Podcast | The Fall and Rise of a Quantum Company Boosting Algorithmic Development – with Sumit Kapur of Zapata Quantum

Konstantinos Karagiannis: Zapata’s back. You might recall some of the headlines about how they tried an AI pivot and eventually shut down in 2024. What followed was an outpouring of support from the industry as we rooted for the affected people and the impressive technology, which seeks to make coding quantum computers better. Find out about the amazing fall and rise in this episode of The Post-Quantum World. I’m your host, Konstantinos Karagiannis. I lead Quantum Computing Services at Protiviti, where we’re helping companies prepare for the benefits and threats of this exploding field. I hope you’ll join each episode as we explore the technology and business impacts of this post- quantum era. 

Our guest today is the CEO of Zapata Quantum, Sumit Kapur. Welcome to the show.

Sumit Kapur: Thank you, Konstantinos. I’m excited to be here. It’s been a great ride we’ve had recently, and I’m excited to tell people about it.

Konstantinos Karagiannis: This is almost the interview that couldn’t happen, but now it can. That’s what we’re going to talk about today.

Zapata, in one form or another, started in a Harvard lab nine years ago with a pure quantum mission. When you look back at those early days, what was the North Star that the team was chasing, and does that same star guide the company today?

Sumit Kapur: The way we think about our company — and we’re structured about this — is, we think about our why, our what and our how. The why has always been about translating and unlocking the immense potential quantum computing has and applying it to real — world applications. We believed then and believe now that organizations can create incredible durable advantage by investing in that application development ahead of the hardware actually being at the ultimate end stage because there’s going to be an evolutionary path on the hardware side, but we just need to make sure that the software side is keeping up.

Alongside that is an idea that we still believe even more than ever now: We believed then that quantum application development would be very hard — and, in fact, harder — than most would have thought. Here we are seven, eight years later after the inception of Harvard’s quantum computing lab, and we’ve done work with folks like Darpa where we were lucky enough to be chosen to lead all three technical areas of the landmark quantum benchmarking program. Those technical areas transcended all the layers of the stack, from the application layer and the utility down to the algorithm and then down to the hardware specification. We did that for multiple high-value domains, and what we learned is that fundamental premise we had that quantum application development is harder than most people think; it was even harder than we thought.

So, our why has actually advanced in the sense that we know we’re more needed than ever. What’s incredible is that right now, this same idea is being emphasized and highlighted by the rest of the industry. There was a Google paper that came out called The Grand Challenge of Quantum Applications in December 2025, which makes the point that right now, we’re seeing systemic underinvestment at the top of the stack. Our desire to unlock the potential of quantum computing by connecting algorithms to applications and by creating solutions for these use cases is, as ever, our North Star. We think we have even greater conviction, and believe the industry has even greater conviction, around it than when we started.

Konstantinos Karagiannis: That’s great. You touched on a lot of the things we’re going to dig into, and it was interesting to see so many people in the industry pulling for you guys in multiple ways. I agree too — that stack is super important, and the coding layer couldn’t be more important, because what are we going to do with these things? They’re cold, they’re fault-tolerant — what are we going to do with them?

Sumit Kapur: Right now, we’re in a situation where it’s actually highly critical for the industry. If we don’t find the uses, quantum computing is going to remain hype. We think it’s not going to remain hype, but it’s going to take some work.

Konstantinos Karagiannis: We’re going to start talking a little bit about some of the turmoil along the way. You transitioned from CFO to CEO during one of the most complex restructurings in the industry. What was that internal light bulb moment where you realized that Zapata’s story wasn’t over but was actually just beginning a new chapter?
Sumit Kapur: This has been the most significant career journey of my life. This is not a new story for me. My career journey before Zapata was at a company called 3Degrees, which is a machine learning platform in the environmental-commodities space. When I joined that company as CFO, it was in some distress, with about $5 million in revenue. We tried to do what we could to fix some balance sheet situations that the company had and to leverage some core technology and build on that core technology — namely, the machine learning engine that allowed us ultimately to become the 800-pound gorilla in that space. My journey there started at $5 million in revenue, and it grew to a billion dollars in revenue over 10 years. It was a wonderful, hyperscaling, everything-I-would-want-out-of-a-career journey. 

When I joined Zapata, [I saw a similar situation]. Unfortunately, the Zapata story was more challenging. First, it’s worth saying that the SPAC dynamic hasn’t treated a lot of companies well, and we fell victim to that as well. In 2023, we made the decision to merge with a SPAC entity and go public. Two things happened: One, we pivoted into AI — specifically, quantum-inspired AI — because that SPAC was focused on AI. Second, we hoped to get a great deal of equity capital out of that transaction. 

The pivot into AI didn’t work very well for us, and on the balance sheet side, we got almost nothing in equity capital. Instead, we brought in $20-plus million of unsecured liabilities to a wide variety of folks. That’s not strange for SPACs, unfortunately. I joined in May 2024, shortly after de-SPAC transaction; the stock had already traded down from $10 to about a dollar, and while we tried to right the ship, the plane ultimately went down.

That was the origin of what happened. Regarding how I realized there were enough parts on the airstrip to put the plane back together and get it flying again, first, the macro timing component was not in our favor the first time. Many in the quantum space entered a bit of a quantum winter for a period.

Konstantinos Karagiannis: Because of AI, ironically.

Sumit Kapur: Exactly. Then there was this wave of technical progress starting with the Willow announcement from Google, and Microsoft declaring 2025 the year of quantum readiness. At the Q2B conference, Scott Aaronson stated we’ll have a fault-tolerant computer running Shor’s algorithm before the next presidential election. He said it a bit tongue-in-cheek, but he was previously more skeptical. The technological progress unlocked a timing moment for us. That’s one ingredient. Second, in talking to folks around the industry, there was a lot of goodwill for Zapata.

Konstantinos Karagiannis: I was so impressed to see it. It was heartwarming.

Sumit Kapur: It was so heartwarming for me too. I had heard we had a bit of goodwill, but it was incredible for me to see firsthand. It was clear we had a niche to fill, and people were looking for somebody to fill it. Then, there were a lot of assets that we held. The IP portfolio is incredible, developed over that seven-plus-year period when nobody was really focused on the top part of the stack. It gave us the privilege to develop some IP that will be important for this space. We also had a portfolio of significant peer-reviewed scientific papers. By the end of 2024 or early 2025, it was clear there was enough raw material to work with to at least try.

Konstantinos Karagiannis: Do you feel that if you look at companies like, let’s say, Classiq, they’re using AI to optimize coding. Do you think Zapata was a year or two too early to the whole quantum-AI idea?

Sumit Kapur: We may have been a year or two too early. You only have a limited amount of runway. It might have been that we were at the perfect time but didn’t have quite enough runway, or that the SPAC transaction just didn’t work out in our favor. We might have been too early, but the good news is, we’re well-timed now.

Konstantinos Karagiannis: When you guys switched to Zapata AI, the first time I saw it, I was like, “What is this?” But when I started this show almost fjve years ago, I called it The Post-Quantum World right away, so I respect it. It was nice to see newsletters like IQT supporting the community by mentioning Zapata folks who were looking for jobs. The community is still small and growing.

What did that day look like — the ceasing-operations day in 2024?

Sumit Kapur: It was not a pretty day. We had obviously seen it coming and had taken every shot we could to try to save it, but we made the difficult decision that we just didn’t have the resources to continue. We did what we could to support folks in their journey afterward. It was the most difficult career day of my life when we had to tell everybody and ask people to find other journeys.

Konstantinos Karagiannis: I appreciate that honesty. Then you managed to restructure over $18 million in debt. How did you get investors to believe in Zapata 2.0? How did you get them to believe in that reboot?

Sumit Kapur: Very slowly. It was a bit of a flywheel. We would talk to a potential investor who would be interested if we could get conversion on the balance sheet side. Then I would go to the folks on the balance sheet side and say I could get financing from credible parties if we restructured. We were recreating the entire business vision of the company and could convey with great clarity the gap emerging in the space. We even had a major global consultancy look at the IP, and they backed us on the fact that it was incredibly valuable. It was a slow effort of building consensus that started to crescendo into a wave.

Konstantinos Karagiannis: Did the naming rebranding have anything to do with the convincing process, like saying it’s going to be “Zapata Quantum”?

Sumit Kapur: There was always the idea that we were going to rename it to something not AI. We were pretty convinced we wanted to keep the Zapata piece, and it obviously had to be Quantum. The Zapata name conveys enough goodwill that we want to keep it, and the quantum space needs us.

Konstantinos Karagiannis: I’d like to shift to one of the major things you do. You were able to secure global patents for quantum intermediate representation (QIR). Why is this universal translator so critical for the next stage of quantum utility?

Sumit Kapur: For quantum applications, they are going to be running on a combination of QPUs, GPUs and CPUs. You’re not going to be doing universal computation on a quantum computer just because it’s going to be more expensive than doing flow control on a CPU. When doing things like machine learning, you’re probably going to use GPUs. Because you have this complexity of needing to coordinate multiple hardware paradigms and connect them to human-readable code at the top layer, it creates the need for multiple layers of translation. We’ve done that in the form of QIR, and it’s an idea other people are adopting as well to make hybrid computing efficient going forward.
Konstantinos Karagiannis: Is the idea to have that work agnostically across modalities — whether it’s an annealer or trapped-ion?
Sumit Kapur: Yes, it would definitely be cross-architecture hardware-agnostic. The whole idea is that you would be able to compile across multiple hardware paradigms and abstract all that complexity away.

Konstantinos Karagiannis: It might need some tweaking based on modality to ensure that you’re getting the best translation for performance.

Sumit Kapur: Yes, that’s the whole purpose: to be able to abstract that complexity away from the people at the top layer who can then be free to code in an abstract, human-readable form.

Konstantinos Karagiannis: And somewhat ironically, will AI then enter that process as a coding agent?

Sumit Kapur: We believe that AI is going to be a key part of the solution. AI will not only represent one of the hardware paradigms at the bottom, it will also represent an ability to help us translate as part of the universal translator and create applications at the top layer.

Konstantinos Karagiannis: People are hoping that eventually, they can just say, “Pick what you need” and the system tells them what it found. How did working with Darpa across the benchmarking initiative shape the software you’re building today?

Sumit Kapur: With Darpa, we went end-to-end on multiple high-value domains like quantum chemistry and computational fluid dynamics. We did this alongside work for folks like BP and BASF. What we ended up doing was very arduous and labor-intensive work to prove out solutions and create resource estimates. It was very inefficient, and we realized we need to do that much more efficiently. We need to expand the size of the straw we are drinking through to empty out this container of potential applications.

Konstantinos Karagiannis: Hardware is picking up pace; we’re likely to pass 50 logical qubits across multiple systems soon this year. What do you feel would be the most immediate return on investment an enterprise could get right now with a hybrid approach?

Sumit Kapur: The very near term is the quantum-native stuff: quantum chemistry and condensed-matter physics. Those are the things Feynman had in mind when he proposed quantum computing. The obvious thing is, those native applications are near-term big-value adds along with one nonnative application: Shor’s. That is a case where you’ve mapped an arithmetic problem onto physics to arrive at a solution. The longer-term stuff is optimization, simulation and machine learning — mapping financial simulation or logistics onto physics. Those are a little bit trailing versus the native side.

Konstantinos Karagiannis: What’s interesting is, you’re not the first person within the last few months to hint that Shor’s might be one of the first things to do. Just a couple of years ago, we swore that was the far-away thing and there would be a hundred use cases before we’re cracking encryption. Now everyone’s basically admitting Aaronson was right — and so was the CEO of IonQ: We’re maybe nine quarters away from cracking encryption.

Sumit Kapur: That’s amazing.

Konstantinos Karagiannis: Let’s say we are in 2030 and we’ve already taken care of PQC. What is the one problem Zapata Quantum has solved that changed the world for the better?
Sumit Kapur: For us, it’s about solving the problem of quantum-application intelligence — giving organizations a clear idea of which problems are most quantum amenable for them, how to solve them, what algorithms to use and when to invest. McKinsey predicts total value creation in quantum by 2035 will be $1 trillion to $2 trillion. BCG suggests 90% of that value will go to 10% of early movers. We’re in an arms race on multiple fronts — geographic, hardware and the application race we are focused on.

John Preskill recently shared a quote from John von Neumann from the advent of the digital computer: “The most important problems that we will be solving with computers are likely to be the ones that we are least able to predict today.” You don’t just want to make things incrementally better; you want to do something completely new. Quantum application intelligence is meant to enable that exploration of high-value-add problems that people might not be thinking about yet.

Konstantinos Karagiannis: I’m not entirely sure McKinsey and BCG can predict what the use cases will be in 10 years. It might be like trying to predict where AI is going to go before the Transformer paper appeared.

Sumit Kapur: Totally.

Konstantinos Karagiannis: That was a great quote to end on. I vote we rename the architecture after Deutsch. Thank you so much for coming on — I genuinely mean it when I say, “Welcome back.”

Sumit Kapur: Thank you. The fact that this industry pulls together and supported Zapata is encouraging.

Konstantinos Karagiannis: That does it for this episode. Thanks to Sumit Kapur for joining, and thank you for listening. If you enjoyed the show, please subscribe to Protiviti’s The Post-Quantum World and leave a review. Follow me on all socials @KonstantHacker. Until next time, be kind, and stay quantum-curious.