PsiQuantum raises $1B to build million-qubit optical quantum computers in Australia and Chicago

liked it. Cool. Like I I made them smile. They're my friends. It it creates a much tighter uh you know, small hometown feel. If I don't like one of your posts, it's got to be really bad. Yeah. Because if what's the quote? There's if there's one person liking one like it's me. It's me. Fantastic.

Uh anyway, we have our first guest of the show or second guest of the show because we already had Bill Bishop on the show in the reream waiting room. We'll bring in Pete from Scantum. Pete, how are you doing? What's happening? Hey guys, how you doing? Doing fantastic. How are you? I hear you have some big news.

Break it down for us. What's new? Introduce yourself and the company, please. Oh yeah. Uh thanks, thanks a lot for having us on. I really appreciate it. I'm Pete. I'm a co-founder of Psych Quantum. We are a quantum computing company based in the Bay Area.

Uh and yeah, I moved here 10 years ago uh to build a quantum computer and uh thought I'd be on the boat back 6 months later and we're still here and we're uh really just like humbled to be now announcing a billion dollar fundra for us to go ahead and actually build giant quantum computers. Incredible. It's massive.

$1 billion. Um, okay. Give us give us the 10 years in uh 2 minutes.

Well, I'll give you the the um prior to 10 year that 10 years, myself and my co-founders spent about a decade in the university system doing what people think of when they think of quantum computing, which is science experiments in a research lab with single photons and chips and so on and so on.

published a ton of papers and then got fed up with that and uh uh some of my colleagues made a bunch of breakthroughs in the architecture for optical quantum computing and so we sort of developed a conviction that we could actually build a million cubit system.

You know Google and IBM and others are all in hot pursuit of useful quantum computing. Everyone today can make systems of tens or hundreds of cubits. Uh but you need about a million cubits. You need a huge system.

And so we sort of uh convinced ourselves that we could actually build that quickly through leverage of the semiconductor industry and we moved to Silicon Valley uh basically to build the organization that would do that.

Uh that meant taking weird devices like single photon sources and single photon detectors uh that you know you can't find in a regular manufacturing process and putting them into a commercial fab.

So we spent north of hund00 million and many many years putting superconductors and single photon devices into uh global foundaries fab 8 which is uh very big very mature commercial semiconductor foundry in upstate New York last remaining bastion of pure play semiconductor manufacturing in the United States. Yeah.

And um so yeah like big emphasis on semiconductors basically for many years. We were doing other stuff as well kind of architecture and algorithms and systems and so on but really spending money on semiconductors. Now we have these wafers.

We can make like high volume uh we can manufacture in pretty high volume the devices that we need. And so our emphasis is sort of switching towards building data center like machines.

And so last year we announced about a billion dollar Australian deal with the Australian government for us to build our first system just outside Brisbane. A few months later we announced that we'd be the anchor tenant of a new $500 million quantum computing campus on the south side of Chicago, former US steel site.

Used to make a million tons of steel a year. We are honored to be uh breaking ground there in a very very short period of time. and uh yeah just thrilled to be at this level of maturity and sort of readiness now to actually realize these systems. How does the Australian government even assess something like this?

I feel like the government is not known I mean at least the US government's not really known for having like tons of technical expertise internally. Uh famous examples of like the government struggling to build uh a CRUD app for you know registering your healthcare information.

uh it feels like the technical due diligence to understand how to underwrite that investment and then actually get any sort of value out of it. I mean you you have to rewrite systems uh to actually write something that's that's going to run on a quantum computer. Correct. Like what what was that what was that deal like?

What was that what is Australia optimistic about in having a billion dollar facility? Yeah, I mean governments uh struggle terribly to build cred apps, but like the entire might of Silicon Valley also struggles to build a good to-do list app. Like there is still no good to-do list as far as I'm concerned.

Uh I have a sort of uh provocative thesis that building to-do lists is actually harder than building quantum computers, but that one will get me in trouble.

But you know governments have excelled obviously in Manhattan projects and James Webb Space Telescope and you know space shuttle and you name it like uh semiconductors uh the internet uh and and and you can also look at companies like SpaceX and Tesla and TSMC and ASML like the pillars of our economy and our advanced society.

those companies wouldn't be the companies that they are today without literally billions of non-dilutive government support in every case. Yeah. Uh and so I don't think it's unusual that a government supports this.

I think governments increasingly understand that hard technology and this kind of frontier technology is sovereign and strategic and uh dangerous if you don't uh pay attention to it. We see that very acutely with semiconductors and AI supercomputers. Yeah. Um and then you asked about the sort of diligence.

So the chief scientist of Australia um spent about 18 months together with a pretty big team uh evaluating what we're doing. They came to our labs and held us by the scruff of the neck and made us uh reproduce measurements that we were claiming.

Um there was a Freedom of Information Act uh request that revealed her emails early in that process and um uh her emails revealed that she was highly skeptical of us when she first met us and the Australian press sort of wrote about this in the frame of like you know chief scientist highly skeptical of psychonum claims which I'm like what do you want her to be?

Do you want to be to be like high? It's her job to be skeptical. And and you know, we were just like super gratified by the end of that process that um uh obviously after speaking to a lot of different different outfits, they they uh they chose us and then in the US.

So, so sorry to cut you off, but I I What is So, so the Australian government, you guys are partnering with them. What is what is the value that they're hoping to get out of this out of this project? Yes. Yeah.

So I mean I think like if you look at semiconductor fabs, if you look at the current effort to get TSMC into Arizona, if you look at some of the sort of horse trading that's going on around the location of AI supercomputers, I think nations now understand that it it is strategically important where these systems are built.

And you know, when the US incentivizes TSMC to build in Arizona, uh the US government isn't buying chips from TSMC. You know, they're not um uh owning the fabs. They're just making sure that that infrastructure is built uh on a piece of land that they care about.

And to First Order, the Australian thing is is structured the same way. It's worth saying that Australia as a nation has been historically forwardinking in investing in the uh academic foundations of quantum computing.

So a lot of the basic research two of my co-founders uh a lot of the early work on optical quantum computing and error correction uh all came out of Australia. And so it's pretty poetic and uh and touching actually that it's heading back to Australia where it all started.

What's the uh what are some of the use cases that you're excited about? People talk about like oh you'll be able to break Bitcoin that'll be like disruptive or crypto or cryptography but uh is there ord or or quite profitable? Yeah, man. Who knows you guys? Uh it's the same as like asteroid mining gold.

If you get if you bring a you know 10 trillion dollars of gold to the to America, what what happens?

But uh uh in the context of AI and supercomputing like actually like delivering business value is is there sort of like a near-term scenario where um quantum computing can drop the cost of inference or unlock a different scale of training or or will the the architecture of what we are doing with AI fundamentally change once a million cubit quantum computer comes online.

Yeah. So we're not going to replace GPUs. We're not going to replace AI supercomputers. Uh and although there is sort of some papers that talk about quantum machine learning, that stuff is really early at the moment. If someone tells you they know how to do quantum machine learning today, you should be pretty skeptical.

Sure. What we're excited about is, you know, the way I think about this is that language models now, I visualize them as this sort of expanding frontier that's just like eating through everything we do with computers. And by the way, everything we do in our lives and so on.

Uh there's a known limit to that beyond which we don't expect conventional computers will go. And on the other side of that threshold is chemistry, material science, drug discovery, fuels, catalysts, fertilizers, semiconductors, like the microscopic foundations of our physical world.

where conventional computers uh have a really really hard time doing their simulations and where we don't have the training data in general to train AI models that will reach far into that virgin territory. And that's what we're really excited about with a quantum computer.

A quantum computer calculates from first principles rather than making approximations from from data. And through that kind of first principles calculation, we think we can reach much much further into unexplored territory.

The other thing that's maybe interesting is that at Psych Quantum, we make the highest performing phonics in the world against a whole bunch of figures of merit.

And of course, phonics is increasingly being viewed as a potential solution to what is otherwise uh maybe a difficult situation in the scale up of AI supercomputers where we run out of power, run out of space, uh run out of money, etc. Mhm.

what what were what were the key unlocks uh that allowed uh you guys to put together this billion dollar round? It's certainly the larger than most IPOs. Yeah. Yeah. So, we I mean it's pretty surreal to be clear, right? Like this is the only job I've ever had.

Uh the only other real job I've had is working in a garden center stacking compost. So, yeah, I should be clear like this isn't this doesn't feel normal. It feels pretty crazy to me.

Um, I'd put it also in context with other things that have happened in Silicon Valley where people have spent a way larger amount of money uh before they get to a product on uh things that are debatably more or less valuable than a quantum computer, you know.

Um uh but the um the unlock one of the biggest things that we did when we raised our series D with Black Rockck and Tamasc and others um Bailey Gfords we told them that we were going to unlock one of the hardest problems for us which is a new material for optical switching.

We had to build an enormous crazy machine here in the South Bay. Biggest molecular beam epoxy tool in the world. uh molten titanium inside, liquid helium inside, huge risk for us and without that there's no way we're building an optical quantum computer.

Uh the shipping company dropped that into the Alama Bay and broke it. So we had some trouble during co uh but we got it done on time. We delivered. We got it built and then the other thing that we've done is it still at the bottom of the ocean or did you get it back? No, we it it just like fell and cracked. Okay.

Yeah, we were we were like late on the schedule. Scuba diving equipment and go down get it. Not quite. John John scuba certified. I I like scuba diving, you know. I like treasure. Willing to help out. That's amazing. I mean, there's got to be some amazing stuff down there like out on the on the port at Alamita.

It's got to be some beautiful stuff. But then the other thing we did, you know, the chandelier like the golden chandelier that politicians love to take photographs with. We got rid of that. Okay, got it. uh that that's in the trash.

What do the energy requirements look like over the next like decade if we assume that there's a million cubit quantum computer running? Are are we going to see a similar race for uh just like intense energy production that we're seeing in the AI super computer buildout?

No, it's like it's like the same story but told backwards is what you should expect with quantum computing. So like yeah AI and conventional data center I mean it's pretty helpful for us given that we're talking about building big machines.

Yeah, 10 years ago in Silicon Valley, if you were talking about supercomputers or HPC or like building a big computer, yeah, people would like fall asleep in the meeting and throw you out of their office, right? Uh it's nice that we're now surrounded by people who are sort of brave about building big machines.

Um but uh we are going to build a big machine and then the good thing is that the 3 months into building it, it's already going to be out of date, right?

like we we're at the beginning of that sort of densification miniaturaturization power efficiency journey whereas AI and conventional supercomputers I think are coming to the end of a you know multi-deade journey of uh more and more efficiency and so on.

So it's actually pretty exciting that we should be able to further miniaturaturize and densify these systems over time. Yeah, it's fascinating stuff. Uh thank you for hopping on the stream. I learned a lot.