to really apply the correct things at the correct places. And maybe I could tell you a little bit about that.

Dave Vellante

>> I'd love to learn more about it because this is the number one topic. We were in a round table, actually an analyst Q&A session with Secretary Wright and Jensen. And it was an ask-me-anything. It was fascinating for literally an hour, hour and a half. It was just awesome. And of course, you hear so much about energy and power being a limiting factor in advancing things. And if you hear about the potential for nuclear, how we're behind, say China, so it's vital for our national competitiveness. What is the state of that? Share with us your

Tom Beck

>> Well, the demand is increasing rapidly, partly due to the rapid rise of computing. That's placing demands up to perhaps 10% of electricity demand in the coming years or even higher with the growth of electrically-driven transportation. So it's not a decreasing demand, and we need more electricity. So I've always been a proponent of nuclear energy, both fission and fusion. Fission is working today, and we can come up with more efficient technologies. But this project that we're working on is fusion energy. And that's the energy like in the sun, it's due to the fusion of nuclei, deuterium and tritium that gives off a huge amount of energy within the reactor, but you have to get to very high temperatures to make that happen. But the other thing is that in a fusion reactor, the fuel, deuterium, we have plenty of that. In the oceans, there's almost infinite amount of available deuterium. Most of them are hydrogens, but deuterium is abundance

Dave Vellante

>> Plentiful.

Tom Beck

>> Yeah. Tritium is not. We don't have tritium sitting around and we have to make it. And it turns out we're going to be making it in the reactor itself. The resource that you need to make that is lithium, which also goes into batteries. And you don't need a small amount of lithium, you need a lot of lithium to make this work, because if you want a one gigawatt reactor, you need to produce about one pound of tritium per day to run that reactor. That doesn't sound like a huge amount-

Dave Vellante

>> No, it's a lot. I understand ...

Tom Beck

>> but it's a non-trivial fraction of our current store of tritium in the U.S. today.

Dave Vellante

>> It's a whole nother dimension to the supply chain.

Tom Beck

>> Yes.

Dave Vellante

>> It's a very complex vector off of what you would think of sand goes in and silicon comes out, but there's a whole nother

Tom Beck

>> Right. It's very complex. So this project we're working on that uses AI, and quantum, and HPC is to design better molten salt materials that can convert ... take the neutrons from the reactor, hit the lithium, create tritium. But then the question is, what happens to the tritium in this complicated liquid and how do you get it out?

Dave Vellante

>> Is that a known unknown or-

Tom Beck

>> It's an unknown at this time. I mean, there are ideas. We have ideas. We're thinking about ways to alter the chemistry in the material to get it into a form that will come easily out of the liquid. These are high-temperature liquids, molten salts, which were actually developed here for nuclear reactors at Oak Ridge. And the tritium, we want it to form molecular tritium, which is a gas, like hydrogen gas. And then it might bubble out of the liquid and we can more easily collect it and get it back into the reactor. But this is a big technological problem. The pieces are not that difficult to think about, but getting it all to work together is really tough.

Dave Vellante

>> So coming back to nuclear, so we hear a lot about fission, small nuclear modular nuclear reactors. Fusion is many years away, perhaps even decade plus, not really sure-

Tom Beck

>> Yep....

Dave Vellante

>> right? So are you optimistic, particularly about the near term with nuclear fission? There's obviously regulatory concerns. The current administration seems to be trying to facilitate that.

Tom Beck

>> Absolutely.

Dave Vellante

>> It's a key competitive dynamic in the world stage, isn't it?

Tom Beck

>> Absolutely. Yeah. I mean, I think we can make safe fission reactors. If you really analyze the history of energy, we have had some bad accidents that have gotten a lot of attention as they should, but in terms of human health, the impact of nuclear energy is probably the safest course of energy that we have had so far for human impacts. So we have to be very careful, but we have the engineering in place to handle these things, I believe. For fusion, I mean, there are definitely radioactivity issues that occur there too. You're producing tritium, which has not as long of a lifetime, but you have to deal with tritium. Tritium will diffuse through metals and it can be difficult to handle that atom. And so you have to take accountancy of that, but you also ... It's not that difficult of a problem to encase that in something that is not dangerous. But the real problem is technology for how to build the thing and make it work.

Dave Vellante

>> It's interesting, Tom, that you said you've always been ... I think you said you've always been a fan of nuclear. And I remember just sort of when you think about the era of the '70s and the stigma that was attached, I remember one of my engineering professors, thermodynamic class, I remember it was very negative on nuclear, almost scaring us as students. And that sort of sentiment sort of held us back, hasn't it? And now we're-

Tom Beck

>> I think it has, yeah....

Dave Vellante

>> at a competitive disadvantage. The work that you're doing here hopefully can catch us up. Can we close the gap on the world stage with say China who has an abundance of power?

Tom Beck

>> They're moving rapidly, both on fission and fusion. But I taught a course at the University of Cincinnati on overall energy concepts. And I wanted to really learn. I mean, even a lot of physical scientists don't really know the details of world energy, the sources, the consumption, all of this. And I taught this course for three years, and I went through cycles of despair and excitement about the possibilities of energy. And I came to realize that really we need nuclear energy to supplement at least for say a hundred years or a couple hundred years to get us through a transition to other forms of energy, I think.

Dave Vellante

>> Well, the potential is incredible, but you know what's my big takeaway here? Is that Oak Ridge, you are on the cutting edge. You have to be in the cutting edge. Yep. How do you stay on that cutting edge? Obviously, you got talent here, you got big brains, there's potential collaboration, there's obviously support from your sponsors, et cetera. But I wonder if you could explain that dynamic.

Tom Beck

>> Well, this is why I decided to come to Oak Ridge. I mean, academia offers really excellent opportunities. You can pursue your ideas, but you tend to be in a more isolated space. And you come to a place like this, and here I am interested in fusion, but I'm working on the chemistry of fusion. I can go talk to fusion experts down one hallway. I have the world's leading computational scientists here that we can work on applying AI and quantum to this problem. And a lot of people come into this space just to work on Frontier. So you get partnerships in so many wide different ways that it's really an exciting space. Part of the problem then is how to keep focused. We have too many things going on and we need to really keep our eye on the prize. And I was really interested at Los Alamos. I was a postdoc there. So I know the whole history of the Manhattan Project. That was a truly focused project and we have many needs today, but I think that should be an example for us of how to focus on really big impact items.

Dave Vellante

>> Well, speaking of focus, I mean, you've got a background and experience in biology, chemistry, astrophysics. I mean, how do you stay focused on what really ... what domain excites you the most right now? Is it nuclear? Is it other material science-

Tom Beck

>> I would say it's two of them. One is nuclear and one is biology. I have a background in biophysics. We are trying to simulate large scale biological systems also on Frontier, and that's also potentially using quantum device and quantum computers to study very localized quantum effects. Like in an enzyme, you might have metal atom that catalyzes the reactions. We need to understand the electronic properties of those atoms. And you need high level quantum mechanics to do that. But we did a simulation about a year and a half ago to ... basically, it was just a water simulation. How many water molecules can we load onto Frontier if we load up the entire machine on one simulation of liquid water? And what we found is that we could handle about 455 billion atoms. So that's an interesting question, is if you take 455 billion atoms and make a cube, how big is that cube of water? And it turns out it's about 1.6 microns on a side. And that's about the size of a bacterium.

Dave Vellante

>> Wow. Okay. So I had to go to my notes here. This system behind us is Cray Frontier systems, 9,800 nodes and 77 cabinets. And I think a two-compute cabinet is 8,200 pounds, which is

Tom Beck

>> Right. And that's one of those boxes, those-

Dave Vellante

>> Right ...

Tom Beck

>> refrigerators back there. Yeah.

Dave Vellante

>> So that is amazing. And my understanding is you really would love for the folks that are using the system to load it up, right? I mean, it's there for-

Tom Beck

>> And they do. We have very big simulations that some of them take over almost the entire machine. A lot of them take up 20% of the machine. And that's an interesting issue with the quantum work. As we were talking earlier, the space you operate in can be very large. And you generate a lot of this data, but then you have to process it when it comes back, like to get the energy levels for the system that you're studying. So that's a place where this machine can be used also, is to do certain parts of the calculation that are very hard to do on the quantum device.

Dave Vellante

>> With quantum, do you think there's ... thinking about bottlenecks, is there the right amount of emphasis on qubit counts? Is there too much? Should there be more focus? Not in the technical community. I think you guys understand it well, but in the media world, should we be more focused on hardware and software integration, algorithms, or is it the right focus to be lasering in on qubit counts and stability?

Tom Beck

>> I think all of those things matter, but in a cartoon sense, the biggest problem to me is the space you work in, it's called Hilbert space in a quantum computer. You have these qubits that have two states and you take 2 to the nth power, that number gets very big. So the potential space that you're operating in is vast, but the pipes coming in and going out are small. So I would say I/O is really a major issue. So the problems we tend to work on are ones that don't require a massive inflow of data so far.

Dave Vellante

>> Interesting.

Tom Beck

>> I think that's one of them. It's like you have this vast ocean of possibility, but you can only access it through small pipes.

Dave Vellante

>> That's really fascinating because ultimately its data is going to be the engine that fuels whatever computing platform. And you mentioned I/O, I mean we saw two weeks ago, maybe it was three weeks ago now at GTC, an entirely new reference architecture for I/O because that has become the new bottleneck. And it's always a constant balance between compute and networking and I/O and

Tom Beck

>> Bandwidth. I call it a Ferrari in a traffic jam. That's what you have kind of. When I taught this energy course, I put a picture of i5 out in California with bumper-to-bumper cars. That's just not an efficient way to do business.

Dave Vellante

>>

Tom Beck

>> And even on Frontier, there are issues of bottlenecks of transfer data and the bandwidth to get data from memory into the processor.

Dave Vellante

>> I usually ask this question in a 12 to 18 month timeframe, especially now because things are happening so fast, but I'm going to stretch the time horizon to the end of the decade. If we sit down at the end of the decade, what would you most want to have proven about quantum, its role, and its efficacy?

Tom Beck

>> I like to operate in the theoretical space, but then have a down-to-earth side.

Dave Vellante

>>

Tom Beck

>> Also, I mean, more of the chemistry type thinking to solve a problem. And I think practically it's really the error correction problem in quantum computers. I think that has to be beaten down to fewer physical qubits for each logical qubit. I think there's promising evidence on that, but really, how to do that and then how to solve the I/O problem onto the quantum machines. And a really big problem is how are we going to scale up to large scale computing with quantum devices? And I go both ways on that. I think there may be possibilities for having one big coherent state of a million qubits, but that as a physicist with all the noise around us and noise in the background radiation and all of this, it's difficult to see how to do that. So I tend to believe that it will be more of smaller quantum units that are scaled up in a way that's similar to GPU, the GPU story that we saw.

Dave Vellante

>> And of course you probably won't be able to make that decision definitively until you solve the error correction problem, which it looks-

Tom Beck

>> Right....

Dave Vellante

>> like AI holds some problem. It could be-

Tom Beck

>> Absolutely holds the problem.

Dave Vellante

>> It could be one of the most useful applications for AI.

Tom Beck

>> Exactly. Ironically, yes.

Dave Vellante

>> Right. Probably more so than chatting. Well, Tom, thank you so much for-

Tom Beck

>> Well, thank you....

Dave Vellante

>> spending some time in theCUBE with us and helping us celebrate World Quantum Day. Really appreciate it.

Tom Beck

>> I thoroughly enjoyed it. Thank you.

Dave Vellante

>> And thank you for watching. This is Dave Vellante. We'll be right back with more great content right after this short break.