 Hello, I'm Dario Gil, and I'm speaking with you from the Yorktown Heights Research Laboratory. It serves as the headquarters of IBM Research, and today is a great pleasure for me to share with you a perspective of the quantum computing era that is emerging, and the implications that it will have for science, for discovery, and for business at large. It is clear that the world is being challenged with ever more complex problems to solve. The coronavirus and COVID-19 is a powerful example of the challenges we confront in humanity as we try to understand some of the deep functioning of nature and the implications it has for all of us in our societies. Here in the context of dealing with a structure that is no more than 100 nanometers inside, we're struggling to figure out ways how to deactivate the functioning of the virus. Now we know that computation can help us. If we look at the inner structure of the virus, we see these spiking proteins, and these proteins can be modeled and understood, and therefore we could use them to then explore different kinds of compounds that could deactivate how they function. Here, supercomputers can come to help us. We can use these most powerful computers to model these protein structure, and we can use another powerful force, which is the powerful force of joining the resources and the capability across institutions. In moments of crisis, it is particularly important that we all come together, and that's why we were very proud to coalesce and create in partnership with the federal government and other institutions in the tech sector, as well as Academia, the COVID-19 High Performance Computing Consortium, where 30-plus members have aggregated over 400 petaflops of computing power and over 100,000 nodes to go and pursue a broad portfolio of projects that can help us understand the pandemic, understand the evolution of the virus, and accelerate the pace at which we can develop antivirals and ultimately vaccines. What this tells us is that when confronting with exponential problems, we have enormous computational demands. In fact, exponential problems require exponential computation, and today we're fighting the pandemic within the context of the computational paradigm we have, the world of bits, but there's another computational paradigm that is emerging and is the world of qubits. So moving forward, qubits and qubits will be foundational computational paradigms with which we can tackle complex problems in the world. So let's try to understand the difference between them and where the power of qubits rise. If we look at the left-hand side, it's a way to describe, and we're going to use these diagrams throughout the presentation to unpack the power. We're seeing a circle and two dots. When the circle, when the dot is in the North Pole, that's a zero, when it's in the South Pole, it's a one. That's a classical bit to state. Now, a qubit is the unit of information in quantum world, and here we get to have, let's depict it as a sphere, a zero and a one, and we can have them in this superposition state, but we have a special trick that is available to us. Not only we can have them positioned in this sphere, but imagine that each one of this, now it's like a little moon that is in this sphere, and we can rotate the moon. So here you're seeing that the bottom sphere has been rotated slightly, so that's why you see a little bit of blue and pink, and we could rotate it all the way. So here's a 180 degree shift. So we have this notion of phase. Quantum affords us three superpowers that are exhibited in the qubits. They are the powers of superposition, the power of interference, and the power of entanglement. So let's unpack those. The principle of superposition is actually quite straightforward to understand. So let's start on the left-hand side. We have our North Pole zero or South Pole one, and if we add those two states, now we have a superposition qubit that has both a zero and a one simultaneously, North and South Pole. Now the right diagram is the same, but now we're gonna do a little trick. We're gonna apply a gate, a quantum gate, and what we did here is remember, we get to change now, we get to rotate these little moons, right? So we've rotated it 180 degrees, so now we have a one. So now when we do the addition of zero plus one, we end up with a superposition state, but notice that it's different than the one on the left. We have still a zero and a one, but now with a rotated state for the one. Now that is gonna be very powerful to then leverage a second really interesting thing, which is the principle of interference. And this is not an operation that we get to perform in classical bit world. So here in this example, we have a zero and a one on the left plus this new state that we created, which is also a zero and a one, but with this rotated moon. When we do this now, we get to interfere those two states and notice something really interesting happens. We've canceled the one state and we've ended up with that zero state. We've done that because we've interfered the states. So the next opportunity that we have right now is how many states can we create? And the beautiful thing in quantum is we get to create an exponential number of states. So let's look at a system that has five qubits. In this case, we're depicting these qubits as independent qubits, right? You're seeing them on the left, zeroes and ones, and they're getting multiplied. We have five of them. So the number of states that we can create now in our quantum computer, depicted on the sphere on the right, it's two to the power of the number of qubits we have. In this case, we have five. So two to the power of five is 32 states. So we're seeing 32 of those little moons. Notice that in this case, we did not incorporate any phase. They're all pink. The powerful aspect of using this property called entanglement in the quantum world is that we can create now states like the one depicted on the right, where we have these combinations of pink and blue. Where that end state cannot be described as the independent product of these five qubits, in this case, since we have 32 states. What this tells us is actually something very profound. And that is that you could have an end state in your quantum system that you're using to process information that cannot be described as the independent components of the qubits that make the whole state. There's a lot in there. It's a very profound statement. And it is the basis of the power and the complexity of quantum machines. So let's bring that all together to give you an intuition for how an algorithm works in a quantum computer. What we do is that at the beginning, let's say in this case, we have a five qubit quantum computer is that we prepare the computer to be in a superposition of states. Two to the power of five, 32, you have those 32 dots that you see on the left in pink. Then we have to encode the problem. We have to inject data into the quantum machine. The way that happens is that that encoding gets done through entanglement. So notice that what we've done now is that some of those states are now changed in phase. You're seeing those depicted as blue and some are pink. Now in our machines, not only we have the superposition, we have exploited the property of entanglement to do the encoding of information. And the principle of interference comes from the fact that we can now take these states and combine them and interfere them with one another in such a way that we get to cancel things out and maximize the right answer. So many things fall away and the right answer gets maximized. We perform a measurement and we get our result. Notice how very different that is from classical bit-based operations. And there is a really important relationship between this entangled property and the amount of information that we can process. And this relationship is exponential. What this table shows is a number of classical bits, zeros and ones that are required to represent that complex entangled state that I just described. And here's an amazing number. By the time you have a hundred perfect qubits that are entangled with one another, if you needed to describe them using zeros and ones, you would need to devote every atom of planet Earth to store those zeros and ones. Clearly that's not possible. In fact, by the time you have 280 qubits, you would need every atom of the known universe. So that, it tells you that this exponential relationship between the special property of entanglement and the representation of information. Why does this matter? Well, it matters because it turns out that despite how powerful our classical computers are, the reality is that there's a vast set of problems in the world that they cannot tackle. From an information theory perspective, we would say that really classical computers can tackle easy problems, things that don't have an exponential number of variables in them. Remember, in the case of the virus as an example, that modeling nature is an example of an exponential problem. But so are other problems like factoring and problems in optimization. In fact, there's a vast number of problems in the world of mathematics that have that character and those problems are deeply important for business the world over. For optimization and in the future and machine learning and developing new molecules and chemistry and the life sciences and beyond. Now, the reality of it is that you say, well, that's very interesting and the theoretical underpinning but can you actually build these machines and access these machines? And we also recognize that for most of history to access computing power, you had to build it and maintain it yourself. But now we're in a situation where even these very special machines, quantum computers can be accessed by anyone in the world. In fact, the system looks as follows. You can sit in front of your terminal, write your program, you can send your zeros and ones and when they get here to the laboratory where we are as an example, we convert those zeros and ones to microwave pulses. They operate about five gigahertz and we send them down a cryostat that operates about 15 millikelvin temperature and we get to use those microwave pulses to manipulate these qubits, to perform those superposition and entanglement and interference operations and then return the result to the user who is using a regular classical computer. This is what the inside of these beautiful machines look like. I say beautiful because they're really gorgeous pieces of engineering and science and you're seeing the inside of this cryostat, that golden chandelier and those wires are the means by which we send those microwave pulses down. At the very bottom of that golden chandelier is where we have the quantum processor itself where we perform those operations. So if we go and look inside, this is what a quantum processor looks like. In our case, in IBM, we use superconducting technology. Specifically, we use a qubit device called the transmon qubit. And what you're seeing is at the core of it, if you look at the bottom right-hand side, we have a device which is called a Joseph of Junction that it's about 100 nanometers by 100 nanometers, roughly the size of the coronavirus, in fact. So 100 nanometers by 100 nanometers and essentially allows us to create an artificial atom with a ground state and an excited state that is the basis then with which we can couple it to other qubits and you're seeing on the right-hand side those wiggly lines that we've labeled microwave resonators and that's what allows us to couple the qubits with one another and then perform these entanglement and interference operations that give quantum computing its power. We build these quantum chips here in the laboratory from which I'm talking to you today in Yorktown Heights. Now, with those units of information, the very important thing that is being created is a rich and vibrant world of quantum circuits. This is what's really next in the world of quantum. So let's explore them a little bit. A quantum circuit is gonna be the unit of value for your business. The qubit is a unit of information, but a circuit is the unit of computation. So let's try to understand what is a quantum circuit and let's look at its anatomy. This is what a quantum circuit looks like. So let's decompose this into pieces. The qubits are depicted by these horizontal lines. In this case, we're depicting a circuit that has four qubits that you see here shown. Now the gates are how we control the qubits and you read the gates from left to right. So it's the sequence of placing these gates that is the basis of creating a circuit. To construct these circuits, we have many quantum gates available to control these circuits. So this is analogous to the way we do controls in the world of bits. So in the world of bits, we're all accustomed to that behind the scenes, we have these ands and ors and knots, et cetera. So similarly in the world of qubits, we have a different set of operations. Remember, some of those operations are the very idea of doing the rotation to one of our little moons that I was depicting before, right? Introducing a phase change. And I showed before a Z operation that performed that rotation that allows us to do that interference operation. So let's do a comparison between a classical circuit and a quantum circuit. Here in a classical circuit, what you witness here is that you don't get to see blue moons, right? There's no rotations. Everything here is pink. And that you got these single states that in this case, we're seeing moving the sphere. But notice what a contrast is gonna look like when we actually implement a quantum circuit. So let's look at what a quantum circuit looks like. So in this case, we're implementing Grover's algorithm, which is a basis for performing search operations. First of all, notice the complexity has increased greatly, right? We have the superposition states. Now you're seeing that we get to perform these operations where we do rotations. So here you're seeing blue. We're performing operations of interference. You're seeing some of those little moons getting bigger and bigger in size that is telling us we're converging to some important answer. And that is how we get finally to the answer. So if we put them side by side, just visually, you can get a sense that a classical circuit can never exploit the vast computational space that a quantum circuit can. It's the visualizations give you this intuition of the richness with which we can access this large state. And the beautiful part of this is that we can have many classes of circuits to do different functions. So for example, this is a circuit that is designed to model financial risk. Here's an example of a circuit for the world of chemistry, to measure the energy of lithium hydride that is important for the development of new battery technology. Here's a circuit that is a unique form of a classifier, a new approach to do machine learning applications. So what we're very excited to share with all of you is the fact that libraries of these circuits are emerging and that there's gonna be this richness of circuits that allow you to perform different applications that we are gonna be able to embed in the programs that we all use and create to create business value. Libraries of circuits that get embedded in classical programs that now combine the best of classical programming and the best of quantum power. It is important to realize that not all circuits have equal value and that there will be circuits that you can implement classically far, far better because if you're not exploiting that computational space, if you're not exploiting these ideas of superposition and interference and entanglement in the proper way, you don't get there. Now, we're very fortunate that some of the best minds in the world are thinking really hard about this problem about what are the classes of circuits that give us a path to quantum advantage. And I'll highlight this seminal paper that was published last year in Science by Sergei Bravy and his collaborator. Now, Sergei works here in the Yorktown lab and he is absolutely one of the top theorists in the world in quantum. And this paper, what he and his collaborator shown is the fact that you can have shallow circuits, meaning these kind of quantum circuits that don't have a massive amount of depth to them, that have a provable theoretical advantage compared to their classical implementation. This is the first time where we have a mathematical proof that these classes of circuits have proven advantages compared to classical ones. So guided by these theoretical insights that are so profound, it is how we are chartering the path of creating these libraries of circuits. And this is how you will consume them. You're not gonna need to learn a new quantum programming language. You can use your favorite language and from there call upon these circuits that implement these special functions, from there map them to the right quantum systems that we have available on IBM cloud and then return the answer back to you as the programmer. Now, I know you hear that some are telling you that you gotta learn, you know, a new quantum programming language. We believe that we need to exploit the richness of all the programming languages that we have today, whether it's Python or C++ or whatever really you love. And that circuits is what will provide you the quantum value and behind the IBM cloud, you will get to benefit from both the library and a large portfolio of hardware designed to implement those circuits efficiently. This is not an aspirational future. We really have many quantum systems on IBM cloud. Here is the growth of the number of quantum systems we have made available on IBM cloud and to the community since we launched the world's first service in May, 2016. You can see that at present, we have 18 quantum systems that we've made available through the cloud. So let's go take a look at a few so that you can see what they look like. So welcome to one of our quantum computation centers here in Yorktown. Remember, I showed you that golden chandelier before where you saw those wires that went down all the way to the quantum processor where this is what a quantum computer looks like, a superconducting one, when it's encased. We have to protect it for the environment so that's why it shielded. And all the way at the very bottom here is where the quantum processor sits. And as I mentioned, this is one of the coldest places in the universe, right down here, 15 millikelvin. Let's listen for a minute to the sound of a quantum computer. You hear, like, this chirping sound, tintin, tintin, tintin. That is linked to the flow of gases that we pump through the system to be able to extract energy from the system and cool it to the very low temperatures. Let's go take a look at some of the other key components of a quantum computer. In this case, is the control electronics. Here, this custom made control electronics is how we convert the zeros and ones to microwave pulses at about five gigahertz. And we send those pulses into the quantum computer to perform the quantum operations and to execute our quantum circuits. So let's just quickly take a final look and some of the other systems. Here are the four systems out of the 18 that I mentioned to you. And let's take one last look and one of the other ones. This is one of my favorite places in the world, right? It is so wonderful to see a whole new paradigm of computation and one of the very first quantum computation centers that have ever been created. And this is how the community has grown since May 2016. What you're seeing here is a visualization of the number of quantum circuits executed in the IBM cloud quantum portfolio. And look how it has grown over the years. Over 180 billion quantum circuits have been executed with over 230,000 users all over the world. If you look now from an institutional perspective of companies and universities who have joined the IBM Q network by far the largest community of institutions collaborating and exploiting the power of quantum computing, you can also see the fantastic growth and the excitement that is behind the scenes. And what are the cases that are being explored by these community of institutions and businesses? Well, a broad variety of problems ranging from chemistry and developing new materials to optimization to artificial intelligence and to scenario simulation. Let me give you just a few examples. Daimler, it's a wonderful partner in the IBM Q network. And in their case, they're very interested in the power of quantum to look at advanced battery materials. And the possibilities of using quantum for manufacturing defect inspection. In this case, I'm highlighting on the bottom right the example and example of one of the circuits that are being used to perform these use cases. Or let's take out the case still staying in the field of chemistry and materials of ExxonMobil. In this case, they're very interested in modeling the thermal reaction modeling for petrochemical R&D. And here, we're seeing some of the example of the circuit that is being used to perform these calculations. And by the way, we developed these circuits through a collaboration between IBM and our partners as we pursue this work. Let me give another example for a different industry. This is in the case of our wonderful partner JP Morgan Chase, where we work on the problem of optimization. And here, we're looking at the potential to find better solutions to more efficiently than what we can do classically to improve things like options pricing and fraud detection or portfolio optimization. And you're seeing here an example of circuits that have been developed for European derivative pricing. Now, when you look at the capability today, I think it's also a very important question to say, well, what is the roadmap? How much better are these systems and how much more complex are gonna be the circuits that we're able to execute? And for that, we have developed a very important metric called quantum volume. Sometimes you may read or hear in the popular press the number of qubits and as a measure of progress. I'll tell you straight that just the number of qubits in the machine is not a good measure of the power of a quantum computer. There's another axis that is vitally important and that is the error rate in these qubits. These qubits and these qubit machines are not perfect today and they have errors that are introduced by the coupling of the system to the external world. The very power of these qubits of entanglement and those properties are also make them both powerful and delicate to the intersection and connection to the outside world. So you need to lower the error rate while you increase the qubit count. So that's where it's depicted here. What you really wanna do is you wanna move in the diagonal and quantum volume is a measure that incorporates both of those metrics and some key additional considerations. And the good news is that we are in a new exponential here. And what we have demonstrated experimentally and by building systems that we have deployed to all the clients of the IVNQ network is that we are at least doubling quantum volume every year. We've gone from a quantum volume of four to eight to 16 to 32 on a yearly basis and we are committed to at least double it, right? Year on year basis. If we keep this space or even faster, we are gonna see really spectacular results with quantum computing. Let me close with a reflection on the urgency of science, the urgency of tackling some of the most complex problems in the world. And I'll put that reflection that a revolution in discovery and a revolution on mission critical applications is gonna be power by the most exciting time in computing in 60 years, where what we're witnessing is the convergence of bits, neurons and qubits. Bits, the classical world that we've all known and the computers we've known. Cubits, this new fundamental way to process information that we explored today. And the world of neurons and neural architectures that are the basis of artificial intelligence. It is not that one will eat the other. The most profound implication of what is happening today in computing is the convergence of bits, neurons and qubits. Now this convergence is gonna be orchestrated through a hybrid cloud architecture and on top of it, to mask the complexity of the underlying infrastructure, we are gonna be assisted by artificial intelligence in the way we program, AI assisted programming. The consequence of all of this coming together will be nothing short than a revolution on how science itself is practiced and the rate at which we could perform accelerated discovery and a whole new class of intelligent mission critical applications. Thank you.