Let's dive deep into the state of quantum computing with one of the leaders in the space. Today we're joined by Dr. Alan Baratz, the CEO of D-Wave, in a conversation brought to you by D-Wave. I've been waiting a long time for us to do a deep dive into quantum, into what the state of the technology is, where it's going, how it's already having an impact today. And I'm very excited to be able to do this show today. Finally, for the first time on our channel, we're going deep into quantum. And Dr. Baratz, you're the perfect guest to be here with us to do it. So, welcome to the show. Great to see you.
Thanks, Alex. It's a pleasure to be here and thank you for taking the time to have this conversation.
My pleasure. By the way, folks, we are going to be doing a version of this conversation at D-Wave Qubits, the Qubits event that's going to be happening in Florida at the end of this month on January 27th and 28th. So, we'll share some more details about that. But first, let's get right into the subject matter here and talk about quantum. So, Dr. Baratz, I'm going to just start, you know, really broad and give us a jumping-off point for those who have heard about quantum but aren't deeply familiar with where the technology is and where it's going. So, let's start with this. What is quantum computing? Is it theoretical? Is it something that, you know, might be coming down the line over time? What is the state of the technology and where is it going?
Sure. So I'm going to give you an answer that probably you won't expect, but that is that quantum computing is energy-efficient computing for solving very hard computational problems. And I say energy-efficient computing because quantum computers consume very little electricity and yet they are very powerful. They use quantum mechanics, for example, superposition, to be able to perform computations orders of magnitude faster than they can be performed classically. And that opens up exciting new opportunities across many different industries and across society.
And so then what is quantum mechanics? I mean, we know that with like typical computing, you know, you have your CPUs and switches and things like that, but what is quantum mechanics and talk a little bit about how quantum computing actually works.
Yeah, so I mentioned superposition, that's really the best place to start because that's in some sense the most straightforward way to explain how quantum computing works and why it's so powerful. So classical computers, today's computers, whether they be CPUs or GPUs, store information in bits and a bit can be either a zero or a one. Quantum computers store information in qubits and qubits can be zero, one, or both at the same time, or any combination of zero and one. And when the qubits are in that state where they're a combination of zero and one, we call that superposition. Now, why is that so important and so valuable? Well, if you think about the way classical computers work, when they're solving a problem, essentially they're able to look at one potential solution to the problem at a time, intelligently iterating their way through to find the best possible solution, but one solution after the next because the bits can only be zero or one. So at any given point in time, the collection of the bits in the computer show one possible solution. With qubits, because we can be in superposition, because we can have a combination of zero and one for each unit of information, we're effectively able to operate on many different possible solutions at the same time. And that's what allows us to intelligently move to find the best possible solution very quickly, orders of magnitude faster than you can do it with classical computers.
And then talk a little bit about just the mechanics of this. I mean, we talk about quantum, you know, is it like, you know, deeper into the particles of matter that we don't fully understand yet? And how is it possible that something can be, you know, both a zero and a one at the same time?
Yeah. So that's the nature of quantum mechanics. You know, it's been called spooky. It's very difficult for most people to understand, frankly myself included. But it's proven to be the way the universe operates on the micro scale. So there are many, many interesting properties of quantum mechanics that we are bringing into the computing environment. I talked about superposition. Another one is entanglement. Very interesting. If I perform some action on one qubit, that can impact another qubit that might be very far away. That's entanglement. And we use that as a part of how quantum computers operate. Now, as far as the physical realization of quantum computers, there are a number of different underlying technological approaches that are being pursued today in the development of quantum computers. One is superconducting. The other is trapped ion, another is neutral atom. Another is photonics. So there are many different technological approaches that are being used to pursue the development of quantum computers. Each of these can be built in such a way that we can introduce the quantum mechanical characteristics like superposition, like entanglement, like tunneling and use those characteristics to solve hard computational problems.
Interestingly, there's a big debate going on right now around which is the best technological approach for leveraging quantum properties and building highly performant quantum computers. Is it superconducting? Is it trapped ion? Is it neutral atom? And I have a very specific view on that which hopefully I'll share with you a little bit later.
Yeah, I definitely want to get into your perspective on what the best approach is and where we're seeing progress. And as far as I'm aware, there's already some breakthroughs that are happening already and usefulness of this technology. But you know, before I get to that, and I also want to talk about D-Wave, but the questions are just, you know, spurting out here. So let me ask you this. With machine learning, we had an idea of why machine learning was able to break out, right? There was this idea of, well, it was a conceptual thing, but then eventually there was enough data and enough computing power that these theoretical approaches like deep learning started to show results and then we had this machine learning boom. And I'm just wondering because we're hearing so much about quantum now and its potential, is there something that's happening in the world of quantum computing that is a similar reason for why this technology could break out? Like what is the parallel for the fact that there's like been more data and more compute in machine learning? What is the parallel with quantum computing?
Yeah. So the thing that everybody is focused on with respect to quantum computing is what's been called quantum supremacy. The idea is, can a quantum computer solve a problem that cannot be solved classically, period. Now, there have been a number of attempts to demonstrate quantum supremacy. Some of them have held up, some of them have not. But in every case, the computation has been a contrived problem of no practical real-world value. But there is exactly one demonstration of true quantum supremacy on a useful real-world problem that was done by D-Wave, by our company. It was published in the peer-review journal Science about a year ago. What we did was we demonstrated that our quantum computers are able to compute properties of materials in minutes that would take nearly a million years on the fastest supercomputers in the world. So that's true quantum supremacy. We are performing a computation on our quantum computer that cannot be solved classically. But importantly, it's on a useful real-world problem in the area of material simulation. And this is the first and frankly still only demonstration of supremacy on a useful real-world problem. So we are at the point actually today where at least at D-Wave we have quantum computers that have kind of made that transition to be able to demonstrate useful quantum supremacy and that has also allowed us to achieve the point where we actually have customers that are using our quantum computers as a part of their business operations today, including for example one of the largest international airlines, one of the largest chemical companies, one of the largest mobile cellular carriers, one of the largest payment companies. So our quantum computers have not only achieved useful quantum supremacy, but they are commercial today. Customers are using us today as part of their business operations.
Okay. And I think this is going to come as a surprise to many of our viewers because the general perception is, and we hear it all the time, like I was recently at Google and they had one of their research events talking about how their quantum computer was able to do like one or complete an equation or complete an algorithm 10,000 times faster than a typical computer. But then this next sentence is always what is like, well, it's well in the future and we're not there yet. But you're talking about real-world applications that are in existence today. And I think this comes down to a little bit of the difference in approaches that you hinted at earlier on. There's an approach called annealing and an approach called gate. And we're seeing some results with the annealing method already. So can you unpack that a little bit and explain the two different approaches and why we're seeing some results with one and not the other yet?
Sure. So just like a few minutes ago I talked about a variety of different underlying technologies that can be used to develop quantum computers, there are a couple of different architectures that can be used to develop quantum computers. The two main architectural approaches are called annealing and gate model. The two approaches that you mentioned just a minute ago. Now these are very different architectural approaches to quantum computing. Annealing quantum computers, frankly, are an easier technology to work with. They're much less sensitive to errors and as a result we are able to solve important problems without the need for error correction, which is quite different from gate model systems which are very sensitive to errors and as a result you really require error correction before you can do anything useful with those computers. Now, we can talk in a minute about where we are relative to error correction on gate model quantum computers, but for annealing quantum computers, they're commercial today. Our current generation annealing quantum computer, it's called Advantage2, is a 4,500-qubit quantum computer. It's the largest and most powerful computer in the world. It does not require nor does it have error correction, but yet as I said, it's the only quantum computer on which we've been able to demonstrate supremacy on a useful problem and it is the only quantum computer that is being used by customers today commercially as a part of their business operations. Now there are many, many important applications that annealing quantum computers can solve. They're very good at solving a class of problems known as business optimization problems. So workforce scheduling is a business optimization problem. Manufacturing plant floor optimization, optimization problem. Logistics management, optimization problem. Many, in fact, frankly most of the important hard problems that businesses need to solve are optimization problems and annealing quantum computers are very good at solving those problems. While at the same time we now know that gate model quantum computers are not very good at solving those problems. So for optimization, it's annealing. Gate model, on the other hand, is good at solving a different class of problems. Gate model quantum computers are very good at things like quantum chemistry for developing new drugs or fluid dynamics. This is a different class of problems, a very important class of problems, very valuable class of problems. Gate model is very good at solving those problems and annealing not so good. So we have a bifurcation in the application environment today where there are problems that will always require annealing, there are problems that will always require gate. Both are very important. And D-Wave, a very important point that I'm now about to make, is the only company in the world with a dual platform strategy and plan. What I mean by that is we're the only company in the world developing both annealing and gate model quantum computers to address the full market for quantum. Our annealing systems commercial today in the market. We've got a strong roadmap to continue enhancing those systems into the future and we're now also focused on developing gate model systems. In fact, we just recently announced an acquisition in this space that we believe has catapulted us into frankly a leadership position in R&D for gate model quantum computers alongside annealing.
All right, so let's go a little bit deeper into both. I want to start with annealing. You mentioned it's really good for business process optimization and you've talked in our conversation so far about how quantum computing is really good for energy-efficient computations that just blow the traditional computers out of the water when they work well. When it comes to business process optimization, I'm curious if you can share a little bit about what quantum optimization and optimization done within an annealing computer can enable a customer to do or can enable a company to do in a way that maybe traditional methods like machine learning can't. Because with machine learning, like with the predictive analytics and the optimization side of AI, that's been something that we've seen in market for a long time. It's done a pretty good job. So when it comes to the advantage that quantum might have over some of those traditional processes, what do you see and what do your customers see?
So let me start by giving you a concrete example or two of customer applications that we have in the market today. And then maybe I'll come back and share some thoughts on how AI and quantum relate to one another and potentially work together synergistically. So, let's take BASF, one of the world's largest chemical companies. They need to fill orders for customers for a variety of different types of chemicals that frankly are created by a manufacturing process in their facilities starting with a variety of raw materials that come together to form those chemicals. Okay. The challenge is to optimize how the plant floor operates to fill those orders as quickly and efficiently as possible. Leveraging our quantum computers, they have been able to reduce the time required to do the production scheduling from what was 10 hours down to seconds. Okay, so a very significant reduction in the time to do the scheduling. Well, if the computer is running for 10 seconds versus 10 hours, that's consuming a lot less electricity. So there you have an example in the real world of the use of quantum computing to deliver business value and reduce electricity consumption. Another example is a large mobile carrier in Japan, NTT Docomo, has used our quantum computers to optimize cell tower resources. Basically, how the cell towers interact with the mobile phones to essentially determine which towers should be in control as the phones are moving about. Leveraging our quantum computers, they're actually able to support up to 10% more phones per tower by optimizing how the control signals flow, which means a significant reduction in infrastructure cost for them. So, another example of real business value through computational optimization leveraging quantum computers.
All right, let's now talk about AI and quantum computing and how they relate to one another. I like to think about it as falling into two categories. One, AI and quantum computing are very complementary, each good at solving different portions of a problem. So for example, you might use AI to predict demand for some future products and then use quantum computing to optimize the supply chain to meet that demand. Here we have AI and quantum computing working side by side, each focused on the portion of the problem that it is best at addressing. The second area where there's synergy between AI and quantum is the use of quantum to improve how we do AI model training and inference and specifically being able to do it with less electricity consumption. I mean, if you ask anybody what's the biggest issue with AI today, they'll say power consumption. I mean, that's why we've got companies out there talking about buying nuclear power plants. Okay. Well, what if you could train models better, faster, and with a lot less energy consumption? That would be very transformative. Well, we at D-Wave have the first commercial example of that. It's work that we did with a company called Shionogi. They developed a large language model for generating molecular structures and they're looking for new molecular structures that are well suited to human drugs and they've been working with purely classical GPU-based training up until recently. They've now introduced our quantum computer into the model training process and they are finding that they are getting better models trained much faster. And by better, I mean many more of the molecular structures are well suited to human drugs. So here we have quantum computing benefiting AI by making training and inference even more efficient and consuming less power.
