Microsoft just announced a massive quantum computer breakthrough that uses an entirely new state of matter. The new quantum computer uses topological superconductors to create stable qubits with low error rates.
Topological superconductors enable stable qubits by utilizing Majorana zero modes to protect quantum information from decoherence.
The result: Microsoft should have a fault-tolerant usable quantum computer this decade.
As in, before 2030.
Watch the show here, and subscribe to my YouTube channel. You can also subscribe wherever podcasts are found …
In this TechFirst, we talk with Microsoft’s head of quantum hardware, Chetan Nayak, who has been working on solving this problem for literally 19 years, and he talks us through the technology and what it means for quantum computer. He explains the methods to measure this new state non-destructively, the novel architecture that leverages it, and Microsoft’s ambitious roadmap towards building a fault-tolerant quantum computer within this decade.
The conversation dives into potential future applications, the integration of this technology into global data infrastructures, and the transformative possibilities it holds for various fields, including chemistry, materials science, and beyond.
What you’ll find in this episode
- 00:00 Introduction to Fault Tolerant Quantum Computing
- 00:48 Understanding the New Phase of Matter: Topological Superconductor
- 02:10 Properties and Applications of Superconductors
- 03:11 Creating and Engineering Topological Superconductors
- 05:16 The Significance of Topological Superconductors for Qubits
- 09:54 Measuring Quantum States with Quantum Dots
- 13:03 Building and Testing Quantum Devices
- 19:43 Future Roadmap for Quantum Processors
- 19:53 Unveiling the Quantum Roadmap
- 20:34 DARPA Collaboration and Engineering Milestones
- 21:23 Fabrication and Demonstration of the Eight Qubit Processor
- 21:43 Accelerating Quantum Progress
- 23:22 Scaling Quantum Computers for Practical Applications
- 27:04 The Long Journey of Quantum Research at Microsoft
- 33:24 Future Prospects and Challenges in Quantum Computing
- 38:10 Quantum Computing’s Role in Addressing Global Issues
- 42:32 Reflections on a 19-Year Journey
Transcript: Microsoft’s new topological superconductor quantum computer
(note: this is AI-generated; it will contain errors)
John Koetsier (00:01)
Microsoft says they have an amazing new development in quantum computing that might just change the entire field. I’m super pumped to hear about it. I have almost zero details about it except that it’s in and around topological quantum computing, but we have a global expert here to chat with us. His name is Dr. Chetan Nayak. He runs the quantum hardware program at Microsoft and he’s been running on this path for apparently almost two decades.
Let’s just start here. What’s the big news?
Chetan Nayak (01:06)
Yeah, so I would break the story into four pieces. So the first is we’ve got a new phase of matter. We created a new state of matter, okay? It’s called a topological superconductor. Now we’re used to states of matter like ice, water, steam. You change the temperature continuously in ice and as you heat it up,
You know, it changes continuously, like the density changes a little bit, property changes a little bit, right up until you get to the melting point and then something sudden happens. And very, very drastic and actually discontinuous. It becomes water and it flows. And you would, you know, the ice has a certain rigidity associated with it. You can build an igloo, you can build an ice sculpture with ice. You can’t do that with water, right? So something discontinuous has happened. And then again, when it boils and becomes steam. There are other, we think of those as the kind of three states of matter, but there’s actually others.
Because when you look at a solid, like you look at iron, for instance, or a permanent magnet, like these rare earth-based magnets, actually, as you warm them up or you change their properties, they could lose their magnetism. They can be magnetic or non-magnetic, and they would also go through a phase transition between, say, magnetic or non-magnetic solids. And the properties are really different. It’s very discontinuous in going from one to the other, and you would have different applications depending on the material. It turns out that there’s a… What’s that?
John Koetsier (02:31)
And there’s plasma.
Chetan Nayak (02:35)
plasma there’s also indeed there’s also it turns out superconductivity is another state of matter within solids you have solids that are metals and you know as you change the temperature the resistance changes a little bit typically as you cool them down they become better and better conducting electricity okay you have less and less heat dissipation less and less energy loss you cool them down there with many metals the first one that was discovered was mercury but it’s others aluminum lead tin lots of metals
As you cool them down at some temperature actually they stop continuously getting better and better and actually just the resistance falls off a cliff and goes to zero. Just continuously. And then below that it’s just zero. And that’s a superconductor. A superconductor is not just a really really good conductor. It’s not the colloquial use of super. It actually is a perfect conductor. It has no resistance and it actually turns out to have something called the Meissner effect that expels magnetic fields. And that’s an incredible state of matter.
And we’ve discovered a new state of matter called the topological superconductor. I should say we’ve created or engineered it because as far as we know, it doesn’t occur naturally in the known universe. We engineered it by putting together different materials, combinations, things that individually wouldn’t have these properties, but they’re combinations. So we have a material stack in which we’ve taken a semiconductor called indium arsenide and aluminum metal, which is super-under-aluminum. We put them together.
And this hybrid structure is more than the sum of its parts. It’s got properties that it wouldn’t otherwise have. And then we actually have to take it to extreme conditions because it’s not enough to grow this and synthesize this combination. But we actually have to not only keep it very cold, and by very cold I mean 50 millikelvin, so much colder than interstellar space, right? So much, much colder than, know, five one hundredths of a degree Celsius above absolute zero. So really, really cold.
John Koetsier (04:24)
Ouch.
Chetan Nayak (04:36)
And then we have to put a big magnetic field on top of it. So we crank up a magnetic field. we really are taking this into extreme conditions. But when we do that, it actually becomes this really novel state of matter called the topological superconductor. And what that enables it to do, and what that means is unlike most conventional superconductors, which really, really
John Koetsier (04:51)
And what does that mean? What does that do?
Chetan Nayak (05:01)
care about whether they have an even or odd number of electrons. I mean, they could have millions of electrons, but whether it’s an even or odd number actually makes a difference because the electrons form these pairs called Cooper pairs. And those Cooper pairs, their ray functions line up and they actually endow a superconductor with a certain kind of rigidity, is why they have dissipationless electrical current conduction, electrical conduction. A topological superconductor actually doesn’t mind having an odd number of electrons. Normally,
You can find that unpaired electron and there’s an energy penalty associated with it because it wants to pair up with something and there’s some kind of frustration when it doesn’t. I know and it’s Valentine’s Day so maybe appropriate, huh?
John Koetsier (05:35)
It is Valentine’s Day when we’re recording this.
Chetan Nayak (05:45)
The remarkable thing about topological superconductor, okay, is that actually it’s fine with an odd number. Okay, there’s no problem whatsoever. There’s nowhere you can find the, it hides the unpaired electron. There’s nowhere you can find the unpaired electron. It actually gets subsumed into that sea of electrons that you have in the wire. So we have a wire. The number of electrons isn’t even fixed because Cooper pairs can come in or out.
you’ve got pairs dancing in and out of this, the numbers fluctuating but always fluctuating by even numbers. And in that process this can accommodate that unpaired electron in a way that you basically can’t find it, you can’t discover it. And the way does it is there are these objects called Myron-Azir modes, I should describe them as quasi particles or features in the system.
that are at the endpoints of one of these wires. And when I say nanowires, these are a few microns long, but tens to hundreds of nanometers wide. And the Myron-Azure modes enable you effectively to hide that unpaired electron. They’re shared. mean, we use these various words to explain the math. So they’re all kind of, you know, a halfway language that’s trying to take the math and put it in colorful language. But in some sense, that unpaired electron is hidden
John Koetsier (07:07)
Yes. Yes.
Chetan Nayak (07:14)
It’s shared between two distant objects, these two myron and zero modes. If you looked at just one, you wouldn’t find it there. You look at just the other, you wouldn’t find anything there. You look anywhere in the middle, you don’t find it. The only way you could see something is if you actually brought these together, okay? And in the process of bringing them together, you then if you bring them close together, then they’re not really far separated. You know that combined they have, they either do or don’t have that unpaired electron. You could check. Now you have a place where you can look.
John Koetsier (07:29)
Mm-hmm.
Chetan Nayak (07:44)
But when they’re far apart, there isn’t any such place. So this is a new state of matter that has these, these things are Majorana zero modes. That’s a person’s name, Majorana, Ettore Majorana was a Italian physicist who did brilliant work when he was young, disappeared mysteriously in 1938. He boarded a ship from Palermo to Naples, never got off.
John Koetsier (07:47)
Mm-hmm. Mm-hmm.
Chetan Nayak (08:12)
Nobody or was never or I should say nobody knows what he did at the end of that trip. People speculated he went to a monastery. Others said he escaped to Argentina. Some say he committed suicide or fell. Nobody knows. you know, it’s a person’s name. And what’s remarkable is you now got this amazing thing where you you basically have hidden the fact that there’s an even or odd number of electrons in this wire.
And that’s fantastic for that new state of matter. It’s fantastic for making qubits because what you really want in a qubit is you want to have two states which you can really well protect so they don’t decohere because you want to be able to make superpositions and you want your superpositions to not be like Schrodinger’s cat where the cat is actually either dead or alive. It’s not really going to be in a superposition because it gets entangled with the rest of the world and it decoheres. You want to have something which doesn’t do that.
John Koetsier (08:50)
Mm-hmm.
Yes.
Chetan Nayak (09:10)
And when it’s so well hidden that you can’t even tell if it’s an even or odd number of electrons, that enables it to have that kind of quantum superposition and to not decohere. So that’s great. So that’s the first piece.
John Koetsier (09:21)
So the big news there is
that you can have stable qubits.
Chetan Nayak (09:27)
Yep, the big news there, yeah, so let me build on it. So that’s part, so it’s a bit of a structure with some, you know, with some layers to it. So I think that’s like the ground floor of this is, is new state of matter, okay? Topological superconductor. And we made this thinking like, hey, this is gonna be great for qubits, okay? So then you’ve got to do something with it to get to qubits, okay? You’ve got the even-or-oddness that’s hidden. You don’t want it be too much of a good thing though.
because if it’s so well hidden, how are you gonna do anything with it? How are you gonna know if it’s even or odd, right? Like you do all this work to hide it and then what? So you have, we don’t want the environment to measure it, but we wanna measure it when we want to, right? And not when we don’t. So we gotta figure out some way to actually access this information and do something with it. Okay, how are we gonna do that? Well, we take one of these wires and we build a device where we’ve also got a quantum dot, okay?
John Koetsier (09:59)
Exactly.
Mm-hmm.
Chetan Nayak (10:26)
QAMDOT you should just think of as a small parallel plate capacitor with some reservoir of charge on one of the plates.
We couple that quantum dot to the wire, okay? And when we do that, electrons can hop from the dot to the wire when the dots tuned up correctly. And by hopping from the dot to the wire and back, they actually hop to the Myron and zero mode and back. What they can do is they can hop to zero mode and back, but they can hop to one of these Myron and zero modes and then go all the way through the wire and come back from the other side. And in so doing, they actually get to sample the entire wire and count and they…
They pick up a minus sign for every electron you go by. And then if it’s an even or odd number, you’ve either picked up a minus sign or picked up a plus sign. So, and yeah, it’s a non-destructive in the sense that the qubit doesn’t go away. What you do is you project it on, you’ve now measured whether it’s even or odd number and it stays in that state at the end of the measurement. If you measure it to be even, it’ll stay even and so on and vice versa. So it’s what you call a
John Koetsier (11:13)
Is that a non-destructive measurement mechanism?
Chetan Nayak (11:33)
quantum non-demolition measurement in that sense. we have a way measuring it by coupling it to a quantum dot. Now, of course then the question is, okay, great, that sounds great on paper. Can you build this thing with the quantum dot? you? Yes, and that turns out that the quantum dots, its energy levels will change a little bit and as a result, its capacitance changes a little bit, okay? So if you have a capacitance change, those are things you can measure.
by putting it into a resonance circuit like an LC circuit and then seeing how the response of that circuit changes. So now there are the outlines of an idea of how to make a circuit that can measure a quantum dot and that quantum dot when it’s coupled to our wire can measure this even or oddness. But of course that capacitance shift has to be large, large enough to measure. If it’s this tiny thing then, because after all,
What we’re trying to do is we’re trying to take a device that’s got 10 million electrons or so and we’re going to say does it have 10 million or 10 million or one electrons? And we’re not even going to directly measure it. going to just, we’re just going at second order, we’re going to take a quantum dot. We’re going to lightly couple the quantum dots so it doesn’t destroy our qubit. Lightly couple the quantum dot to it and then see, okay, this quantum dot has got this, you know,
John Koetsier (12:38)
That’s a very small difference.
Chetan Nayak (12:58)
little bit of the imprint of the nanowires. At this point, it’s like we’re taking a nice wine and we’re hoping for some just notes of blackberry or something at that point, right? And that we’ve got to be able to detect.
John Koetsier (13:11)
And ideally you want to do this a lot
of times quickly.
Chetan Nayak (13:15)
Right. So, but we, but we got to do it fast. Right. And we want, and we actually want a single shot measurement. So we want to be able to check once single shot and no even or odd. And it turns out it’s possible. So we built this device. This is a paper that’s coming out in nature on digital version on the 19th of February and then in print on the 20th. And what we showed is actually that little tiny imprint is
John Koetsier (13:22)
Mm-hmm.
Chetan Nayak (13:46)
is actually it’s not that tiny. Okay, it’s not just a little bit of bouquet on the wine, it turns out that it’s actually big, it’s a femtofarad, you know, in units of capacitance, which on the scale of these things is big, and enables us to measure this. It’s a noticeable perturbation on that circuit. We see the response of that RLC circuit change and
Voila, we can measure whether it’s an even or odd number of electrons. So that’s like number two is the second, we’ve got the ground floor, the foundation is we’ve a new state of matter. The first floor is, is what this paper that’s coming out in Nature is, we can actually use it and we can measure even or odd number of electrons in one of these wires. And we can measure it rapidly and we can measure it fast. Great, now we actually monitor this thing. Okay, and we monitor this state and actually,
John Koetsier (14:32)
Mm-hmm.
Chetan Nayak (14:41)
It doesn’t actually live forever if it’s even or odd. Okay, you knew there wasn’t going to be stable, infinitely stable, you know, right? Yeah, so, but but it lasts really long time. Actually, when we met where we measure, we monitor it, we do see that it flips, it jumps like every two milliseconds or so, which is actually a pretty long time. Okay, we can do these measurements in in microseconds, 10s of microseconds, and this is lasting milliseconds. And ultimately, we’re to do this measurement, we’re going to make it faster and do it in a microsecond.
John Koetsier (14:48)
Not magic. Yes.
Chetan Nayak (15:12)
But so, so that’s great. We understand that there are some flips. We can understand what causes them. They’re actually caused by, there are some of these excited unpaired electrons caused by infrared radiation that manages to get into our device. We shield it really well. And so we can keep these things, you know, at a manageable level, but some of it sneaks in. It does cause this. And so.
We do understand something about the error processes and the jumps and that kind of thing. So that’s the first floor. Or Europeans would call it the zero floor, but we call it the first floor in America. Yeah, exactly. So, okay, so now what have we done with that? So going beyond what’s in the Nature paper, because that was submitted for peer review a while back. It goes through the peer review process, which is great that it’s a rigorous and extensive process.
John Koetsier (15:49)
Yes, exactly. And programmers.
Chetan Nayak (16:05)
But what we’re sharing at this meeting in Santa Barbara next week at the Station Q meeting is what we’ve done since then because we want to share the very latest results. And so that part of the story is that we’ve built on that and we’ve taken two of these wires, OK, and two of these wires we can assemble into a qubit. That qubit has two wires. It’s got four of these Myron and zero modes. And what’s cool about that is we can do the measurement that we said before, which is measure one of the wires. We measure one of the wires. We monitor it.
we see some jumps. Okay, so that’s great. But we don’t just wanna see these random jumps caused by errors. We want to change the state of the qubit intentionally, right? And what you would normally do with the qubit is if its state was pointing in some direction, there’s something called the block sphere that is a way of geometrically describing the state of a qubit. If the qubit state were in this direction, which is let’s call it zero, you would drive it. You would drive the qubit.
with microwaves or something like that, and you would rotate it over to, let’s say, the equator, right? We don’t do that. What we actually do is we do a measurement, and we do a measurement not on a single wire, we do a measurement that bridges across two wires. So there’s myron and zero modes at the ends of two of the wires. We do a measurement that goes around like this. So that measurement that bridges in two wires actually discreetly puts you into onto the equator.
Okay. So rather than doing this kind of rotation, which actually is sort of an analog sort of thing, we actually just discreetly do a measurement and put it onto the equator. That measurement is a measurement that involves two Myrana zero modes that bridge across. You’ll get two different answers, okay? So you might actually go this way or you might go this way, either way on the equator. We just need to record and then…
John Koetsier (17:34)
Interesting.
Mm-hmm.
Chetan Nayak (18:01)
correct later in processing which way it went as long as we put it along the equator. And so now we’ve shown, okay, we have a qubit in which we can actually do something. It’s a qubit with four minor on a zero modes. And now we see, you know, the first pieces of our architecture, which is we’ve got to be able to manipulate, we are manipulating quantum information through measurements by different kinds of measurements, okay, which is a unique part of our architecture.
John Koetsier (18:28)
Yep.
Chetan Nayak (18:31)
that the standard qubit architecture involves doing a lot of unitary gates to manipulate the quantum information and then do measurements at the end or maybe do measurements just to diagnose errors which you’ll then correct. Our architecture is based around almost exclusively is based very highly on doing measurements. That’s more than 99 % of what we do is measurements. Single qubit measurements, we’re gonna be doing two qubit measurements and so on.
John Koetsier (18:43)
Mm-hmm. Mm-hmm.
Chetan Nayak (18:58)
It’s all based on measurements, because measurements also do changes to states in Hilbert space, and particularly for these like 90 degree type of rotations, including in higher dimensional spaces, measurements are really effective ways of doing that in a very discrete way. And that’s what you need for quantum error correction. That’s what we natively do with this kind of architecture. So that’s like the second floor in this story is we’ve got a topological qubit that we…
John Koetsier (19:20)
Very cool.
Chetan Nayak (19:27)
on which we can do two different types of measurements.
John Koetsier (19:30)
Excellent.
Second floor in Europe or second floor in the States? Excellent. What’s the third floor?
Chetan Nayak (19:34)
I’m giving the American version. Okay. No subtitles and no dubbing.
And I’m doing it at 1.5x for free. Okay, so cool. And that is, that’s not in the Nature paper, that’s unpublished work, but we’ve shared that with a number of experts in the field. And this is gonna be presented at this meeting next week, okay? With the audience of people from across the field and across the industry. Great, now we have a roadmap.
John Koetsier (19:48)
All good, all good. So what’s on the next floor?
Chetan Nayak (20:11)
that builds on that, that takes that basic qubit and makes multi-qubit structures and builds out towards a useful quantum processor. We are putting on the archive next week a roadmap paper that shows the roadmap of what we’re doing of building up the test. The next step is an eight-qubit chip. The step beyond that is a larger processor heading out towards a fault-tolerant quantum computer. So that means one that’s robust and has very, in which
the entire system has a much lower failure rate or error rate than the individual pieces. So you’ve kind of bootstrapped to something that has error rates less than one in a million. that is the basis of that roadmap and all of the engineering and architecture underneath it is the basis of contract we just signed with DARPA. So you may have seen that last week we signed this contract with DARPA. We were one of the…
first companies to go into phase three of the US2QC program, is part of the quantum benchmarking initiative. We’ve entered into that contract with DARPA. so that roadmap and all the detailed engineering underneath it, the roadmap is pretty high level on the paper and kind of focusing on the physics aspects, but all the engineering and architecture underneath it is the basis of our interaction with DARPA. So that’s kind of the next piece of the story is that we were building on this with a roadmap that
heads out towards building a useful quantum computer. And then the last, what’s sitting up on the roof of this building is we’ve got a, we actually have fabricated one of those eight qubit processors. And we demonstrated that we can fabricate it, that a lot of the engineering challenges, the routing and packaging that you need to be able to make something like this, we’re able to demonstrate.
that’s the next step in our journey. So I would say, you know, what I think we’re showing is that we are accelerating progress. I think that this development of this new qubit is one of those, you know, moments of acceleration. And I think DARPA feels the same way, hence their contract. That is based around the belief on their part and ours.
that we’ll have a fault-tolerant quantum computer, a real fault-tolerant quantum computer in years, not decades. So we’ll have it this decade. And once we have that, that’s the thing we’re gonna build on to get up to utility scale.
John Koetsier (22:43)
Amazing, amazing. Well, you can take a breath. So four pieces of news, very, very cool. I’m going to try and summarize what I think they are. A new state of matter, which is cool. A way to measure it via quantum dot that’s non-destructive, which is also super cool. A way to load information into it and process it essentially by measuring it. Again, very cool. And also a way to make multi-qubit structures that you’ve actually
proven out by making an 8 qubit processor. So that’s pretty impressive news. That’s incredibly impressive news. And it’s a totally different architecture than any other quantum computer, which is also incredible. It’s another path to a similar destination, but on a totally different architecture, like VHS and beta or something like that. Super interesting. What surrounds us? How big are we talking about? So the typical quantum computer that I’m used to.
maybe a fridge size thing in the massive room, all these refrigerators around it. How, what’s the size and scale scope of what you’re looking at here?
Chetan Nayak (23:48)
Yeah.
roadmap is based on the assumption or really the strongly held belief more than assumption that we need this thing to be pretty small okay and the reason is if you don’t have a fairly small quantum processor chip and mean like a million qubit chip as an example if you don’t have a fairly small quantum processing chip then you are not going to put it in a single fridge you are necessarily networking multiple fridges
And then you’re not only solving the problem of building a quantum processor scale, you also have to solve the quantum networking challenge. You have to solve two super difficult problems. Okay, so we want to solve one difficult problem. What’s that? Yeah. This is easy solution, This is the easy solution. And so we don’t want to have to try to solve two fundamentally very difficult problems.
John Koetsier (24:27)
So this was the easy solution is what you’re saying. So this was the easy solution.
Chetan Nayak (24:46)
our roadmap is based around having a single module quantum computer, number one. Number two, a secondary reason aside from the difficulty of the quantum networking challenge is we’re not gonna want just one quantum computer, we’re gonna want a thousand. And the reason is we’re gonna use the quantum computer to let’s say find the ground state of some molecule that maybe is important catalyst for carbon capture, nitrogen fixation to make better fertilizers or non-PFAS coolants.
So there are number of chemistry problems and there’s a number of materials problems that are really near and dear to my heart because I started out by studying superconductivity, high temperature superconductors. We want to be able to discover better superconductors that have more practical applications. We can’t simulate any of those things right now. We can do some approximations and they give us as much insight as the approximation allows. A quantum computer is actually the processor for doing that, right?
It is the processor that is designed to do those kinds of things. like, okay, GPUs are great at matrix multiplication. It’s why people love them in AI and also love them for Bitcoin mining and various other things. But you know, what is a quantum processor good for? is for simulating quantum systems. I mean, it can be good for other things too, like potentially breaking RSA, but it’s good for, you know, simulating quantum systems. That’s what excites me the most. You don’t want to just simulate one molecule and then take a month.
John Koetsier (26:04)
Yes.
Chetan Nayak (26:13)
find out it wasn’t the one you needed and then simulate another one, you’re gonna, you wanna do a high throughput thing where you’re simulating something parallel, thousand different molecules, and hopefully one of them has the properties that you want, or hopefully a few of them, and then you will in the lab go and test them and see how well they can, et cetera, et cetera. Right now we have to do all that screening, almost all of it just experimentally, which of course is costly, slow, you know, it’s energy consuming and time consuming. So we wanna be able to simulate all those things to find
John Koetsier (26:29)
Mm-hmm.
Chetan Nayak (26:43)
to find promising materials and promising chemicals and compounds, we need to have a thousand of these. So if a single million cubic quantum computer is an entire data center, well, it’s some huge machine networked between multiple fridges. Okay, then you need a thousand of those data centers, then it looks really hard.
John Koetsier (27:08)
Yeah, we’ll just put them all,
we’ll buy Denmark and put them all in Denmark or something.
Chetan Nayak (27:13)
Okay, about that one. Okay, But so we want we want our we want our million cubic quantity to be in a single fridge, and we’re going to want 1000 of those. And so that was really a guiding principle behind this from pretty early on, you know, I’ve been on this journey, this is the longest running R &D program in Microsoft history, actually. So
John Koetsier (27:39)
How long?
Chetan Nayak (27:40)
19 years, 19 and a half years actually. We’re coming up on a really long journey here. Yeah.
John Koetsier (27:42)
19 years.
so much
faith and so much investment over so many different, you know, leaders. That’s almost unbelievable.
Chetan Nayak (27:53)
is true. It’s true.
We’ve had multiple CEOs, we’ve had multiple chains of command and reporting structures. I’ve had multiple bosses over that time. But it actually is pretty unique. If you think about Microsoft as a company, when I joined, which is 2005, when we started the quantum program, it was one of the five largest companies in the world by market cap.
at that time. 10 years later in 2015, by then we had a new CEO. It was still one of the five largest companies in the world by market cap. And here we are, almost another 10 years later, again, one of the five largest companies in the world by market cap. I don’t know if in the history of capitalism that’s happened very often. And through that entire time, the company has been willing to take big bets, right? mean, we do have to be honest, this is…
It’s high risk, high reward stuff because not only is quantum computing necessarily, you know, high risk, reward. We decided to do it in this way where we had, where we were insisting on reinventing the wheel many times over. You know, I sometimes use this analogy that, you know, we decided not to build a computer with vacuum tubes. And we decided we’re going to build it with transistors, right? But to do transistors, what people don’t always appreciate is, well,
John Koetsier (29:14)
Thanks.
Chetan Nayak (29:21)
people were just barely understanding what a semiconductor was in the 1930s. mean, they knew there were metals, knew there were insulators, and semiconductors kind of somewhere in between. And you would kind of tune it back and forth, which eventually people realized, that’s good for a switch. you were barely understanding what semiconductors were. it took a lot of understanding. People didn’t even know about the electron early in the 20th century. And Millikan did his experiment, the oral drop experiment. We measure the charge of a single electron.
So, you know, had to discover the electron, you had to discover and understand semiconductors, you had to invent the transistor, then the integrated circuit and so on. We had to, in some sense, go through a, you know, slightly sped up or hopefully very sped up version of that path of discovery because we had to invent a new kind of material, topological superconductors. We had to essentially discover a new kind of particle. Like for us, it’s not the electrons, it’s…
John Koetsier (30:12)
Yeah.
Chetan Nayak (30:20)
these Majorana zero modes are the basic particles for us. had to figure out how to do something with, just as with semiconductors, people had to figure out how am gonna gate these things to actually take them between some metal and insulator effectively is what you’re doing. We had to do that by coupling quantum dots and figure out how to read out between zero and one, even an odd number of electrons. We had to.
make our transistor, which is a qubit involving two of these topological wires. And then we had to make multi-qubit devices, which, you know, seeing in our eight qubit chip and you see what our roadmap has to do. So we really were reinventing the wheel in a lot of ways. And you know, that is part of the reason for the nature of the journey. Now people could say like, people could easily say sort of.
John Koetsier (31:06)
Yeah. Yeah. You kind of went from, you kind of went from
a laden jar to a lithium ion battery in 19 years. Wow. And I guess you were 12 when you started at Microsoft is, I guess you were 12 when you started at Microsoft.
Chetan Nayak (31:15)
already putting it. So, so although it’s a long journey. What’s that?
Very funny. Thanks. I appreciate that. But I’ll say this, you know, over the 19 years, I have been learning something all like I would have gotten discouraged if I didn’t think I was learning something right. And I think as a team, we’ve learned things.
I think we are operating much more effectively as a team. This is not an individual thing. So, you know, I’m talking to you about this work, but actually, if you look at our nature paper, there were 160 some odd co-authors on it. And, you know, which I think sometimes people have said, oh, that’s a lot of co-authors. I can tell you not one of those was a courtesy. Like we didn’t put co-authors on, oh, let’s be nice to this person. These are all people who I know and who contributed to this work.
John Koetsier (31:54)
Wow.
Chetan Nayak (32:09)
because of the fact that we had to reinvent the wheel many times over, because we have so much of an effort on creating these materials, because we had to invent new ways of measuring materials, because we had to design these devices that take advantage of a new state of matter. And so there are so many things involved, but it is a big team and it really has to function as a team. So I think we’ve gone through an evolution of learning a lot about science and engineering.
but also learning a lot about how to function as a team and how to be a team accomplishing something like this. And I’ve seen the progress over the years. We have gotten a lot smarter and it’s clear that we are going through in a phase of acceleration right now. But even through all of the years of one step forward, sorry, one step back and two steps forward, even through all these years, I could see the direction of progress and that’s what kept me going.
John Koetsier (32:58)
Thanks.
That’s pretty cool. Um, and, and, and I want to get to some of the personal feelings, uh, as well, maybe a couple, a couple of things, uh, before that. So you’re, you’ve manufactured an eight qubit processor now. Um, you’re going for a million. said, uh, that’s pretty impressive. You want to have that in one fridge. That’s pretty cool. I mean, it’s funny cause you’re talking about, know, we don’t just want one of these. I was thinking about IBM back in the day when they are starting to build those massive.
room scale computers and the CEO at the time estimated there’s a worldwide market for five. And he probably wasn’t wrong. I mean, we look at them and we think, ha ha, what a silly, you know, what an idiot, right? He was right at the time in terms of that size and that scale. Now go, go forward, get out your crystal ball, go forward five, 10 years, whatever you figure out there’s remaining challenges, right? There’s many remaining challenges. There’s, there’s not just mass production, but all the other.
Chetan Nayak (33:45)
Yeah
John Koetsier (34:03)
road speed bumps that you’re going to hit and all that stuff. But assume you solve all that. What do you think that looks like in the world? How many quantum computers do we have? Is it the same as classical computers? Is it a fraction? Will it ever be in my hand or will it always be somewhere that I access in the cloud?
Chetan Nayak (34:21)
Yeah, think let me take it in reverse order. I think it’s always gonna be something you access in the cloud because quantum computers really, they all inevitably involve pretty extreme conditions. I talked about ours, you know, 50 millikelvin temperature, two Tesla, one to two Tesla magnetic field. So these are extreme conditions. So these are gonna be in the cloud. And it’s probably not a consumer device, right? One thing I want to emphasize is quantum computers are not gonna be good for everything. There are things
John Koetsier (34:38)
Mm-hmm.
Chetan Nayak (34:50)
that classical computers today are very good at, that quantum computers are always going to be terrible at. Just like there are certain things that quantum computers, when they get to that scale, will be great at, classical computers will never be able to do. So we have to think, I think we have to think about quantum computers as a very particular processor. Just like we have CPUs and GPUs and other things, have ASICs, applications, specific integrated servers. Quantum computer is definitely a specialized processor.
for certain applications. We don’t know the full extent of what those applications are and we won’t until we start playing around with them. But we know that they’re really good for simulating quantum systems. And there are a lot of problems that involve simulating quantum systems because, know, chemistry and materials enters into 96 % of all manufactured goods. So it is a big part of what we do. So I think that while we can today glimpse some of the things for which quantum computers will be important.
And for some of those problems, you are going to want a million, a thousand machines, each of a million qubits. I have no doubt that there are going to be applications that we haven’t even discovered yet. I think that that may be for search within a particular space and you will want several of those kinds of, you know, thousand computer data centers. So I do think that quantum computers, there will be a day
when quantum computers are widely used for certain things.
John Koetsier (36:22)
Mm-hmm. Mm-hmm. Mm-hmm. Cool. And then maybe any guesstimate right now as to when you will get to this 1 million qubit quantum computer.
Chetan Nayak (36:37)
Yeah, so what I can say is we have a contract with DARPA to build a fault tolerant quantum computer. So that’s the first point at which you can say, okay, this is a computer that with error corrected fault tolerant qubits, you’re doing real and useful computations. That is low tens like 10 to 20 logical qubits where each logical qubit has this one in a million or better error rate. Okay, that’s in the thousands of qubits. Okay, a few thousand qubits. But that is the point at which you
you do understand a lot of the underlying physics and engineering and you are ready at that point to scale up to something like a million. So I think the fault tolerant, useful fault tolerant quantum computer we’re gonna have this decade. I think that’s years, not decades away. And I think that the utility scale quantum computing building on that at that point, obviously it’s gonna depend on resources investment and so on, but that is the point at which I think you have a line of sight.
And I do think that those things, once you see line of sight on things, I think that’s when you’re talking years, not decades.
John Koetsier (37:43)
Chayden, this is kind of crazy. This is kind of amazing because quantum computers for so long have been over the horizon, hard to predict. And you’re saying sometime this decade and it’s already 2025, just getting used to it not being 2024. And so that’s not too much more time. We already have these massive changes coming into our lives from technology in terms of AI.
massively changing how we work, how we communicate, what we do, what we need to know, what we don’t need to know. Other thing, what work can be done by humans, what work can be done by machines. We have this looming, there’s probably a hundred companies globally right now that are building humanoid robots, right? And that’s not far away. There’s two of them that have actual paying jobs right now. Two of those companies that have them actually working in factories and logistics centers. Now you’re talking quantum computing is, it’s on the horizon.
It’s visible and that’s going to vastly. lot of things in science and engineering materials, as he talked about other things we don’t even know, don’t even understand that we’ll apply them to. there’s a lot of change coming.
Chetan Nayak (38:52)
There is a lot of change coming and it’s interesting. you know, I think it’s hard to predict exactly when we’ll have artificial generalized intelligence. We’ve seen a lot of improvements in AI, but that’s one of those things that people have looked at over the horizon and said, you know, maybe it’s not so far away. Obviously there’s a lot of challenges, but, you know, I think we maybe are seeing some glimmers that that’s not so, that’s not, not, you know, that’s not, that’s not just a subject of science fantasy, or science fiction, I should say, or fantasy.
John Koetsier (39:20)
Mm-hmm. Mm-hmm. Mm-hmm.
Mm-hmm.
Chetan Nayak (39:22)
I would
say, you know, and this is a little bit, but that’s both really cool, exciting and also scary, right? Because like artificial intelligence is one of those things, it’s a technology that’s destabilizing and could fall into the wrong hands, could get out of hand, could get uncontrolled, just like nuclear weapons, right? And the truth is we have been extremely, I don’t know, lucky, fortunate, wise enough to not, you know, have a nuclear war.
John Koetsier (39:29)
Yep.
Chetan Nayak (39:50)
and something that threatens civilization over the, you know, since the first bombing bombs in Hiroshima and Nagasaki, you know, but it does worry me, you know, because I don’t think we can, we can assume that it’s never going to happen. So that’s one of those technological things that that’s caused for concern, but it has been also a great opportunity because nuclear power is one of the way, you know, these data centers are all extremely power hungry, right? And we are talking about nuclear power plants. So
John Koetsier (39:59)
Yeah. Yeah.
Chetan Nayak (40:19)
So I think, know, Brad Smith wrote this book, Tool or Weapon, you know, a lot of these technologies give us a lot of reason for optimism and a lot of reason for concern. And we have to be attentive to both. And the third thing I was gonna, sorry, I’m just gonna keep going. You know, with CRISPR Cas9, there is this ability to potentially edit genes, right? And maybe change what it is to be human in a fundamental way by editing genes. Again.
John Koetsier (40:38)
Mm-hmm.
Chetan Nayak (40:46)
That’s not something that’s going to happen today. Maybe it’s on the horizon.
John Koetsier (40:52)
And maybe quantum will be impacting
that and maybe quantum will be impacting that as well as the nuclear stuff you’re talking about. still have to get rid of that nuclear waste. That’s a material science problem in a lot of ways. And maybe, maybe quantum will help with that as well. there’s a huge, there’s huge opportunity as well as some challenges. I swear, I’m sure Microsoft PR didn’t think we’d get onto those topics, but Hey, last thing I want to ask you and go ahead.
Chetan Nayak (41:15)
You know, but I’m but I was actually
going in the direction that you’re going in. I love you brought that yourself, which is if I look at all those challenges and I think, you know, things having to do with nuclear reactions and property of nuclei, maybe something that quantum computing could actually have an impact on nuclear physics. And when we think about power generation and the amazing power costs associated with training these big AI models, these data centers, well, that’s when I think about better coolants, about super about higher temperature superconductors.
for conduction electricity without loss, better batteries for storing that energy. So I look at a lot of these things and I say, these are things that quantum computers could be useful by understanding fundamental materials. And then as you mentioned, CRISPR-Cas9 and the genes and where there are…
John Koetsier (41:53)
Mm-hmm. Mm-hmm.
Chetan Nayak (42:10)
new chemicals and new reactions to be discovered that might have an impact on how we look at the fundamentals of life. That all of these things that are tied together at some point rely on the quantum nature of the universe we live in. When you really push them down to Feroz’s first principles and you say, why is this? And then you say, okay, great. And then, no, really why? And you keep pushing them down.
they do lead us back to the basic physical laws and most of the constraints on these things end up coming down to basic physics of what you can and can’t do. And quantum mechanics is an inevitable part of that. So I think that quantum computing really can, quantum computers can be a tool that help us.
John Koetsier (42:57)
Cool. Last question. You’ve not only do something that very few people can do, you also have had the privilege of having an experience that very few of us will have, which is working on a project that is high risk, high reward, uncertain ending for 19 years. But let’s face it, it’s longer than that because you went to school and you studied and all these other things before you even came into this position here with Microsoft.
You’ve been working on this project for 19 years and now you’re seeing fruit. How’s that feel?
Chetan Nayak (43:34)
It’s really, well, first of all, it’s incredibly exciting to, you know, there have been times when I’ve seen some of the data from these devices matching the predictions and definitely spine tingling. There definitely been a lot of moments where I got chills. As I said, it’s been a long journey, but it’s been a fun journey because I felt like me personally and us as a team have been learning and growing better and have been making progress this whole time.
even if not always in a straight line. And I look at what we’ve accomplished and I look at the opportunities we have ahead of us and I’m extremely excited.
John Koetsier (44:12)
Wonderful. I guess that’s the old adage, right? Find a job you love and you won’t work a day in your life. Thank you so much, Chaitin. I really appreciate your time.
Chetan Nayak (44:19)
Thank you, John. It’s been my pleasure.