Conference Transcript - INFQ 2026-06-02

Mark Lepozis, Analyst — Evercore ISI

Thanks for joining us today. I am, my name is Mark Lepozis. I'm the Senior Semiconductor Analyst at Evercore ISI. Very excited to have Matt Kinsella, who is the CEO of Inflection. For not, for those of you in the room who are not living quantum every day, Inflection is the first publicly held neutral atom-based quantum computing company with a full-stack strategy, spanning not just the neutral atoms but also quantum software optical atomic clocks inertial sensors and quantum RF antennas and Matt has been involved in the company going on all the way back to his days as cold quanta participating in its seed round in 2019 and joining the company full-time two years ago so Matt thanks for joining us today thanks for having me here Mark so you know if I was to start off you know when we started first doing work when I was on quantum computing I have to say I was overwhelmed with the the technology and the different lexicons and if you could you know from a portfolio manager standpoint there's six or seven quantum modalities yep you're a neutral atom yes just to level set like how should investors think about the advantages of neutral atom versus the other modality sure well I can tell you how I think

Matthew Kinsella, CEO

about it today and I can also tell you how I thought about it when I first

Mark Lepozis, Analyst — Evercore ISI

I think that's great, I think it's great because things are changing so rapidly.

Matthew Kinsella, CEO

They are, yeah. And so I was, before I came full-time at inflection, as Mark said, I was the first investor in And I was at a firm called Maverick Capital for 18 years and was a public equity investor, used to read your research all the time, and then helped start our venture capital effort in 2014 and then seeded inflection, like Mark said. And what I found so fascinating about the neutral atom quantum modality, or even if I... When I say quantum, does everybody have an idea of what that means? It's like, yeah, sort of. Yeah, we're talking about the world of the very small, right? So the atomic and the subatomic levels, and the really weird quantum mechanics that rule the day down at those very small levels. And they are very different than what we experience on a daily basis in our macro world But if you can take advantage of those weird quantum mechanical principles, you can turn them into useful products And so that's that's what quantum is and why quantum matters and the reason why turning them into useful products matters is because they can Do things that classical products can't do or they can do things that classical products can do But with orders of magnitude improvement over how those classical products do the job So, when I was going down the quantum rabbit hole in 2017, I saw a lot of the different modalities, and what I noticed was all of those modalities were really just good for one thing, and that was building a quantum computer. But the range of things that you can take advantage of these weird quantum mechanical principles to build is much wider than just computing, but you needed a very flexible underlying quantum technology to address those different types of use cases, and that is where Neutral Atoms really shines. it is the most flexible quantum modality so I met a professor named Dana Anderson who was the founder of inflection and he was a professor for 40 years at the University of Colorado and you don't think of this you don't think of your Colorado is like necessarily like the premier academic institution but for atomic physicists it's like the best in the world Dana left Caltech to go to Colorado as like an upgrade so he had spent his life designing basically pioneering neutral atoms and what I found so interesting about them at that time was that flexibility and so we could build a quantum technologies company that wasn't just do we get a useful quantum computer or not but you could actually point this very flexible neutral atom modality at some more near-term applications like timekeeping like antennas like quantum sensing and then we started to put the R&D in place to discover whether neutral items was going to be an interesting candidate for quantum computing. So what I found so interesting back then was the flexibility. What I find so interesting now has been, and if I reflect back on the initial seed thesis, I had a pretty good idea that we could build a good quantum sensing company on neutral atoms, but it was unclear as to whether neutral atoms would be a great quantum computing modality. Today, what's been so interesting has been the progress that neutral atoms have made on the quantum computing front since we started investigating if they'd be good at quantum computing back in 2018 and really fast forward today neutral atoms even though they're kind of the new kids on the block are Winning in most of the metrics that matter to get to useful quantum computing. So there's a lot to like about neutral atoms The flexibility is one of them the ability to monetize in the near term but really the fundamental scaling ability of turning those atoms into qubits and then scaling those qubits up to the the numbers of qubits that are useful to do calculations gotcha and can you um so on that

Mark Lepozis, Analyst — Evercore ISI

topic or just to extend from there can you just talk about like in the broad brush strokes like how uh how investors should be tracking you and your milestones and what you know where are those opportunities to monetize along the way yeah two probably most critical metrics

Matthew Kinsella, CEO

to focus on for our company one is revenue and and i think that will show you can we execute on the commercialization of some of the more near-term quantum sensing applications and then the other is logical qubits which are really the in my mind the only metric that matters for quantum computing and so we have guided to at least 40 million dollars in revenue this year so i would hold us to that and then we've guided to hitting 30 logical qubits by the end of this year and our roadmap calls for 100 logical qubits by the end of 2028. So I think if you focus on those two things, that'll make sure we know how to execute and sell our sensors in the near term and then continue to drive the R&D to get to useful quantum computing.

Mark Lepozis, Analyst — Evercore ISI

Gotcha. And I think from my standpoint, what's exciting is when we were doing the work, we plotted the qubit scaling against the transistor scaling, and it's at or maybe even better than Moore's Law, double the transistor density, and it seems like if you're going from 12 to 30 to 100, that may be even faster than that kind of transistor density that improvements that we saw with Moore's Law. Is that a right way to think about this?

Matthew Kinsella, CEO

I haven't seen that specific analysis, but I'd love to see it, actually.

Mark Lepozis, Analyst — Evercore ISI

You said you used to read our research. I'll make sure. I would love to get back on the distribution list.

Matthew Kinsella, CEO

Yeah, exactly. It intuitively makes sense to me because one of the reasons why it's hard to keep track of all the progress in quantum is because it does scale double exponentially. And so it would make sense that it's going faster than the exponential scaling of Moore's And that has to do with the fact that as you add a new physical qubit to a system, and just to differentiate physical versus logical, physical qubits are the physical manifestation of quantum bits. But in order to make them useful, you have to error correct them and then take clusters of physical qubits and turn them into useful logical qubits. As you add another physical qubit to the system, that system scales not just exponentially, but because they're all entangled and they're all in superposition, you actually see a double exponential scale.

Mark Lepozis, Analyst — Evercore ISI

And so it makes sense.

Matthew Kinsella, CEO

The computational horsepower. Exactly, the computational horsepower. so the fact that that translates into a faster scaling of logical qubits mapped against Moore's

Mark Lepozis, Analyst — Evercore ISI

law does intuitively make sense that and i think that's maybe one thing that investors don't understand when they think about this industry you know if i think about adding processor cores to cpus they've started to hit a law diminishing returns they kind of scaled linearly until you get to four cores and then it kind of would level out and i think gpus scale linearly with the number of GPU cores correct but quantum scales exponentially with the number of qubits added which is a really exciting if you're if you're if you're adding qubits at the rate of Moore's law the computational horsepower would grow exponentially faster that's exactly right that's where the and it's hard it's hard to imagine the end use cases I

Matthew Kinsella, CEO

imagine right we can suppose about what some of the more obvious ones are but i think we all have to be very um humble for lack of a better term in in knowing that use cases we haven't even thought of will become right probably the primary use cases for quantum computers so when we talk

Mark Lepozis, Analyst — Evercore ISI

about um you know getting to a hundred uh qubits by the end of 2028 what are what are the long poles in the tent. What physics or science challenges have to be solved and to what

Matthew Kinsella, CEO

extent is that engineering? To get to 100 logical qubits at the highest level you need a lot of high quality qubits. So you need quantity and quality and then you need good error correction software. So to get to 100 logical qubits, we've solved the quantity piece. We have 1,600 physical qubits in our system. We have solved the quality piece. We are at 99.73% gate fidelity. So eventually we need those to be five nines, but at 99.73 we're above the threshold where we are, they're high enough qualities to get to 100 logical qubits. And so now it really comes down to two things, controlling those qubits with better precision, and that's all done with lasers and then improving the error correction software that we utilize and so to get to 100 logical qubits there is not really scientific breakthroughs that need to happen it's engineering at this point in time to go to a thousand logical qubits and then 10 000 logical qubits and beyond we are still in the realm of engineering but some science that needs to take place and then scientific breakthroughs can accelerate that whole timeline if and when they happen but that's it's it really if you think back to when did quantum computing go from if to when it was it was really still pretty speculative until late 2023 and in late 2023 neutral atoms showed logical qubits for the first time so it was the first time anybody had created a logical qubit on planet earth and so until we could prove that you could go from physical qubits to useful logical qubits it wasn't certain that quantum computers would eventually be useful we've passed that kind of binary threshold of going to zero to one logical qubits and then we did we did 12 last year we'll do 30 this year and we'll do 100 by the end of 2028 and and that to me was the

Mark Lepozis, Analyst — Evercore ISI

big change from if to when these got useful can you i mean this can be you know we're talking quantum mechanics and qubits and it can be abstract for a lot of investors as they try can you drive it down to like on a concrete level like what what near term are the applications where you're getting the 40 million dollars in revenues from like what extent is this you know hitting milestones with government um programs or versus like commercial applications how to think about

Matthew Kinsella, CEO

that and i'm sorry i didn't do it this time i usually will bring one of our quantum chips with me to make it more real but I left it back in Boulder and unfortunately when I take them on the road I usually break them so it's like I'm destroying some of our capital equipment by taking them on the on the road with me but I I would imagine I'm holding a quantum core in my hand and it looks kind of like a chip but it also sort of looks like a vacuum tube so like a little bit of a vacuum tube chip and it's about that big and inside that chip live atoms and in our case we use rubidium or cesium atoms and all of the products that we make are fundamentally using lasers to control those atoms that reside inside that core and taking those quantum mechanical properties and turning them into products and so we build clocks by exciting the atoms and taking the energy transition of the outer valence electron and that energy transition is sort of kind of one of the weird things of nature the most stable frequency reference there is out there and so you can create the most precise fastest is ticking clocks on planet earth by manipulating the atoms in that way. So that's where a lot of the revenue comes from today. Some of those more near-term sensing applications, clocks being one of them, where you can just kind of run a replacement cycle on the existing precision clocks that are out there with clocks that are a thousand times more precise. We also build antennas very similarly, instead of using that transition of the electron as a frequency reference, if we can put that electron way out in orbit. So think about if the atom looks like the solar system, this is now Pluto and that atom now becomes an antenna and it has like magical receptive properties where you can receive the entire electromagnetic spectrum with extreme precision so you don't need huge antennas to receive low frequency long wavelength signals. You can receive them in something very small. That's a very interesting use case for both military and commercial applications. That's a lot of the monetization today. we are sending our quantum gravimeters into space alongside NASA. They're our

Mark Lepozis, Analyst — Evercore ISI

biggest customer right now. A quantum granimeter? Gravimeter, yes. And so there's a whole new lexicon. It's a whole new lexicon, yeah. There are classical

Matthew Kinsella, CEO

gravimeters in space today that are sensing changes in gravity on the Earth's surface, and you can infer interesting things by changes in gravity. It could be the melting of polar ice caps, it could be the depletion of aquifers underneath the Earth's surface. It could also be the building of facilities underneath the Earth's surface. Anything that fundamentally changes the mass on the Earth's surface is interesting information to collect. Quantum gravimeters can do that with, you know, call it a thousand times more precision than you could with classical gravimeters. So that's what we're working with NASA to send into space right now. But there's a few of the applications that are outside of the computing realm that are useful today that we are monetizing. And we've also sold three quantum computers as well. So that's in the revenue mix too.

Mark Lepozis, Analyst — Evercore ISI

gotcha um okay so there was news a week or two ago you signed a letter of attempt with the Department of Commerce to receive a hundred million dollars yes what's the what is the money earmarked for and maybe maybe before that like what does this mean that the government is funding you I think you were the only neutral atom company that got fun there were two neutral atom

Matthew Kinsella, CEO

companies there was a private neutral atom company at us well if you go back November September October November of last year the Department of Commerce put out a call that basically said they were redirecting some Chips Act money to fund some fundamental technical areas that the US had to maintain competitive advantage in and it kind of broke down into photonics chips and quantum they basically then said okay all quantum companies across the globe show us what you have and tell us about what you can do. And then they worked alongside NIST, which is the National Institute for Standards and Timekeeping. Some of the smartest human beings in the world work there. And they did a deep technical overview of basically every quantum company. And they down selected to a small number of companies that they wanted to partner with going forward. The interesting thing is, is that this was a more broad call than just quantum computing. It was actually quantum sensing plus quantum computing. And we put in proposals for both. And originally I had thought that the Department of Commerce would skew more towards quantum sensing because it was more near-term. But one of the interesting takeaways from this is I believe they effectively saw what everyone was doing in quantum computing across all companies and came to the conclusion this was more near-term than they had historically thought. And so what they've done is they've kind of down-selected the companies they want to work with funded the milestones that we've told them we need to get through to get to useful quantum computing, and allocating that $100 million to the achievement of those milestones to get to those useful quantum computers as fast as possible. And so I think, to me, there's a lot of interesting things embedded in what just happened, one, the US government declaring quantum mission critical for the country, two, kind of coming that quantum computing is sooner than they had historically thought and then three having done a very deep technical overview of kind of every quantum company out there and then don selected

Mark Lepozis, Analyst — Evercore ISI

the ones they wanted to work with and the government is taking a stake they are so it's

Matthew Kinsella, CEO

not you're just a traditional 100 million dollar grant there is in exchange for 100 million dollars they get uh they're investing in the company basically gotcha okay we're we're kind of running

Mark Lepozis, Analyst — Evercore ISI

up on time we got a couple of more minutes i think uh you know when when you read the news stories on inflection. NVIDIA comes up, obviously an incredibly important company, right? It's nice to be aligned with them. Can you tell us what's the relationship between you and NVIDIA?

Matthew Kinsella, CEO

Sure. They have an unbelievable quantum team. And what they're doing for quantum to me is very smart, especially based upon what their core goal in life is, which is to sell more GPUs. And so I know that the company believes that as quantum proliferates into the world, that provides them with an opportunity to sell more GPUs. And so what they're doing is focusing on a couple of things. Their most recent announcement, the Ising models, were really how do you use AI to accelerate the time to use for quantum computing? And a lot of that has to do with error correction. And the detection of errors within a quantum computer is actually a quantum process. but tracing those errors back to their causes is a process that AI can do pretty effectively and so Using their models and tweaking them to our specific computer has helped us accelerate our error correction Then the other thing that they put out into the world is called NVQ link and that is linking GPUs and quantum computers together so that they can work together to solve hard problems and that's where I I believe they think their opportunity to sell more GPUs comes into play. Because I don't think QPUs, quantum processors, are going to replace GPUs at all. I think they'll work hand in hand to solve a whole range of problems that we just haven't been able to, nor ever will be able to solve with classical computers. So at the end of the day, the most powerful GPU cluster is boiling things down to zeros and ones. But the problems of nature really can't be boiled down to zeros and ones, because that's not how nature works. It's a heuristic for how the world works or how the universe works. If you want to try to model how molecules are going to come together and how the electrons in those molecules are going to come together, that's a quantum mechanical process that would take a trillion years for a classical computer to try to model. Quantum computers can do those because they exist and compute in the same way that nature And so that will then open up that class of problems for compute to be thrown at. But quantum computers aren't going to be able to do it on their own, not to be done with GPUs. So NVIDIA is trying to get quantum computing useful as fast as possible, and then connect GPUs and QPUs to solve those really hard problems.

Mark Lepozis, Analyst — Evercore ISI

So there's kind of a synergism between quantum computing and GPUs, kind of like we've seen more recently between CPUs and GPUs.

Matthew Kinsella, CEO

Exactly. And I think that's how you'll see the data center play out in the future. Just like GPUs have been layering into the CPUs that already existed in the data center, we'll see QPUs layer in on top of CPUs and GPUs just to open up the types of problems we can solve and we'll send workloads into the data center and it'll get partitioned and sent to the part of the stack that is most appropriate for that problem and we won't even care if it's quantum solving this just is something we couldn't have done before outstanding I think that's gonna have to

Mark Lepozis, Analyst — Evercore ISI

be the last word okay Matt thank you for joining the great insights yeah appreciate it a lot of fun yeah thank you