This interview examines integration of quantum processors with high performance computing, HPC, focusing on middleware, workflows and system challenges. Amir Shehata of Oak Ridge National Laboratory explains middleware design, cross-stack coordination and approaches to hybrid quantum-classical applications. Shehata discusses the OpenQSE initiative, handling diverse qubit modalities, timing and latency constraints and the role of the Lustre file system in data workflows and they emphasize the importance of preparing software teams for integration now rather than waiting.
Key takeaways: Shehata identifies robust error correction as the pivotal milestone toward useful quantum advantage. They highlight OpenQSE's effort to define interoperable specifications, the need for middleware that supports varying hardware timing and the likely role of Lustre and classical telemetry for syndrome data and artificial intelligence, AI driven error-correction workflows.
Recorded as part of theCUBE's World Quantum Day coverage and presented by Hewlett Packard Enterprise
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
HPE World Quantum Day. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for HPE World Quantum Day
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for HPE World Quantum Day.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
HPE World Quantum Day. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to HPE World Quantum Day
Please sign in with LinkedIn to continue to HPE World Quantum Day. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Amir Shehata, Oak Ridge National Lab
This interview examines integration of quantum processors with high performance computing, HPC, focusing on middleware, workflows and system challenges. Amir Shehata of Oak Ridge National Laboratory explains middleware design, cross-stack coordination and approaches to hybrid quantum-classical applications. Shehata discusses the OpenQSE initiative, handling diverse qubit modalities, timing and latency constraints and the role of the Lustre file system in data workflows and they emphasize the importance of preparing software teams for integration now rather than waiting.
Key takeaways: Shehata identifies robust error correction as the pivotal milestone toward useful quantum advantage. They highlight OpenQSE's effort to define interoperable specifications, the need for middleware that supports varying hardware timing and the likely role of Lustre and classical telemetry for syndrome data and artificial intelligence, AI driven error-correction workflows.
Recorded as part of theCUBE's World Quantum Day coverage and presented by Hewlett Packard Enterprise
HPC Systems Engineer, Quantum-HPC GroupOak Ridge National Laboratory
Dave Vellante
Co-Founder & Co-CEOSiliconANGLE Media, Inc.
HOST
In this interview from HPE World Quantum Day 2026, Amir Shehata, HPC systems engineer in the Quantum-HPC Group at Oak Ridge National Laboratory, joins theCUBE's Dave Vellante to discuss the software infrastructure challenges at the heart of integrating quantum computing with classical HPC systems. Shehata, whose background spans networking, open source software and kernel development, explains why quantum integration is fundamentally a middleware challenge — requiring deep understanding of both application requirements and hardware constraints across the full...Read more
exploreKeep Exploring
How does your team stay on the cutting edge of quantum computing, how do you decide whether to sample many hardware approaches or focus on a few, and what do you mean by “software for quantum” (hardware/software integration, applications, or both)?add
How do quantum processing units (QPUs) differ from GPUs in availability and usage, and what challenges does that create for efficiently sharing a QPU among multiple applications?add
Will there be hybrid workflows in which classical (including accelerated/HPC) and quantum computing work together within the same application, or will workloads be separated to the platform best suited — and how would such integration work?add
What is the current state and future direction of the Lustre file system (including recent developments like The Lustre Collective), and how might Lustre be used or adapted for quantum computing workloads?add
What are the networking and latency requirements when connecting a quantum system to classical error-correction hardware and to an HPC like Frontier?add
What is the OpenQSE initiative (Open Quantum HPC Software Ecosystem) and why was it started?add
>> Welcome back to theCUBE's celebration of World Quantum Day, made possible by Hewlett Packard Enterprise. My name is Dave Vellante. I'm here with Amir Shehata, who's the HPC Systems Engineer in Quantum HPC Group here at Oak Ridge National Labs. Amir, thanks for spending some time with us. Appreciate it.
Amir Shehata
>> Oh, thank you for having me.
Dave Vellante
>> So that's a mouthful in that title. You got HPC in there, you got quantum in there, you got engineering in there. What do you do? How do you spend your time?
Amir Shehata
>> So currently I primarily work on HPC-QC integration. So it's about discovering what the challenges are with integrating such different technologies together. So my day is spent exploring, talking to different quantum vendors. I started an initiative called the OpenQSE or Open Quantum HBC Software Ecosystem, which touches or includes many vendors and universities and other labs that are working on the same problem because we're not the only ones working on it, there's so many other people working on it. So basically it's about thinking about the problem, thinking about the design of how to solve this problem, talking to other people who are working on the same problem, trying to figure out how we can coordinate and work together.
Dave Vellante
>> So we were just talking to your colleague, Tom, about how you all here are on the cutting edge. And I'm interested in how you stay in the cutting edge of the quantum. There's so many different approaches to quantum. You got on one end of the spectrum annealing and the other end of the spectrum, you have other super high performance technologies. Do you sort of sample them all or do you focus on a few? How do you decide where to spend your time?
Amir Shehata
>> Well, of course it's difficult to touch on all of them at the same time, but what we're trying to do right now is we have a system onsite. We have two systems on site, actually, and we're targeting, looking at those systems. So trying to analyze them, trying to dissect them in a software type of way to understand how we can more tightly integrate with them, how we can put them into our workflows.
Dave Vellante
>> So when you say a software type of way... So first of all, for the audience, you've got a background in networking, open source software, kernel development, Lustre, and probably much more. So I probably missed some things, so fill in the blanks there. But when you talk about the software for quantum, is it hardware, software integration? Is it applications? What do you mean?
Amir Shehata
>> It's all of the above. So you have to start at the top of the stack and go all the way down to the bottom of the stack. So you're starting at the application level. I'm not an application guy, but there's a bunch of people that are working on applications and algorithms in the lab. So we're working with them, trying to understand their requirements from the application perspective. And then from these requirements, we're trying to figure out how to design our middleware in order for us to integrate with the bottom layer, which is the hardware. And of course, when you deal with different types of hardware, you have to understand the different requirements those hardwares impose on you. If you're working with super conducting, the qubits have shorter life cycles and they degrade fast, so you have tighter timing that you have to deal with. And if you're working with neutral atoms or another type of slower modality, then there's different timing constraints there. So your software stack that you're developing has to handle all these different types of requirements that are being thrown at you from the hardware side.
Dave Vellante
>> So would you say your area of focus is in the middle, in the middleware, or is that right?
Amir Shehata
>> Yeah. But of course, as the middle you have to really ingest the requirements that are coming from the hardware and you have to ingest the requirements that are coming from the applications. So yes, we were in the middle, but you have to really understand the other layers.
Dave Vellante
>> So when you think about what's happening with the software stack, the entire computing stack with AI, compute storage networking, IO all changing. The data stack is changing. The applications, the world thinks that SaaS is no longer necessary, it's probably not true, but nonetheless, there's some disruption going on there. So the entire stack has changed as a result of the move to extreme parallel computing or accelerated computing, as some like to say. Do you see a similar transformation? I guess so, quantum requires its own unique stack, right? So are we going to see a similar transformation? How will it leverage the AI software stack? I wonder if you could give us some insight there.
Amir Shehata
>> Well, it's funny when you talk about new technology, you think everything is new, but it's not. You're actually utilizing many of the things that have been developed. So a key thing that when we do when we look at the new software infrastructure that we have to create is we actually look at previous technology that was integrated in the ecosystem and see how they did it and we'll try to learn from them. For example, when I talk about accelerators, one of the biggest thing is GPUs. GPUs are the accelerators right now, and they sort of shifted the software paradigm to deal with them. Much of the lessons there, we can learn from it. So we're looking at GPUs, we're looking at smart mix, for example, seeing how smart mix were integrated, what kind of life cycle they went through, and we see how we can learn from these in the QPU world, or the quantum world.
Dave Vellante
>> Are extreme parallelism concepts sort of portable or quasi portable, or?
Amir Shehata
>> It's a little bit funny. The difference between GPUs and QPUs is QPUs are very scarce. You want to have one system that can run one circuit at a time on site. So you have different types of challenges. You don't have many of them like GPUs that you can parallelize, but you have one QPU system and you're trying to fire at it a whole bunch of different circuits to execute. Now, when you have one system, you really have to think how can we utilize it most efficiently so that it's not dedicated to say one application that might use it once in a while. So you need to see how you can get multiple applications using the QPU at the same time. And that kind of introduces different challenges you have to consider.
Dave Vellante
>> Do you think we'll see... I guess we'll almost definitely see hybrid systems, but do you see a hybrid workflow where you have classical computing, parallel, or call it accelerated computing, high performance computing and quantum all working together in a similar workflow? Do you see it as more the right workload for the right job, sort of separating those? How do you see that playing out?
Amir Shehata
>> Funny you say that, we're actually working on this right now. So there are different patterns for these type of workflows. You can think about it from an accelerator pattern. You have an accelerator, you have one big application like a chemistry, quantum chemistry application, that you know some bits of it can benefit from running on quantum, and you're on the application, and then you have those quantum kernels and, you've shipped those over to the quantum to execute. But that assumes that your quantum kernels can execute within a time that is reasonable so that your HPC doesn't remain idle forever. But lots of the trends right now are trending, especially when you go towards the fault-tolerant and error-corrected systems, that those circuits can take a really long time. So you might think about workflow engines where you separate your applications into steps, quantum steps and classical steps. You can do your classical pre-processing, you can collect some data from there, then you free your HPC resources, then you move towards quantum resources where you run your circuits on the quantum that can take a day, two days, whatever, and then you get the results from there and then you go onto your post-processing steps.
Dave Vellante
>> So that's sort of a new software layer that will essentially have to be invented to create that balancing, if you will.
Amir Shehata
>> Right.
Dave Vellante
>> When you think about the road to exascale, Frontier behind us, you had to balance the system, obviously you had hardware performance and you had to balance that with networking and IO and the like. And then of course you had to integrate the hardware and the software, then had to be software written on top of that. All of that was sort of relatively new to get to exascale. How do you see the journey to exascale and the journey to quantum? What are the similarities? What are the differences generally, and then specifically interested from a software standpoint?
Amir Shehata
>> There's definitely going to be parallels between the two, but quantum is really, really at the beginning. So it's hard to compare it to such a mature technology. When we integrated GPU and Summit, for example, GPUs were mature technologies. They had their own software stack, they had algorithms. We're still trying to figure out right now what applications will be useful, will have some use from quantum technology. So we're very early in the life cycle right now.
Dave Vellante
>> We're kind of in the 1970s or '80s? How far back do we have to go to-
Amir Shehata
>> 1960s, I would say.
Dave Vellante
>> '60s, okay. So, all right, so yeah, when the first concepts were-
Amir Shehata
>> Exactly, right.
Dave Vellante
>> That's where we are with quantum today.
Amir Shehata
>> Yeah.
Dave Vellante
>> Of course, you have things like cloud computing and AI is advanced so that maybe things can accelerate.
Amir Shehata
>> Exactly. We're not in the 1950s. You have a ton of different technologies that came before us that you can learn from, you can accelerate your progress, right?
Dave Vellante
>> I want to ask you about Lustre. So we were talking off camera, Lustre sometimes gets a bad rap from those that want to compete with Lustre, but Lustre's proven great for highly parallel workloads. You've got to have expertise, but it's open source, a big community around it, it's mature, it's proven. What's the state of Lustre? I know you're excited about some new developments that are going on there, and how does it apply to quantum?
Amir Shehata
>> That's an interesting thought. So Lustre is going through a new stage in its life right now. So there's The Lustre Collective, we chatted this about this before.
Dave Vellante
>> Sure.
Amir Shehata
>> Basically, the founders of Lustre like Andreas Dilger and Peter Jones, they started their own Lustre company called Lustre Collective, and they view it as good for the open source community. Now they can develop features that are provided and they can support multiple different customers, so to speak. And Lustre is very prevalent. Frontier runs Lustre. You can almost like... A lot of the supercomputers run Lustre.
Dave Vellante
>> Virtually every HPC runs Luster, right?
Amir Shehata
>> Yeah.
Dave Vellante
>> I mean, it is the standard, is it not?
Amir Shehata
>> Yeah, and you can think about it, when we do integrate file systems with quantum systems, Lustre will be there somewhere, right? You're going to be reading all this vast amount of data from the qubits to do error corrections, and you might want to learn from those error rates that are coming out of the quantum computer to be able to develop AI models that help you do better error corrections. So where are you going to store this? You'll need a file system. Lustre can be there.
Dave Vellante
>> Do you think... I would imagine a specialized purpose-built file system will emerge for quantum. At the same time, the Collective, for example, is going to evolve Lustre so that when the time's right, it will be able to work on quantum computers. How do you see that playing?
Amir Shehata
>> There's an interest. So when we talk about quantum, really, you're talking about a quantum computer is mostly classical with some quantum element into it. So when we read qubits or when we read syndromes of qubits, we're actually reading them into the classical world and then writing from the classical world into Lustre. So I don't imagine that you're going to be integrating a pure quantum technology into a file system. You're going to have some classical beside it that interfaces with the quantum through control electronics, et cetera, specialized equipment that can handle the quantum side of things, but the interfacing with the file system will be classical.
Dave Vellante
>> That's good news for Lustre then-
Amir Shehata
>> Yeah, exactly....
Dave Vellante
>> since somebody who knows Luster as well as you do. So are we further along than people think in quantum? Not as far along, because we were talking earlier about still early days, but Jensen made the comment a couple years ago now that quantum is decades away, and of course all the quantum stocks dropped, and then he backtracked, and then all the quantum stocks went up. So what's the truth? Are we further along? Are we where we should be? How would you describe that?
Amir Shehata
>> So I'll give the honest answer from a software perspective, right?
Dave Vellante
>> That's why we love when engineers come on.
Amir Shehata
>> Exactly. Where... The companies' roadmaps are out there for people to look at. What we're trying to do is we're preparing the software world to accept those technologies. So when they are ready, you can just jump on and start using them. You don't want to wait until they're ready and then start thinking, "Well, how are we going to integrate them with our systems?" Because that's going to be too late. There's going to be another 10 years down the road or whatever. Don't take the years for-
Dave Vellante
>> Yeah, yeah. Not literally, but.
Amir Shehata
>> But like a long while down the road. So you need to start early so that when those technologies mature, now you're ready to actually use them.
Dave Vellante
>> What kind of milestones should observers look for as signs of substantive progress in quantum that should give us indicators that it's here? I mean, I know there are some things here and now that we could say, but going forward, what are the things that we should be looking at?
Amir Shehata
>> I think the biggest thing is error correction, because right now quantum is noisy and they can't live for a very long time. So you can't do deep circuits that actually give you meaningful results without the system descending into noise and then basically you don't get any useful results out of it. So error correction is key, but error correction is not just suddenly you're now going to flip a switch and suddenly the system is going to be error-corrected. There's going to be a curve as you move from the NISC era to the full tolerant error. So as technology matures and people come up with better error correcting codes, that's when quantum is going to really become useful.
Dave Vellante
>> You know a lot about networking, done quite a bit in your career on networking. When you think about, I remember Jayshree Ullal educating me back in 2011 about north-south versus east-west traffic, that east-west was coming and that's basically the premise of her company. And now you're seeing this north, south, east, west explosion, so sort of a new networking paradigm where everything is as synchronous as possible. How will networking be different in quantum?
Amir Shehata
>> That's an interesting question. So when you talk about networking and latencies in data transfer, you have to really divide your system into kind of two pieces. There's the error correction side of things, that's going to require classical that's really tightly coupled with the quantum, and that's going to have to be very low latency, very fast, dedicated connection to grab all the syndromes out of the quantum system and error correct, and do the error correction in the classical and then feedback into the quantum to adjust. That's going to be very fast. Now you're going to integrate say a system like Frontier with that. That might not necessarily be very tightly integrated. You don't need the very fat connections between that system and the quantum system. However, what I would say is whatever software you're going to develop to couple an HPC machine like Frontier, for example, with a quantum system, whatever software infrastructure you develop has to be well-designed and it has to be, support even low latency. So whatever you're going to develop there is going to run for the low latency environment as well. So I don't see, just because we don't need high bandwidth between Frontier and the quantum system doesn't mean whatever software we're going to create is not going to operate in low-latency environments.
Dave Vellante
>> How about data center design? The first data center I ever walked into was very cold, it was a mainframe. That mainframe was, I'm pretty sure of liquid-cooled, had a raised floor. You saw that evolve, hot aisles and cold aisles, and now you're seeing the evolution of data centers now with these AI factories being built. How will quantum data centers be architected? What's going to be different about them?
Amir Shehata
>> And that question really depends on the type of modality you're thinking about. So there are some modalities that can operate in room temperature. We have one of them, like Quantum Brilliance is one of those machines here, and that can just live in the same data center as Frontier, for example. But when you talk about other super conducting systems, they require big fridges that cool down to like millikev and kelvin temperatures and they're very sensitive machines. So if you put them beside a supercomputer like Frontier, the energy and the noise coming and the magnetic field or whatever from a system like that can interfere with the super conducting. So there might need to be separation in geography between a quantum system and an HPC system.
Dave Vellante
>> Where do you land on data centers and computing in space? We saw Artemis launch yesterday, you hear Elon talk about, "Oh, we're going to do one in space, is this an unlimited power because of the sun." Others say, "Well, latency is kind of an issue." How do you guys think about that here?
Amir Shehata
>> Oh, I don't know. I'm thinking about it now. Send me up there, I'll solve that problem for you.
Dave Vellante
>> Truck rolls, rocket roll.
Amir Shehata
>> That's actually interesting because when we talked a little bit earlier about, you don't need really high bandwidth between a Frontier supercomputer and a quantum system. So if you have the quantum system somewhere up there, you might still be able to hook it up with a supercomputer that's down here and the data that's going back and forth might not be very latency sensitive, but you'll still need classical up there that maintains the error correction for the quantum system.
Dave Vellante
>> Can you cool that quantum system in space? Is it cold enough?
Amir Shehata
>> If you build it on the far side of the moon where it's dark or in some, I don't know, crater, that might be interesting.
Dave Vellante
>> Do you actually think about these things, or is this-
Amir Shehata
>> I'm a science fiction writer, so I do think about this sometimes.
Dave Vellante
>> What would you say is the hardest part? I mean, you're a system thinker. What do you think the hardest part of integrating quantum and HPC is from a systems perspective?
Amir Shehata
>> I think it's the workflow model. You're integrating new technology into existing systems. So you want to be able to make it as seamless as possible for the user. You don't want someone to learn everything from scratch. So one of the challenges is to integrate this new technology into a system that is into an existing system and make it as seamless as possible. So that's one of the challenges we are facing because there's new requirements coming in and we need to adopt those requirements into environments like Frontier's environment, for example.
Dave Vellante
>> So I mean, error rates obviously is a problem that has to be solved. Latency we talked about a little bit. It sounds like it's more optimizing for orchestration across those systems.
Amir Shehata
>> Exactly. Yeah, you're right.
Dave Vellante
>> Interesting. What would you say is one of the biggest lessons you've learned so far from your early quantum experience? And what are you hoping to learn in the next 12 to 24 months?
Amir Shehata
>> So it's funny, I think that the biggest lesson is that you can do this on your own. You really need to work with other people that are working the same problem in an open source community that allows free exchange of ideas because if everybody's working behind walls, you'll never learn from each other. So I think this is one of the most important lessons for me as I interact with a whole bunch of different companies and labs and universities that are working on the same problem. So you really need to have those discussions. And that's why we started that initiative, the OpenQSE initiative, which basically addresses that need. It brings a whole bunch of people together that are working on the same technology and we're working on trying to figure out how to solve this together.
Dave Vellante
>> Sorry, that was started out of Oak Ridge.
Amir Shehata
>> Out of Oak Ridge, yeah.
Dave Vellante
>> Tell us more about that.
Amir Shehata
>> So OpenQSE, it stands for Open Quantum HPC Software Ecosystem, and it's an initiative that basically brings in together labs, vendors, universities to discuss the integration problem. The goal of this initiative is to try and figure out how we can standardize that software environment, such that we can make different software stacks interoperable so you don't get vendor lock-in by using one specific software stack. So we're trying to come up with specifications for the different software layers so that we can allow different people to go out and develop their own modules, and then those components can be interoperable at the end.
Dave Vellante
>> I mean, there's a huge advantage for developers. A developer, he or she wants to focus on something that has durability, they want to see their work have lasting effects. So what's your message to those developers? How do you attract them? It's early days, so it's not a lot of software being developed today for quantum, but that's something you have to think about in the future, is it not?
Amir Shehata
>> Yeah. Like workforce, you're talking about attracting the workforce? Is that the question?
Dave Vellante
>> Workforce or just even utility, so where people can actually program these things and develop software that's useful on top of these amazing systems.
Amir Shehata
>> Right. Quantum brings in a completely different programming environment. It's not like if you know how to program and see or Python, it's not directly translatable into programming in a quantum circuit. So it requires a mental shift, but this is the best time to learn. There's so many tutorials out there. You can go on YouTube and open a site and you start learning about this. Lots of the vendors have documentation online to learn from, how to use their systems. So there's a lot of material out there to learn from right now.
Dave Vellante
>> Well, and we'll be programming, and me in English, you speak multiple languages, so in some natural language, right? I mean, that's going to be potentially with AI and all the advancements in coding and coding assistants, perhaps it will be coding quantum computers by speaking to them. Do you see that day as a possibility or are we going to be in front of a C prompt?
Amir Shehata
>> Well, that's an interesting idea. I'm not sure how close we are to that world, but at Oak Ridge, we are working on that exact idea using AI to inform how to develop four quantum systems. But in order to do that, you need to collect data. You need to be able to curate all this data and represent it some in a useful way to the user.
Dave Vellante
>> So we were talking about learnings a little bit. I wanted to ask you about advice. You started to touch on that a little bit with the open collaboration. Where should people focus first? Is it on skills? Is it on infrastructure? Is it use cases? Where would you sort of advise people who want to get involved starting?
Amir Shehata
>> So I think this problem is a systems problem, like integrating quantum into HPC centers is a systems problem. And you have lots of many different people that might have different expertise in different areas. So speaking personally, I think you should capitalize on your own expertise and see how you can utilize your own expertise to solve that problem. So I come from a networking systems background, and that's why I sort of delved into the HPC integration because that's where my primary experience is. Some other people might be learning quantum so they can target the applications and the algorithm development. Some other people might be networking, so they can look at how to solve the tight coupling that we need for error correction. So I think it depends on your expertise. You can capitalize on what you already have and see how we can tackle this new problem with what you have.
Dave Vellante
>> Amir, if I had to ask you what is the one big breakthrough that you'd be most looking forward to that would change everything, would it be the error correction capability?
Amir Shehata
>> I think that's the biggest thing right now. Once you have error-corrected quantum computers, that's when you start moving from toy problems into significant scientific applications that can utilize it.
Dave Vellante
>> So once you get to that point, obviously compute is not going to be the problem. You've solved that with that error correction breakthrough. Does it then become an integration into existing systems, that hybrid sort of workflow that we were talking about? Is that the sort of next real challenge?
Amir Shehata
>> I think it's both, right? You can't really do it in sequence. You can't focus on error correction only. And once that's solved and you focus on integration and once that's solved, you start looking at applications. You have to look at everything at the same time. And that's what we're doing here at Oak Ridge. We're looking at the error correction, we're looking at the middleware integration, and we're looking at the applications as well. And all those have to move in parallel in order to capitalize on this technology once it becomes viable.
Dave Vellante
>> How much of a bottleneck will data movement be? And I guess specifically IO?
Amir Shehata
>> Again, I think you really need to look... When we talk about IO, we need to kind of focus that question on what type of IO. If it's error correction, then you really need very fast interconnect and be able to move this because you're going-
Dave Vellante
>> Real time, essentially.
Amir Shehata
>> Exactly. You need to operate in microseconds, right? If you talk about application, submitting jobs to the quantum computer to run, maybe it's not microseconds, maybe it's seconds, maybe it's milliseconds, but we already have the technology to handle this time scales.
Dave Vellante
>> I asked Tom this question. I'll ask you, I said, usually I ask in 12 to 24 month timeframes, but with quantum, I'll stretch it out a little bit for World Quantum Day. Let's say it's end of the decade. What would make you most happy as to where we would be by the end of the decade? What would Amir say, "All right, we accomplished what we wanted to accomplish." What does that look like? What does success look like end of decade?
Amir Shehata
>> So I'll answer that question in two parts, right? I think as a community, having error-corrected systems is a big milestones. For me as personally, I think having a common software stack that people... Sorry, I want to make this clear. Not a common software stack, a common set of specifications that allow different stacks to be interoperable. I think having that kind of environment will really propel innovation and will allow those systems to integrate more smoothly and will help centers like us that are trying to evaluate different types of technology, we don't have to reinvent the software stack every time we try a new technology. Now we have a common base that we can use.
Dave Vellante
>> When I was prepping to come here, I sort of developed this mental model and I said, "Okay, Quantum, it's kind of this new capability. It's to think of it as a new tool." When I think about Frontier and HPC, I see, oh, that's kind of that and is a foundation, and AI is this sort of glue between the two. AB can help with error correction, and then the value comes from this sort of hybrid workflow that we've been talking about. You've got general purpose systems, you've got HPC systems, AI systems and quantum systems all working together. That's where the world, scientific world, business world is going to see real value. Does that mental model sort of hold water with you?
Amir Shehata
>> I think so. Those systems are not going to be working on their own. They're going to be all integrated together. And like you mentioned, you're going to have some AI models that assist application developers in developing your applications. We see that already right now. You have codex, you have Claude, you have different AI models that assist software developers to write their applications. You're going to have something similar with quantum, and then you'll need the integrated systems in order for those applications to actually make use of the quantum. And quantum, we said it's not going to exist without HPC, so you need this tight integration with HPC to allow quantum to really shine.
Dave Vellante
>> All right, well, Amir Shehata, thanks so much for coming on theCUBE. Really appreciate your time.
Amir Shehata
>> All right, thank you.
Dave Vellante
>> All right.
Amir Shehata
>> Take care.
Dave Vellante
>> All right, and thank you for watching. This is theCUBE's Celebration of World Quantum Day, brought to you by HPE. Thanks for watching, be right back with more from theCUBE.
