Thanks, everyone, for joining here. I'm honored to have the CEO Subodh from Rigetti Computing. I'd love if you could just kind of start with a quick introduction of yourself and maybe kind of a background of the company.
Sure. Thanks for having us, Brian.
Of course. Yep.
I'm the CEO for Rigetti Computing. I've been with the company for a little more than two years now. Before that, my background, basically engineering by training. I did my PhD at MIT in semiconductors, then worked in semi-industry for many years, almost three decades now. Last 10 years, I was the CEO of a publicly traded company in the Twin Cities called CyberOptics. We were doing inspection and metrology for semi-industry and electronics manufacturing. That company got bought out about 2.5 years ago, and then I took this job at Rigetti. Quick background about Rigetti. Rigetti was founded in 2013 by Dr. Chad Rigetti, and that's where the name comes from. We are a full-stack quantum computing development company. That means we do everything from designing the chip to building the chip. We have a fab in Fremont, California.
Our head office is in Berkeley, California. We do the entire stack, the hardware and the software. That's how we started. Right now, we are in our 11th year. A lot of IP around this area, obviously. We have about 230 patents. We are what we call superconducting gate-based quantum computing development company. I'm sure we'll get into that.
Yeah, perfect. Yeah. In fact, that's my next question. If you can discuss the differences between supercomputing or superconducting, trapped ions, photonics, can you talk about the strengths and weaknesses of both?
For sure. Before we dive into the details, a couple of minutes about quantum computing.
Yeah, please. Yes, absolutely.
Quantum computing is an exciting area right now with significant potential long term. Fundamentally, that's because instead of getting constrained by zeros and ones with the classical digital technology we all use, which is CMOS, basically, we can have multiple states at the same time. Also, we can entangle this, what we call qubits, at the same time. When you have n bits with your classical digital technology, your computational power goes linearly by two multiplied by n. Your energy consumption also goes linearly two multiplied by n. When you have n qubits in quantum, we call them qubits, your computational power goes two raised to n, and your energy consumption also gets equally reduced. When n is small, it does not make a big difference. When n becomes 50 or 100, two multiplied by 100 is 200.
Two raised to 100 is practically infinity. That's where the power of quantum computing starts. When you have many qubits and you entangle them, you get this tremendous computing power. I mean, we are not talking 2x or 4x better than classical computing. We are talking a million or billion times faster than classical computing, taking on problems that you cannot even fathom right now, like protein unfolding problem or weather forecasting type problems, where there are millions of variables all interacting simultaneously. Phenomenal potential to solve problems that cannot be solved with classical computing today. Also consume a lot less energy because a 100-qubit chip is practically the power of a supercomputer, which would be the size of this room, if not multiplied by two. We are consuming a lot less energy than a classical supercomputer or an HPC stack.
It is both higher computing power as well as lower energy consumption. That is what the excitement is about. Obviously, one of the reasons governments and everyone is getting involved is because of its capability to encrypt and decrypt. Now, what it means is right now, all of us are using AES encryption, 128-bit, 256-bit. That is how the world is running right now. It works because even with a supercomputer, it would take you 1,000 years to break a single encryption key. With quantum computers, once you have capable quantum computers, you can break a 128-bit or even 256-bit encryption key in a few seconds.
That's crazy.
Unfortunately, that means the negative use for quantum computing may happen before the positive use cases because a rogue company or country, if they manage to get a quantum computer, I mean, we can laugh about it, but the stakes are pretty high, right?
Right.
I mean, if somebody got hold of a capable quantum computer, it would make the current missile-based warfare look like medieval affairs. I mean, you can literally bring a whole country to its knees by shutting down everything. You also mean that to prevent that from happening, you need to encrypt information with quantum computing, right? There are both cases. Anyway, it's exciting. A lot of effort is going on right now to advance quantum computing. Now, within quantum computing, and that gets to your question, there are many ways we can do quantum computing. Fundamentally, quantum computing is we are getting state of the matter in quantum state, and we are doing the computation while it is in quantum state. There are many different ways of bringing things in the quantum state and doing computation with that.
Broadly, there are what we call gate-based quantum computing and non-gate-based approaches. Gate-based is very similar to what we do right now with our conventional technologies. We have 0, 1, and we have gates in between that open the gates. Quantum computing works the same way. When we do gate-based, we have different states between 0 and 1, but we can transfer the information from one side to the other side exactly the same way as we do today. Most of us, including Rigetti, we are in the gate-based camp. There are a few non-gate approaches, like entropy-based approach or annealing-based approach. There is a company called D-Wave that is doing annealing. There is a company called Quantum Computing Inc., which is doing entropy-based approach, a little different than what we traditionally do in the gate-based. Most of the rest of us are doing what we call gate-based approaches.
Within gate-based approaches, there are different modalities of quantum computing. We do what is called superconducting quantum computing, where we are taking basically a chip, taking it at extremely cold temperatures where we are getting superconducting effects. That is when we create the quantum effect and doing the computation. In the superconducting camp, you have us. You have big companies like IBM, Google, Amazon, Microsoft, the government of China, the Chinese Academy of Sciences, and many other smaller companies too, IQM, Finnish company, Dutch company called QuantWare. In Japan, you have Fujitsu, Toshiba. A lot of effort, a lot of big tech and small companies in the superconducting gate-based quantum computing. There are other modalities. Trapped ion is one of them, where there is a company in Maryland called IonQ. There is a private company in Colorado called Quantinuum and a few other companies like that in the U.K., Oxford Ionics.
There's photonics, where you are using laser-based approaches to create quantum effects. There's spin approaches, where you have companies like Intel, GlobalFoundries are doing some work in that area. There are a few other modalities. Frankly, there are pros and cons for each modality. There's no clear reason why one modality is significantly better than another. We like superconducting along with many big tech companies primarily for two reasons. We have the benefit of what we call scalability and gate speeds. By scalability, what we mean is we are essentially using a chip. Once we perfect the chip, we can leverage five decades of semiconductor industry experience. We know how to shrink it. We know how to multiply it using all the technologies that have been mastered by the semiconductor industry.
We all feel confident about the scaling part of quantum computing using superconducting approaches because we are using chip technology. Also, gate speed. We are reliant on electron speeds, essentially, exactly like our CMOS technology is. Our gate speeds, our clock speeds, if you will, are in the tens of nanoseconds range. Our downside, the con for superconducting gate approach is what we call fidelity. Because we are dealing with engineered devices, just like CMOS chips, typically, we get noise when we are there's always a noise component. That leads to decoherence effect. We lose the quantum effect after tens of microseconds. Ideally, we want it to be as long as possible. That's been the challenge for superconducting gate-based quantum computing approaches, the fidelity.
On the flip side, trapped ion, pure atom, those kinds of modalities, because they are dealing with pure ions or pure atoms, their fidelity is intrinsically better. They suffer with scalability. It's not easy for them to go up to hundreds and thousands of qubits and certainly hundreds of thousands of qubits. Also gate speed. Because they are dealing with ions and atoms, they are significantly bulkier than electrons. Their gate speeds are 10,000x slower than where we are, making it impractical to do real-life computing kind of stuff. There are pros and cons. Every modality has its slightly different challenges. Overall, if you look at where the money is coming in right now from an investment standpoint, just look at all the names that I mentioned, IBM, Google, Amazon, Microsoft, government of China.
Almost we are pretty sure the superconducting gate-based modality is probably one of the most likeliest to win just because of the amount of R&D that's going in, effort going in, the benefits of scalability and gate speed. Until about a year ago or two years ago, fidelity was a challenge. If you look at one of the reasons the area is getting a lot of excitement right now is with all the announcements that have been happening in the last few months. We've announced some pretty impressive results. Google Willow chip made a big splash when Google announced in early December. Those were really good results. IBM had some very important announcements. Very recently, about a month ago, both Amazon and Microsoft made some announcements.
Essentially, when you look at all the announcements from various companies in the superconducting camp, including us, what we are saying is we have caught up in fidelity compared to the other modalities. If anything, we have gone ahead of them. The traditional advantages of scalability and gate speed still continue, but we have gone ahead in fidelity. Specifically, Google announced 99.7% fidelity. We announced 99.5% fidelity. Frankly, those are equal to, if not better than, the other modalities like trapped ion and pure atoms. That is what is getting a lot of enthusiasm generated right now in the field.
Interesting. Can you talk a little bit more just about qubits, logical qubits, and just kind of your scalability roadmap?
Sure. Right now, Rigetti, we have deployed our systems on the cloud on AWS Braket as well as Microsoft Azure. Our state-of-the-art system is an 84-qubit system. Our fidelity is 99.5%. As I mentioned, that's a median 2-qubit gate fidelity, which is when the two qubits are interacting. Single-qubit gate fidelity is no problem. We are all at 99.9%, the capable companies. It is a 2-qubit gate fidelity that we monitor most closely. That is at 99.5%. Our gate speeds are in the 70-nanosecond range. That is where we are today. If you look at IBM and Google's numbers, they are kind of comparable regime right now. All of us are in the same general ballpark. Where we go from here, what we have announced is before the end of 2025, we want to bump up to 100-plus qubits.
Importantly, what we have announced is we want to do what we call the chiplet approach this year. That means we want to tile smaller chips together to build a more than 100-qubit quantum computer. The reason we are doing that right now, all of us, us, IBM, Google, other companies too, and even the government of China, we are all doing a monolithic chip approach, which is a single chip right now. That works, obviously. The challenge we see is that as we think about getting to thousands of qubits, we see a potential big problem ahead in the roadmap. We will not be able to control the uniformity correctly enough across the entire surface of the chip area. Same reason why CMOS right now uses chiplets for the advanced technologies.
If you take most of the CMOS applications at 3nm and 5nm, the state-of-the-art GPU technology, the state-of-the-art iPhone, if you will, they are all using chiplets, or most of them are using chiplets for the same reasons. When TSMC or Samsung, those kinds of companies are trying to build large chips, monolithic chips, there's challenges with uniformity. When you are trying to control 3nm-type dimensions across the length of centimeters, it becomes extremely difficult. That is what we are also finding in the quantum world. The same reasons why CMOS is deploying chiplets is why we are also planning on deploying chiplets. We proved that you can keep the quantum effects intact while you do chiplets. We want to prove it in scale right now.
This year, our plan is to get to more than 100-qubit with chiplets at 99.7% type fidelity with specific type of gates or 99.5% fidelity with generic gates and get the gate speeds below 50 nanoseconds. That's what our plan for this year is. Going forward, we'll continue to bump up both the qubit count as well as the fidelity number. The goal is to try to get to a few hundred to maybe like 1,000-qubit, more than 1,000-qubit in the next four to five years, more than 99.8%, closer to 99.9% fidelity, median 2-qubit gate fidelity in the same time period, gate speeds of less than 30 nanoseconds or even 20 nanoseconds, and start doing what we call real-time error correction. Right now, Google Willow was a phenomenal demonstration of error correction, but it was not a real-time error correction.
They couldn't keep up with the speed of computing. It was almost done offline, if you will. That works, but it's not practical from a compute standpoint. All of us, I'm sure we want to get to real-time error correction where the error correction is happening as we are computing. The goal is to get to that point in roughly the next four years or so, at least five years. That becomes a very important point for the whole industry. We call it quantum advantage. That's when we can demonstrate that quantum computers are superior compared to classical computers in terms of performance or cost. Right now, none of us are superior than classical computers. That's why you don't find quantum computers in data centers today. All of our applications are research applications. You asked a very important question, a logical qubit.
All the numbers I mentioned, 100-qubit, 1,000-qubit, those are all physical qubits. Right now, I mean, because the errors, 99.5% sounds impressive. Frankly, when you multiply 0.995 a thousand times, you get a very small number. That is what is going on in real computing. Pretty much, that is not a good enough fidelity. At 95, it was really awful. At 99.5, you start getting real meaningful numbers. Your CMOS technology, by the way, right now is usually between four nines and six nines. What that means is it is 99.99% or 99.99999%. The rest of it, we address with error correction or with redundancy. That is how the current computers work. We want to get to the same point. We want to get to this 99, at least three nines, and start doing error correction, real-time error correction.
We are pretty confident we'll get there. That's when the discussion about logical qubits matters. If you ask me what is our number of logical qubits on our current 100-qubit system, I would say it's pathetically low. It's not even worth discussing right now how many logical qubits we have with a physical 100-qubit system. By the way, that's the same for everyone. If somebody is telling you that they have tens of logical qubits, they are either doing something wrong or they're just flat-out lying. None of us are talking logical qubits because we are not quite there yet. When we get to, let's say, four years or five years from now, we are at 1,000-qubit at 99.9%. We will be talking roughly 50 to 100 logical qubits at that time.
That becomes important because if you look at the state-of-the-art supercomputer today, it's the Frontier supercomputer that sits in Oak Ridge National Lab. That's the world's biggest supercomputer today. That would be equivalent to 50 logical qubits. The power of Frontier is two raised to 50, roughly. That would mean that if we got 50 logical qubits, we could rival the state-of-the-art supercomputer today. That is why we want to get to the thousands and tens of thousands of qubits so we can get to once you are at 1,000-qubit or logical qubits, then you have unleashed a real power of quantum computing. That is where we want to go.
Yeah. OK. All right. To me, there's so much that needs to be done yet to get to quantum. I think about the software, the memory, the storage, the security, compilers. Can you just kind of talk about where we are on that aspect of quantum? You mentioned that you guys are doing some software. Expand on that, if you could.
Overall, we are still very much in the R&D phase of quantum computing. We still have to perfect quantum computers before we can take on real-life applications. Most of the work, we have built systems. We have delivered it to customers. If you go to the U.S. government, DOE labs, you will find our quantum computers. You go to DoD labs, you will find our quantum computers. Same with the U.K. government. They are working. The good part is they are working. People are doing research work on it. All that is good. We cannot demonstrate superiority over classical computing today. We are very much in the research phase. A lot of companies are focusing on the various different aspects. We are a full-stack company, but we are a small company. We are a startup company.
Our focus is increasingly becoming more and more on the chip design and chip fabrication. What we have taken pride in is that we have designed an open modular interface approach so we can essentially incorporate third-party innovative solutions very relatively easily on stack. IBM and Google have opted to stay with a monolithic mainframe kind of an approach, which works. I mean, they are companies with deep pockets and take an effort to do that. Frankly, we think our open modular approach is a better approach, more creative approach in the long term. Good examples of that is there's a company in the U.K. called Riverlane in Cambridge, U.K. They are about the same size as us. They do excellent error correction software. When we saw their error correction is really top-notch, we started integrating their software in our stack.
We could do that because of our open modular approach. Same with NVIDIA. When NVIDIA started working on the quantum version of CUDA, we could relatively easily integrate NVIDIA CUDA-Q in our stack, whereas IBM and Google, they could do it if they wanted to. They have chosen to stay completely vertical right now. We think that's the right way to allow others to develop creative solutions with different parts of the stack. Very recently, just last week, we announced a strategic partnership with a Taiwanese company called Quanta Computer.
Yep. I was going to ask you that.
They are the world's number one manufacturer of CPU, GPU servers right now. They are NVIDIA's closest partners. We believe that their technology marries very well with our technology so they can start working on control systems because a control system essentially is a CPU, GPU server with some RF circuitry in it. Instead of trying to do everything ourselves, we are slowly but surely incorporating creative solutions from outside that accelerate our time to market, allow us to go faster. Applications is the biggest one. That is where we are really going to be counting on more customers. We are going to enable more and more applications. We are doing a few applications work ourselves. We work with customers like Moody's and HSBC and ADIA in Abu Dhabi and a few other customers like that.
Frankly, how many software engineers can you invest in to keep coming up with applications, particularly when you are in the R&D stage? We are going to be dependent more on customers driving those applications. Our job is to make sure the hardware is ready. We demonstrate what you can do with the hardware, the few applications, and then allow customers to implement those applications. Are there some questions?
Oh, go ahead, Justin, please.
For the U.S., you said you're deployed, I guess, on AWS. What are the common use cases that people are using your stuff for currently in the environment? How much does it cost? Like electricity power, how much are they being charged?
Yeah. AWS, I mean, the rates are well publicized on their website. If you just go to AWS Braket, you will see Rigetti Computing right at the top over there. It's like the first few minutes or hours are free. After that, it's per time, second or minute, depending on how much you want to buy. It's relatively cheap. I mean, you don't need that much time to do your problem set.
If you can crack encryption in a matter of seconds, you don't need much time.
We are still far away from that. That is all public domain info. You can see it very easily. Regarding the use cases, I mean, we can see the number of customers who are accessing our quantum computer through Braket, for instance. We do not see what exactly they are doing. That is the right thing. We should not be seeing what they are exactly doing. We can see the workload that is being put on our quantum computer, what we call the depth of the circuit that they are deploying, how many qubits they are using. We can see that there are roughly about 1,000 customers that are accessing it through AWS Braket. We see how much time they are taking. It is not taking that much. I mean, right now, we are basically incorporating all of AWS and Azure customers on one quantum computer.
That's more than enough for that right now. Most of them seem to be for research applications. It seems like that's what AWS tells us. It's mostly academic institutions, government, national labs, some Fortune 500 companies. It's mostly the cutting edge, like the Johnson & Johnson, trying to see what it does for pharmaceutical modeling, those kinds of leading edge kind of customers. It's very much in the research domain right now.
Yep. Fair. Any other questions from the audience? They're also probably hedge funds, right, trying to crack codes and get better performance. Can you expand a little bit on Quanta? I mean, how do they pick you guys? I think there's an investment related to it too.
Yeah. As I mentioned, Quanta is a large company. They had almost $43 billion sales last year, growing very rapidly along with NVIDIA because they are NVIDIA's closest partner to the GPU ecosystem. Not too many people in the U.S. know about Quanta Computer because they are located in Taiwan. They're a classic ODM type company. They have been looking at, I mean, they have always taken pride in the next thing, what's the next big thing. They were one of the first ones to jump into cloud-based CPU and GPU servers 15 years ago. That's why they are in the position they are in today with NVIDIA and other partners. They concluded a couple of years ago that quantum computing is the next big thing after GPU. That's likely to happen in the next 5 to 10 years.
They have been scouting the landscape for what companies are most likely going to succeed, what technologies, and so on. They concluded that superconducting gate-based computing is the most likely one to win for all the reasons I mentioned earlier, scalability, gate speed, and now that the fidelity gap is closed. Within superconducting, they narrowed it to us because of our open modular format that jives in with their philosophy of things have to be open modular compared to the closed-end approaches that some other companies are doing. From our side, we knew that longer term, as volumes pick up and applications grow, we needed contract manufacturers. Right now, we are putting everything together in Berkeley, California, not exactly the cheapest manufacturing cost place in the world. We knew that we had to find the right partners as volumes grow.
Taiwan is a logical place. I mean, most of the CPU, GPU chips are coming from Taiwan right now, as we all know. It makes sense to partner with a Quanta Computer-like company. The transaction is a little more complex. Right now, at the end of last year, Rigetti, we finished the year with about $217 million cash and no debt. Quanta Computer is committing to investing $250 million on the non-QPU portion of the stack. The QPU portion is what we will keep intact, QPU being quantum processing unit. We'll focus on QPU design fabrication. Quanta will focus on more the commodity portion of the stack, which is the control system, the dilution refrigerator, the cables, the chassis, everything. The reason for that is, obviously, quantum computing is considered to be a national sensitive, critical technology.
We did not want to get all kinds of issues involved. We are very closely aligned with DOE and DoD labs here in the U.S. Some of our IP is actually shared with DOE labs in the U.S. We did not want to get quantum technology involved with foreign entities that would have created challenges. We kept it very clean and straightforward. They are going to invest $250 million on the non-QPU portion of the stack. We'll invest roughly a comparable amount on the QPU. In addition, and this is classic, the Asian company JSR, they are investing $35 million in Rigetti, I believe, for 3 million shares, so about $11.59 a share so that we get cash at the end of this month, assuming regulatory clearance. Effectively, we feel that we beefed up our position quite a bit from a cash standpoint.
From $217, we effectively have +$35, +$250. So effectively, we have gone from $217 to about $500 million cash now. That puts us in a very nice position to compete directly with companies like IBM.
Absolutely.
Even though they have deeper pockets, bigger teams, we have excellent technology. We have great IP. So we feel very good about the position we are in right now with our approach and with a company like Quanta supporting us.
Yep. Awesome. Congrats on that.
Thank you.
Let me ask, are there other meaningful partners outside of Quanta that are important like that?
We certainly will continue to look at other partnership approaches because we are open modular. Right now, Quanta, obviously, for control system, Riverlane for error correction. There are other parts of the stack too as we go along, particularly when we start talking about encryption, decryption, applications layer. Even the whole, what we call dilution refrigeration technology, it's a complex technology. That's what's cooling the chip down. Right now, we are relying on there are three or four commercial companies where we are essentially buying dilution refrigerators. It is our single biggest cost component right now when we are building a quantum computer, what we call the DR, dilution refrigeration technology. We are looking forward to other partnership potential opportunities where we can integrate creative solutions from those kinds of companies.
OK.
Yeah. More questions?
Yep. Go ahead, Jess.
With Quanta, when do they expect to have a provide an extra rack system relatively easily? They think that it would happen in 12 months, in 4 months. They're going to spend all this money to figure everything out, to buy the separate GPU. How long do you think they'll be ready? Like, oh, we'll have something that we can show you in 9 months itself, just a whole box of somebody made that would work very easily.
I mean, it's a good question. You're getting into the general timeline for when volumes will take up and when Quanta's.
Those are really good questions.
They understand we are very much in the R&D stage. It is very hard to predict exactly when volumes take off. Longer term, I think there are many studies done. We are talking 15 to 20 years from now when this quantum computing would be hundreds of billions of dollars of business. Everyone's convinced. You can look at reports from companies like McKinsey or BCG. All that stuff exists. Predicting the inflection points in the next few years is always very difficult when you're dealing with R&D kind of things. We understand that. I believe Quanta understands that as well. The way I'll put it is, take a look at the actual numbers. That's better, easier to predict when we will be at 500-qubit or 1,000-qubit at 99.8%.
When we come to this quantum advantage, which is roughly the four to five years from now, that's where we believe the real inflection points start happening. If you say, what is the volume pickup next year, I don't think that's a meaningful discussion at this point, given where we are in the trajectory of R&D. Certainly, four to five years from now, the volumes are expected to be a lot larger. That's really what we are looking at. That's where Quanta is looking at. They are very much looking at it as a long-term investment, just as we are.
Yep. You think it's three to five years or so before there is real commercialization of Quanta?
Yeah. I mean, our view is that you will start seeing quantum computers in data centers in five years or less. I know there's a lot of opinions out there. Frankly, there's a lot of overhyping done in some cases where some companies are saying it's going to be this year and next year. We think that's a little overhyped case situation. There are some other extreme cases, like the famous one being Jensen Huang saying it won't happen for another 15 to 30 years. That we think is a little bit of a pessimistic underhyped side. We are somewhere in the middle in that four or five-year timeline. Fairly consistent with what you heard from Google recently in their earnings call.
They also kind of said the same thing as what I'm saying right now, four to five years, very similar to what IBM is saying. Maybe that's because we are all using the superconducting gate approach. So our timelines are generally jiving with each other. I think those are realistic timelines. Anyone who's saying it's now and next year may be a bit over optimistic. Anyone saying it's 30 years is definitely very pessimistic.
Yep. Yep. I agree. All right. As I started doing research on kind of quantum, I mean, the first thing that popped up to me was national security, right, and the importance of the U.S. leading in this category. Can you just talk about where the U.S. is versus China?
The honest truth is none of us know. It's very tough to get information what's going on at the Chinese Academy of Sciences. At least about three years ago, they used to publish papers and file for patents. They have gone dark on us. We hardly see any papers or patents. Very surprisingly, we saw a paper just published last week, as a matter of fact, in Physical Review Letters, which is a very good peer-reviewed journal. That's our closest look at where they are. Now, Physical Review Letters is a very good journal. We want to trust the data and the people who reviewed the paper. At the same time, the data, when we look at it closely, comes very, very close to Google's data, almost to the decimal points.
Oh, really? Interesting.
That makes a lot of us, a lot of us who did some cynicism and skepticism, even more skeptical about that data as to why is it so close to Google data. The data should not be that close when you do different experiments and everything like that. Now, maybe it was coincidence, maybe not. It is tough to get information from China. We do know they are investing a lot. We do know there are tons of people working on it. Our message to our government is that take this threat very, very seriously. I mean, right now, it feels like the U.S. with the superconducting camp, and we believe we are in that camp right now. We seem to have leadership technology. They are working feverishly on that one. They may be ahead of us, and we just do not know.
Wasn't there an announcement out of something called Origin that came out of China recently?
There are multiple announcements. A lot of them come in their newspapers, like the South China Morning Post. Those are even harder to interpret. We honestly don't know whether they have been done for some geopolitical reasons or for some technical reasons. At least the paper that came Physical Review Letters was a real paper last week.
OK. Any other questions from the audience? We got a couple of minutes. Justin, go ahead.
What do you think about the next, like in the three to four years or five years, whatever the number is? For you guys, have you discussed publicly what's your sort of funding needs? Do you need X amount to get there? Or just in general, if you're doing much of a company, it's just.
Yeah. Regarding the funding, right, as I mentioned, right now, we are effectively between our cash plus Quanta, we have about $500 million. That takes us solidly into the next that's definitely good enough for about five years or so. In the meantime, many governments, including the government in the U.S., there are many big projects that are being initiated, the biggest one being DARPA, which is DoD, has a man on the moon-type project. That is like a $315 million project. They are going to announce this month a group of companies. We hope we are one of them that will be chosen to build this 500,000-qubit quantum computer in seven years or so. That is like the man on the moon-type project. The DoD's goal is to build the world's biggest, best quantum computer in six, seven years.
Their budget is about $300 million for that. That, obviously, we want to be the company which wins that project. If we do, then obviously our funding needs are taken care of. It is not just that. There are other initiatives in DOE. There is this Quantum Initiative Act that is going on right now in our House and Senate. The U.K. government has similar initiatives. You go around Western world. Even if you stay within the friendly, allied countries, there are many initiatives at different government levels that are potential opportunities. Depending on how many of those become real and how many we can win, we may not need to raise more money. We will certainly be opportunistic and continue to look at opportunities. We feel pretty good about our financial position right now.
Awesome. One more question in the corner?
Yeah. You've answered the capital question. Is there enough talent out there to actually deliver based on the fact that it's all in R&D right now?
A good question. I mean, it's always because we are dealing with very quantum physics-related topics, it's difficult. There are not too many universities who are graduating. Right now, we have about 150 people. About 50 of them are PhDs from top-notch schools around the world, many of them from Berkeley and Stanford, which is where we are. And we managed to get them. Our retention rate is really high because we have a very good reputation as a company in this field right now. It is always a challenge to attract. We continue to work with departments like Stanford and Berkeley, trying to encourage them to get more students. The good news is quantum physics and quantum computing is attracting young talent quite a bit. A lot of younger people want to come into quantum computing. They clearly see this is the next big wave of technology.
A lot of undergrads are starting in quantum physics right now because of quantum computing. We are optimistic that in the next 5 to 10 years, more schools around the country will graduate people with quantum physics background.
Is there talent to AI people?
Luckily, in that sense, in an ironic way, because companies like Google, I'm sure you have seen, and Meta are going through some significant layoffs in the Bay Area right now, that's actually helping companies like us attract talent from those companies.
We should probably stop. Subodh, thank you so much. I really appreciate the opportunity to chat with you.
Thank you.