Quantum Solutions with Rigetti Computing
Subodh Kulkarni is CEO and President at Rigetti Computing, and our topic in this episode is quantum computing and its implications for our changing planet. Rigetti Computing, a publicly listed company, builds quantum computers and superconducting quantum processors. It also provides cloud-based access to quantum computing for both commercial and research purposes.
Quantum computing, a field rooted in theoretical physics, is complex and involves the use of quantum bits or qubits, which differ from binary computing's zeros and ones. As the world's computing demand continues to soar with AI and data centers consuming vast amounts of energy, AI also has the potential to unlock climate change solutions through innovations in biotech, advanced energy, planetary modeling, and more.
We wanted to learn from Subodh about how quantum computing might change the game in all of these areas as it matures. Is it possible that quantum computing is an enormous future climate solution? Let's find out.
Episode recorded on Jan 23, 2024 (Published on Mar 4, 2024)
In this episode, we cover:
[02:39]: Subodh’s background and joining Rigetti as CEO
[05:26]: Basics of quantum computing versus classical computing
[09:19]: Introduction to the subatomic world
[11:46]: Physicality of a quantum computer
[15:01]: Rigetti's business model: cloud access and on-premise quantum computers
[17:06]: Qubit development and the potential to follow Moore's law
[18:02]: The power requirements of maintaining a hyper-cold environment for quantum computing
[20:02]: Quantum computing's potential to reduce data center energy consumption
[24:25]: Quantum's strength in probabilistic versus deterministic computation
[29:26]: Rigetti's involvement in DOE’s fusion project
[30:59]: Predictions for quantum computing advancements in the next 10 years
[33:45]: How quantum computing can bolster artificial intelligence
[36:18]: Current state of quantum computing and near-term inflection points
[37:16]: Initial use cases for quantum computing, Rigetti’s commercialization strategies, and government funding
[41:08]: Collaborative projects with major financial and pharmaceutical companies
[42:12]: Potential applications in weather modeling and synthetic biology
-
Cody Simms (00:00):
Today on My Climate Journey, our topic is quantum computing and what implications the technology may have for our changing planet. Our guest is Subodh Kulkarni, CEO and President at Rigetti Computing, a publicly listed company that builds quantum computers and superconducting quantum processors, and also facilitates cloud-based access to quantum compute for commercial and research use cases. Quantum computing is a complex topic that has its origins in theoretical physics.
(00:36):
And I know I won't do justice to trying to describe quantum bits or qubits and how they differ from traditional zeros and ones based binary computing. I'll let Subodh cover this, but I do recognize that the world's computing demand today is seemingly insatiable. AI and data centers are consuming an ever-increasing amount of energy, and also the power of AI is helping us unlock many solutions to climate change in the forms of biotech, advanced energy, planetary modeling, and so on.
(01:13):
And therefore, I wanted to learn from Subodh about how quantum computing might change the game in all of these areas as it matures. Is it possible that quantum computing is an enormous future climate solution? Let's find out. But before we start, I'm Cody Simms.
Yin Lu (01:33):
I'm Yin Lu.
Jason Jacobs (01:34):
I'm Jason Jacobs, and welcome to My Climate Journey.
Yin Lu (01:40):
This show is a growing body of knowledge focused on climate change and potential solutions.
Cody Simms (01:46):
In this podcast, we traverse disciplines, industries, and opinions to better understand and make sense of the formidable problem of climate change and all the ways people like you and I can help. Subodh, welcome to the show.
Subodh Kulkarni (02:00):
Thank you, Cody. Look forward to it.
Cody Simms (02:02):
I am very excited to learn from you today because I almost can't imagine a more complex topic that we could have on My Climate Journey then to discuss quantum computing and what the future of quantum computing might entail for those of us who are spending a lot of our time and energy thinking about climate change and ways to solve that problem. So thank you for joining and I have certainly some questions to dive into, but way more unknowns than knowns. And so I trust you as my guide through this conversation to just help us get our heads around this complicated topic.
Subodh Kulkarni (02:39):
Well, thank you for the opportunity. Look forward to it. I'll try to keep it simple, but by definition, as you said, quantum computing is a complex topic.
Cody Simms (02:47):
All right. Well, why don't we start just with your own personal introduction. So you stepped into the CEO role at Rigetti Computing relatively recently. Maybe just give us a bit on your background and what pulled you into this role and opportunity.
Subodh Kulkarni (03:02):
Sure. So I joined Rigetti just about a year ago as the CEO of the company. Before that, I grew up in India. I did my undergraduate in engineering over there. Then I was at MIT here in Cambridge, Massachusetts, did my master's and PhD in semiconductors. And then I've been in and out of semiconductor industry for more than three decades. Started my career as a fab engineer actually at IBM. More recently for the last decade or so, I was a CEO of a semi camp company based in the Twin Cities in Minnesota as the CEO of CyberOptics.
(03:37):
So what got me into Rigetti, CyberOptics got sold about a year ago. So I was looking for new exciting ventures to do, and quantum computing is certainly an area that I was tracking and looking at. All of us in the semiconductor industry are generally aware of quantum computing and monitor it, but this opportunity came. I thought it was a fantastic opportunity to come to a young startup company in Silicon Valley, which is where I am right now, and take it to the next stage.
(04:04):
Quantum computing started as a research academic kind of a domain. But more recently, as progress is getting better and faster, quantum computing area does need people like myself who come with decades of experience of manufacturing and commercializing products and so on. The company was founded by Chad Rigetti about 10 years ago. He was a graduate of Yale, and then was working at IBM. Separated from IBM because he had some very clever ideas on how to build a quantum computer that is better and different than what IBM was up to.
(04:36):
And that's what the company was focused on. The board and Chad got into separate. Rigetti went public in early 2022, and then there were several issues the company ran into as a public company. It's very different to be a technology startup versus a public company. Somehow the board and Chad separated ways. That's why they were looking for someone like myself with public company CEO experience to come and help the company get to the next stage.
(05:01):
So I looked at the opportunity, thought it was an exciting opportunity to come to Rigetti, bring my experience and knowhow and take the company from what was a scientific startup to more of a mature technology company that is commercializing products and taking it to customers and actually making a difference in the commercial world.
Cody Simms (05:19):
Great. Thank you for that background introduction. I'm going to start with just the most basic question. What is quantum computing?
Subodh Kulkarni (05:26):
Quantum computing is an exciting new technology. It's a very disruptive technology. Fundamentally, think of your standard classical computers as something that uses a what we call a transistor with a zero or one. It's a basic unit of a digital world, a transistor that uses zero/one, and we do all the computations today using those zeros/ones and all different kinds of permutations and combinations that are enabled by the transistors. Quantum computing is fundamentally different than a zero and one world.
(05:54):
We use what is called as a super imposition type principle where you can have zero/one and multiple states in between all at the same time. That's the classic Schrödinger's cat, is it there or not there, type discussion. But fundamentally in the quantum world, we can have multiple states acting simultaneously. But when you go and try to find out what state it is in, you lose it. So it's one of those tricky aspects of nature where something is in a quantum state, but you don't know what exact state it is in.
(06:24):
And that's what quantum computing's challenge of explaining, but also though it unleashes the potential. So realistically what it does is when you have n bits, that's what we're calling the classical computers, your computation power typically goes by two multiplied by n. In the quantum world, when we have N, we call them qubits instead of bits. When you have n qubits, the computation power goes two raise to n, so two multiplied by n in classical versus two raise to n in quantum world.
(06:55):
When n is small, five or 10, that doesn't make a big difference. But as n gets larger, 50 or 100, two raised to 50 or two raised to 100 is practically infinity, whereas two multiplied by 50 or two multiplied by 100 are relatively small numbers. So as we increase the number of qubits, the ability to do computation is exponentially higher than classical. So that gives us tremendous power in terms of performance. We can be literally millions or billions of times faster and better than a classical computer.
(07:25):
More relevant to this podcast and it's an equally important thing is we can reduce the energy consumption by the same amount of order of magnitude. I mean, if you take a classical computer today and look at a supercomputer, those things literally occupy a whole building and consume electricity practically equivalent to a small city or something like that. And we can reduce that by million times or billion times. So it's not only faster, better computation, but also much, much cheaper and more practical from an energy standpoint.
(07:58):
So it's a combination of the two that really unleashes the potential of quantum computing. We could solve problems that are intractable today like weather forecasting or some of those impossible problems to solve with classical computers today. We can take them all in quantum computing. So it really is a new breakthrough technology. After the invention of the transistor in the '50s and '60s, it's the first time we in the computer world are seeing something that powerful on the horizon that has the potential to significantly alter the trajectory of the world in the future.
Cody Simms (08:32):
I can't wait to spend most of our conversation diving into some of what you were just sharing about the energy consumption, the future use cases, and whatnot. Before we go deeper on that, I just want to help both myself and our listeners make sure we understand at least a bit more on the 101 side of what is quantum computing so we can have a dinner table conversation with our friends and family about it and at least hit on some of the key points.
(08:55):
So thanks for explaining the ones and zeros of traditional computing and the qubits or sort of the infinite possibility between zero and one of where a computation may lie. I don't know if I'm even speaking correctly there. One thing I understand about the quantum computing world has something to do with subatomic particles. And of course, I'm now channeling Marvel and Ant-Man to pull on some of that. But maybe explain a bit about what the subatomic world has to do with all of this.
Subodh Kulkarni (09:24):
Well, fundamentally, we are dealing with electrons, photons, those kinds of things. So we are definitely in the subatomic world, and we are essentially manipulating or trying to manipulate those states to get the desired result. I mean, electrons have a spin, they spin all over, and we are trying to capture that spin of the electron. And we are trying to manipulate it and adjust it. That's where the physics starts getting fairly complicated. But intrinsically, it works. I mean, this original concept of a quantum computer, if you will, started with Richard Feynman in the 1970s and 1980s.
(10:01):
He got a Nobel Prize for all that work at that time. And subsequent to that, there are many theoretical physicists, at that time they were just doing everything on paper, that came up with, if those quantum concepts exist, how could we use them to do some computation? And that's not very different than how the classical computer works. I mean, if you just tell someone, "Here's a transistor that tells you if it's a zero or one. Now convert that into a calculator and a computer," it seems like a big jump from a simple device that is making zeros and ones.
(10:32):
But there are various developments that happened along the way such as the definition of gates, the NAND gates, AND gates, and so on that scientists and engineers use to take a simple zero/one transistor device and make it into a powerful supercomputers or your smartphones that we all use today. And the quantum world is not that different. Fundamentally, in the qubit world that we are making today, we have the zeros/ones and all the states in between.
(10:58):
So a little more complicated than the transistor, but we are trying to identify the different states and trying to make sure that the states can be identified from each other. And then we do something similar to what the classical world does, which is gates and stuff like that. I won't get into all the details, but there are different gates for quantum qubits than the classical bits, but conceptually they're not that different. We are still using the same concepts of if there is a charger, how does that transfer to the other qubit?
(11:25):
And if we can transfer it, then how can we use it to do some computations, simple ones, plus/minuses and so on? And then it gets more and more complex from there. But fundamentally, because we have the potential to do zero/one and everything in between, we get a lot more powerful machine for a given device.
Cody Simms (11:46):
And today's quantum computing machines physically look like this high-tech chandelier kind of thing. Can you maybe describe a bit about the physicality of the machine?
Subodh Kulkarni (11:58):
So yeah. I mean, there are various ways of doing quantum computing. We're now getting into a little bit of the details here. So there's probably about 10 to 12 different ways of doing quantum computing. One of them is what we do at Rigetti Computing is called superconducting quantum computing.
(12:13):
So we're essentially using very, very cold devices where they're exhibiting superconducting properties, and then we are using that to the quantum computation. But there are other ways of doing it. There're photonics. There's trap ion. There's atomic. There's topology, spin, and many different ways of doing quantum computer.
Cody Simms (12:32):
I've heard one of the big deltas is that in China a lot of it is light and photon and optical-based, generally speaking, and the US is taking more of this hyper cold approach. Yeah?
Subodh Kulkarni (12:42):
There are different things. Frankly, none of us know exactly all the areas where China is investing in right now. None of us in the Western world, I should say, know what exactly China is doing. They are definitely investing in superconducting quantum computing. They're also definitely investing in photonics and other approaches too, but we don't know the details of what exactly they're doing. They're being very secretive about it.
Cody Simms (13:02):
So describe the Rigetti system then maybe physically.
Subodh Kulkarni (13:05):
So let's go to the superconducting, which is what Rigetti does. It's the same work IBM, Google, Amazon, and a few other companies do. So yeah, you're right. When it's closed and it's working, it just looks like a cylinder with a control stack next to it. It's very boring, to be honest. It's the size of a human being. Nothing fancy. It sits in ambient conditions. You can visit our facility in Berkeley or Fremont in California and you can take a look at it, or I'm sure you can even go on the internet and see what a king quantum computer looks like.
(13:33):
It's kind of a cylinder with a stack next to it. When you open the cylinder, that's when things start looking interesting. Then you see this golden color chandelier-like structure. So what you see is this baseplate, and then underneath that there are multiple baseplates with all kinds of wires and cooling tubes coming down. And essentially what it is is it's a refrigeration system. We call it a dilution refrigerator, and it's taking temperature from ambient temperature all the way down to the chips, 10 millikelvin.
(14:04):
Just to emphasize that, I mean, 10 millikelvin is really, really cold. It's the coldest you can find on earth right now as far as I know. So that's the temperature we take the chip to, which is when it starts displaying the quantum effects. So the big chandelier are cooling tubes, trying to reduce the temperature in stages. There's also the electrical circuitry and signaling that goes along with it. We typically use microwave kind of signals to send us inputs to the chip and get output back from the chip.
(14:35):
The chandelier looks aesthetically very, very interesting and appealing, but at the heart of it, it is essentially a cooling device and an electrical signal transmission device.
Cody Simms (14:46):
And so based on this, Rigetti has built your own hardware device and then you also have essentially a software platform on top of that that is a developer platform. Am I understanding Rigetti's business at least at a high level appropriately?
Subodh Kulkarni (15:01):
Yeah. I mean, in general what we do is we build quantum computers from scratch. So we have a fab where we build a chip. We buy dilution refrigeration components from commercial companies. We put the whole computer system together, including control systems and everything. And then we have two business models right now. One is we offer the quantum computer on cloud so you can access it through our own cloud service or we also offer it on AWS and Microsoft Azure.
(15:28):
So anyone who has access to AWS or Azure can access our quantum computer. So that's one way customers can use our quantum computer. Other more interesting development in the last year or so is many of these researchers want an on-premise quantum computer, because there's a lot more things they can do when it's under their control. It's dedicated to their use. They don't have to be in a line to use it or anything like that. Many people are very sensitive about sending their data over the cloud, so they prefer having an on-premise computer.
(15:59):
So in the last few months, we have physically started shipping quantum computers to customers. So far we have sold only two. One is to the Department of Energy's Fermilab in Chicago and another one to another major national lab. We cannot disclose to who. But we are talking to more and more customers, national labs, universities, those kinds of customers who are interested in physically buying a quantum computer from us. So two models.
(16:24):
You can access quantum computer through the cloud and just pay by the minute or hour, or you can physically get a quantum computer. But the great news is the technology is at a point where it's working. You can try it out yourself. It's doing the calculations that you expect it to do. The challenge is we are still not significantly better than classical computer or cheaper than classical computer.
(16:44):
We can see the potential, but it's still very much in early R&D stages. So we need to keep improving it until it becomes significantly better than classical computers to unleash the advantages that I described earlier.
Cody Simms (16:57):
So far, does there seem to be any sort of Moore's law type of curve with respect to the qubit development pathway?
Subodh Kulkarni (17:06):
Over time, we expect something like a Moore's law will come along. However, today we make 84 qubit quantum computers. 84 is a relatively small number compared to your billions of transistors that you're talking about in a single chip. So we have plenty of distance between the qubits on the chip today. But again, keep in mind that we have the power of exponent. So it's going to be when the qubits are perfect, a computation power of an 84 qubit device would be two raise to 84.
(17:31):
We are not perfect by any means, so we are still far away from that point, but that's the potential you are going to get of two raise to 84. And you don't need more than 50 perfect qubits to rival the world's best supercomputer today. So we may have a Moore's law type some kind of a projection, but it's not going to be the same challenges that the semiconductor industry went through in packing all the transistors up down to the three nanometer and two nanometer levels. So we will have our own challenges, but not exactly the same as a semiconductor.
Cody Simms (18:02):
Okay, let's dive into the climate related questions. I think this is super helpful overview of the space. Helped me get definitely more familiar. Let's start with the hyper cold environment. I want to start on the input side, and then we'll move our way toward the implications side in the future. But on the input side, you mentioned it requires this hyper cold environment. That presumably has some degree of power requirement and some degree of potentially emissions impact to achieve that hyper cold environment in the first place. Can you describe that a bit more?
Subodh Kulkarni (18:36):
Sure. So dilution refrigeration is the technology we used to cool down the chips. It does need power obviously, but it's not too demanding from an energy input standpoint. I mean, we use standard voltages, standard currents. As I said, it's the size of a cylinder, so obviously it consumes more power than your household refrigerator. But size wise and, for all practical purposes, power wise, it's not that much more than what your standard refrigerators are.
(19:05):
So yes, it does power... And once the chip is cold, entire power system, everything is to cool down the chip, once it's cold, I mean, you need to keep it at that cold condition, but it's not like it's energy intensive appliance that needs to be continuously fed with tons and tons of electricity. It's more or less a standard off the shelf technology that we are using, but it does need some power to cool down the chip.
Cody Simms (19:28):
And what we've seen so far in AI, which is using classic transistor style chips, right, not quantum chips, is just this incredible insatiable demand for power, such that some of the largest data centers in the world are becoming power generators. They're doing huge power purchase agreements and working to do on-prem power in many cases. How do you see the data footprints of quantum growing as eventually demand exists for more and more compute here?
Subodh Kulkarni (20:02):
Frankly, that's the part... I mean, everyone focuses on the power of quantum computing in terms of performance and speed and all those things. I personally get a little more excited more because of the power situation than the performance side. And the reason for that is exactly what you said. I mean, today's AI, I mean, you take ChatGPT or OpenAI type examples and they're using state of art GPUs, I mean, they're basically energy guzzlers. If you look at how much power it takes to fire up those GPUs and those devices, it's a lot.
(20:33):
And there's been many people who have studied this, but data center today, data centers worldwide are consuming on the order of 5% of the world's total energy. And a decade ago that number was less than 1%. So we have quickly come from 1% to 5% of the global energy is being consumed by data centers today. And just to give you a feel, the entire automobile industry that we all focus on from EV standpoint and everything is only 10%. So sure, auto is large, but it's relatively flat at 10%.
(21:06):
And here in the data center world, we have come from 1% to 5%. If the current trend continues, data centers will be the maximum guzzlers of energy a couple of decades from now. And obviously we can't do that. I mean, the world cannot sustain itself if that trend just continues and everything stays the same and starts consuming more and more power. So something has to change in the energy consumption of chips. I mean, clearly there are technologies like ARM that are trying to do low power chips and that's great, and I wish them all the success and they should do it.
(21:39):
But beyond low power chips, we need something else, and I believe quantum may provide the answer to that. We consume a lot, lot less energy because of the number of qubits we are talking about. So despite the need for some energy for dilution refrigeration at all, our total energy footprint is orders of magnitude smaller than state of the art classical computers. And that has the potential to really knock off the requirements for data centers.
(22:07):
So instead of just going from 5% to 10% to 20% of world's energy, we could certainly knock that whole number down by orders of magnitude. And that's really the way I think of it. It's the combination of performance and cost and energy consumption that's going to allow the deployment of quantum computers at scale. So clearly it's a potential answer for some of the energy problems the world is seeing right now created by data centers.
Yin Lu (22:36):
Hey, everyone. I'm Yin, a partner at MCJ Collective, here to take a quick minute to tell you about our MCJ Membership Community, which was born out of a collective thirst for peer-to-peer learning and doing that goes beyond just listening to the podcast. We started in 2019 and have grown to thousands of members globally. Each week we're inspired by people who join with different backgrounds and points of view. What we all share is a deep curiosity to learn and a bias to action around ways to accelerate solutions to climate change.
(23:01):
Some awesome initiatives have come out of the community. A number of founding teams have met, several nonprofits have been established, and a bunch of hiring has been done. Many early stage investments have been made, as well as ongoing events and programming like monthly Women in Climate meetups, idea jam sessions for early stage founders, climate book club, art workshops, and more. Whether you've been in the climate space for a while or just embarking on your journey, having a community to support you is important.
(23:27):
If you want to learn more, head over to mcjcollective.com and click on the members tab at the top. Thanks and enjoy the rest of the show.
Cody Simms (23:36):
And just to clarify, the percentages you were citing, which obviously are generally rounded up or down numbers, those are energy consumption, not emissions, right? So there is a delta or a difference obviously in the emission side of it depending on the type of energy you're using.
Subodh Kulkarni (23:48):
Yeah. But as all of us know, just about any type of energy when you're consuming, there is an emission part that goes with it. It may not be immediately obvious depending on if it's green energy or fossil-based energy, but there's always some kind of a negative artifact associated with consuming that energy.
Cody Simms (24:05):
So now talking about some of the outputs that might be possible from quantum computing, let's start with a more general question, what kinds of things do you anticipate quantum being significantly better at doing than traditional computing, even the most powerful AI systems we can imagine today?
Subodh Kulkarni (24:25):
My simplistic way to look at it is: quantum computers seem to be better for probabilistic computation compared to deterministic. So when you're dealing with problems which have... Fundamentally, the question is, what are the chances of something? What are the chances of an economic recession happening later this year? What are the chances of company X going into bankruptcy in the next five years?
(24:49):
What are the chances of this particular transaction on a credit card has a fraud element involved with it. All those kinds of applications are extremely difficult for classical computers, because you've got billions of inputs and transactions and all that stuff to sift through and try to find out where the problem may be. Those kinds of applications fit very well with quantum computing and how the quantum computer works. On the flip side, problems that are deterministic in nature, like you have the equation, you know the answer, you just crunch the numbers.
(25:20):
And the faster the speed, the better. Those problems may very well stay with the current state of art CPUs and GPU technologies because they're deterministic. There is no problem. [Inaudible 00:25:30] So another way to look at it is I tend to think of a quantum computer working like a human brain, and think of a human brain as we have millions of neurons that are all entangled with each other all the time.
(25:44):
When we get any input, a few of the neurons get excited because of that input, the remaining neurons stay idle, and they all simultaneously act on that input and outcomes to output. I mean, neurons don't have zeros and ones and don't transmit a signal one at a time, one at a time, sequentially. So quantum computer works something like the way a human brain works. So we have n qubits and they're all entangled. And we take the inputs, put it on those n qubits, and we look at the output coming out of those n qubits.
(26:15):
So very analogous to some extent in the way human brain works. And then you say, "Where is a human brain better than a machine," assuming our speed circuit. And intrinsically, we can deal with this softer side of things where probabilities are involved, where judgment is involved better with a human brain than with just a fast machine. And I think that's the way quantum computers will evolve.
(26:38):
Obviously will be better than human brain in terms of the accuracy and all that stuff. But fundamentally, because our qubits are entangled and they can deal with complex streams of inputs, they may work better than classical when we are dealing with those problems.
Cody Simms (26:51):
So that would lead me to, if I pulled on that thread, I would conclude that quantum may be better at, for example, solving biological problems than chemical ones, where there's this factor of mutation and unknown and accident that could come into play versus something that is fairly, like you said, deterministic, where you expect some kind of reaction to happen.
(27:16):
So I would think in that case, quantum might be good at something like complex earth system models of how might climate change based on a bunch of different inputs happening at once or better understanding of ocean currents and potential future changes that may happen in our ocean currents. Am I on the right track here?
Subodh Kulkarni (27:37):
I think you're on the right track. A lot of very high quality researchers are working on exactly what we are talking about right now, so I don't claim to be an authority in this area, but they are reaching similar conclusions that extremely complex problems with many, many different input data streams where interaction is difficult to predict and model, they belong better to quantum, so like biological systems classic, protein unfolding, folding, that leads to drug discovery. Another problem. It's hard for classical computers to deal with those models. Those may fit better with quantum.
Cody Simms (28:11):
We're starting to see in climate tech lessons being learned from biotech, particularly on the pharmaceutical and therapeutic side now starting to be applied toward plant-based product delivery and discovery.
(28:25):
And so if you can use, essentially it can improve your ability to do synthetic biology or biological manipulation and understand outputs in order to lab grow something as an example, that could be a natural product that doesn't require you to go out and cut down trees or cut down the natural world because you can build it in a lab. That seems like the kind of use case I'm hearing from you that quantum could be helpful with.
Subodh Kulkarni (28:49):
Exactly. I think you latched onto the right one. There are many, many modeling experiments we can do without disturbing something in real life that can make your experiments a lot more effective when it comes to the actual experimental part. So that's exactly what many people are looking at quantum computers as a potential way to reduce the wastage in the world. Right now a lot of experiments that get done could be avoided if you do the right kind of modeling a priori.
Cody Simms (29:14):
Animal testing as an example, right? When I think of the big moonshot technologies that are being developed in the 21st century, I think of quantum computing and I think of fusion power. Do you see anything in the fusion realm? Whether it's understanding the potential instability of hydrogen gas, as an example. I don't know. Are these the kinds of things that you see coming together like peanut butter and jelly in the future potentially as well?
Subodh Kulkarni (29:41):
I'm glad you asked that because actually we are working right now with Department of Energy on the fusion project. So clearly Department of Energy and in specific the Lawrence Livermore National Lab here in the Bay Area, they are working on exactly what you are talking about right now, fusing two hydrogen atoms. In their case, they take hydrogen and deuterium, but fundamentally, it's the same. You're recreating the reactions that exist on stars like the sun on earth now, and you're trying to create energy from that.
(30:10):
And that's one of the holy grails of mankind. I mean, if you can come up with fusing atoms and generating energy on earth, it would solve a lot of our energy issues that we have today in the world. And so there's a lot of research going on in fusion. We are involved with Department of Energy right now to try to understand the merits of using quantum computing.
(30:31):
Even though our quantum computers are still very early in research stage, we are trying to understand how some of the fusion models can be used on quantum computers and what's the value of deploying quantum computers in fusion. Still early research, but some very interesting work coming out from that area. We certainly hope that quantum computers play a significant role in enabling and helping the fusion researchers long term.
Cody Simms (30:55):
What do you think the next 10 years looks like in your space?
Subodh Kulkarni (30:59):
It's exciting. We are still early in the quantum computing, but the exciting part is unlike some other scientific areas that don't pan out and they never really materialize, in the quantum computing world, it is working out. I mean, we have machines that are working and people can use them for computation. The challenge is closing the gap, but we are literally at 84 qubit and what we call fidelity. We are right now at median fidelity of 98%. We are starting to approach parity with classicals.
(31:27):
Once we get to several hundred qubit at 99.5% fidelity and we think we'll be there in a couple of years, then we can exceed the power of classical computing. That is super exciting. So in the next five to 10 years, for sure, those of us who are in the quantum computing space right now feel that we will get to what is called narrow quantum advantage and eventually quantum advantage and quantum supremacy. And that's where things really take off, where we can take on the problems that we discussed earlier, the biological problems or fusion type problems.
(31:59):
We certainly see 10 years from now quantum computers coexisting. I don't think they're going to... No technology is ever that good to just disturb and completely change the existing technology. Those things take a decade or two far eventually. But 10 years from now, I see a world where CPUs, GPUs, and what we call QPUs, quantum processing units, will coexist. The QPUs will take on problems that are more probabilistic in nature, difficult for CPUs and GPUs to tackle complex data streams, and that kind of stuff.
(32:33):
And that will go to the QPUs. Pure speed will stay with CPU, and parallel stream, that kind of stuff will stay with GPU. So the three will coexist 10 years from now. Hopefully the combination of the three in the hybrid world, you are getting a lot more bang for the buck, if you will, in terms of computational power, but also energy consumption. I certainly think it's going to be an exciting decade for QPU development and how that gets into the mainstream of computational world.
Cody Simms (33:01):
It feels to me like we're at a bit of an inflection point or a precipice, if you will, where if you look at what started the Industrial Revolution, it was really harnessing thermodynamics and harnessing a little bit of control over the ability to manufacture and produce things. You look at the Information Age and it was about harnessing electromagnetism, harnessing compute, the ability to process things beyond what our fingers could touch.
(33:27):
And now we're getting into this subatomic world that allows, like you said, potentially almost infinitely more powerful algorithms on understanding probabilities. What does that unlock? Do we have a sense of what this new age we could enter might be?
Subodh Kulkarni (33:45):
The way I try to look at it simplistically is, I mean, AI, we are already beginning to start seeing the tremendous power of AI and what it could do. But I think all of us who use AI or close to it also see the limitations of things like hallucination, the energy consumption, the GPUs, and many issues that come with AI. So we have some limits on what we can do today. I'm sure it will continue to improve. My view is quantum computing gives it a huge boost. So you're right, we are going through inflection points as we speak.
(34:19):
So this current inflection point that we are going through in quantum computing where we came from basically theory on paper and demonstration of something in a lab setting to something that can be shipped to customers, that they can start using it. So significant inflection point we are going through right now. And then there will be another inflection point in my view in five years or so when we actually exceed the power of classical computers. And the combination of AI and quantum computing I think could really help us.
(34:45):
I mean, I look at the industrial world at large and the amount of inefficiencies we have, be it wastage or energy emissions that we touched on. And there's a lot of things that we know as mankind we are not doing it correctly. We are doing it because it's practical, it's the cheapest way. There's needs for goods and services and all those other things are overriding and that's what we are doing. But assuming machines can be designed and they are doing a much, much smarter job than what we can, I can see how we could truly take the world to the next stage.
(35:21):
We will reduce inefficiencies. Hopefully we can tackle some of the fundamental problems that we are dealing with at a global level right now in terms of distribution, logistics, those kinds of things. I mean, the world has enough food as we all know. The problem is distributing it to the places where people can access them. And a lot of it is because of the lack of optimization. I mean, it's not that we don't have physical devices to ship food from point A to point B. But it's how do you optimize?
(35:49):
How do you make sure that the right amount of food is in the right place to be taken to the right location? And those kinds of problems, hopefully, we will be able to take on with AI plus quantum computing, some combination. So I think [inaudible 00:36:03] by nature and I think the power that we are creating hopefully will tackle some of the most fundamental problems we as mankind deal with today.
Cody Simms (36:11):
It strikes me that water faces a similar problem. We have enough water. Sometimes it's too much of it somewhere, and sometimes there's not enough of it somewhere.
Subodh Kulkarni (36:17):
Yeah, exactly.
Cody Simms (36:18):
So with all of this, where are we today?
Subodh Kulkarni (36:21):
From a quantum computing standpoint, we are still in infancy. I don't want to get too excited and say that we are there. I think we are entering an inflection point, as you correctly pointed out. And we are getting from small lab experimental devices to physical devices that are actually starting to be shippable to customers where they can do their application development and so on. But still a relatively minor inflection point.
(36:46):
Roughly five years from now, I think we'll go through a bigger inflection point where we will say just a quantum machine, QPU, that is significantly better than your CPU, GPU in terms of performance or cost or energy consumption, some combination of those things. And that's what will really unleash, and then the next decade after that will be pretty exciting. So I think we are still infancy as far as quantum computing is concerned with a couple inflection points ahead of us, one pretty soon or we may be there right now and one probably three to five years from now.
Cody Simms (37:16):
With Rigetti, you're running a public company. So there is pressure to have it be a business that generates revenue and appeases shareholders and makes them excited to continue to invest. You mentioned one part of commercialization for the company, which you said was relatively new, which is actually selling quantum computers right now primarily to US government departments.
(37:38):
If you want to share more on that, great. I'm also curious what some of the initial use cases are from a commercialization perspective that you're seeing people want to explore. Financial probability calculations feels like kind of an obvious one to me, but I'm curious right now, where do you see industry leaning into quantum to try to get smart and prepare for the future?
Subodh Kulkarni (38:03):
You're right. We are a public company. Our stock trades. We have to worry about investors and all the usual things that any public company is expected to deliver. Near term, when I say near term, I'm talking the next three to five years, certainly our monetizing capability is in allowing our quantum computers on the cloud, which we're already doing and expecting more demand coming from the cloud. But more importantly, this selling on-premise QPUs primarily to national labs and universities.
(38:29):
Certainly we are dealing with DOE and DOD labs in the US, but we are also dealing with other sovereign government labs. Some of our dealings with the Government of UK are [inaudible 00:38:39] but we are dealing with other Western European and some Asian countries as well. The exciting thing is, I mean, everyone at the national lab level recognizes the power of quantum computing and many governments are increasing the budget for quantum computing.
(38:54):
Here in the US itself, despite all the disagreements that we deal with in Washington, DC, one area that there seems to be bipartisan support is funding for quantum computing and AI and protecting those two as crown jewels of the country, if you will.
Cody Simms (39:08):
I was going to say there's clearly a national security interest there and it feels like cybersecurity is a big part of that. Yeah?
Subodh Kulkarni (39:13):
Exactly. So there's full recognition right up to the senior levels of White House administration and certainly congressional representatives that AI and quantum computing, an area where we as a country are ahead and there's tremendous potential in these technologies. Let's make sure we continue to fund them correctly and let's make sure we put the right safeguards in place, inadvertently give up the technology to some adversarial countries.
(39:41):
So that's great and enables more funding coming to companies like us from DOE, DOD, and other national labs, but Western European countries and some Asian countries have come to the same conclusion. So the amount of budgets that are getting opened up for quantum computing, we're talking billions of dollars now. And organizations like IDC have quantified the number to be like $7.6 billion by 2027.
(40:06):
So that's what a small company like us that has to worry about financial situation is focused on. So we have great technology. We'll continue to improve the technology in the next three to five years. But at the same time, there is this potential seven to $8 billion market opportunity in five years, and we certainly want to be one of the key players, if not the leader in that marketplace.
(40:27):
Assuming that money is for real and we can play an important role, then funding takes care of itself because we can clearly sell devices, make money off it, and then the company becomes self-sustaining. So for the next two to four years, we still need money.
(40:43):
Hopefully we'll be able to raise money through the public market or through strategic investments, and then get to the point where we are cashflow neutral, and then really unleash the power of quantum computing. So that's our plan right now.
Cody Simms (40:55):
And outside of the government side of things, I mean, the commercial use cases that would jump to my mind would be pharma, wonder drug development. We talked about finance. AI training by itself is probably a whole thing, I'm guessing.
Subodh Kulkarni (41:08):
That's now the right areas. I mean, clearly, and you can see some of our press releases in the last few months, I mean, we announced collaborative projects with Moody's in UK and Standard Chartered and HSBC and pharmaceuticals and certainly the fact that we are so closely involved with DOE and DOD and the Government of UK. I mean, we are dealing with national security type things that they're working on.
(41:30):
Some of our work with DOD is done by NASA actually, so we work very closely with NASA people, and they have some very interesting use cases. So certainly I think from a commercial side, pharmaceutical and financial industries are the first two I can see who will enter this space because of the nature of the problems they deal with. Engineering type, typical deterministic type calculations are likely to be later.
Cody Simms (41:54):
My hope is in the world that I spend a lot of my time thinking about, the types of things that find their way to be able to rely on what you're building or what you and your peers in the industry are building would be things like synthetic biology, food, catalysts for different forms of clean energy reactions.
Subodh Kulkarni (42:12):
Weather modeling, right?
Cody Simms (42:13):
Yeah. We talked about fusion. All of that feels like areas that the cleantech or climate oriented world could lean into in addition to all the different modeling exercises we talked about around atmospheric modeling and whatnot.
Subodh Kulkarni (42:25):
Yeah, that's the exciting part about having machines like this coming along.
Cody Simms (42:29):
Well, Subodh, I'm so grateful for you joining us. I know waking up every day and thinking about climate change may or may not be something you personally think about, but in building your business right now, you have a lot of just very operational things to do that this is certainly a conversation for the future. But it's really helpful for me and hopefully for our listeners to get our head around the potential of where the world may be going.
(42:54):
And a big part of MCJ is helping people have some optimism that there is a chance the future could actually have some positive things in store for us. And my hope is a lot of what you're helping to enable can be a major factor there.
Subodh Kulkarni (43:09):
Certainly. Thank you for your interest. And I'm an optimist. I think fundamentally the world has always solved the most challenging problems with innovation, and that's what we are here for. We are innovating a new technology, and hopefully it's available and the right people get involved and we will start using it for the right reasons.
Cody Simms (43:27):
Well, thank you for your time with us.
Subodh Kulkarni (43:28):
Thank you, Cody. Appreciate it. Bye.
Jason Jacobs (43:31):
Thanks again for joining us on My Climate Journey Podcast.
Cody Simms (43:35):
At MCJ Collective, we're all about powering collective innovation for climate solutions by breaking down silos and unleashing problem solving capacity.
Jason Jacobs (43:44):
If you'd like to learn more about MCJ Collective, visit us at mcjcollective.com. And if you have a guest suggestion, let us know that via Twitter, @mcjpod.
Yin Lu (43:57):
For weekly climate op-eds, jobs, community events, and investment announcements from our MCJ Venture Funds, be sure to subscribe to our newsletter on our website.
Cody Simms (44:07):
Thanks, and see you next episode.
