Thierry Labro: When I see the name of the company--SandboxAQ--I understand that, in a way, you are proposing a sort of sandbox solution based on AI and quantum technologies.
Fernando Dominguez Pinuaga: You’re right. The A and the Q in our name stand for AI and quantum. The group started in 2016 inside Google X, playing with both technologies. And we discovered all these applications. We spun out of Alphabet four years ago.
Since then, we’ve been working on AI and quantum technologies in simulation, life sciences, financial services, navigation, the Department of Defense, communications and cybersecurity--so, in many different fields.
In terms of quantum, we used to say that quantum technologies are not really ready yet. What does that mean in your case? Do you have a special high-performance computer? How does it work?
Yes, that is a very good question. You’re right--quantum computers are not ready. They are not business-ready. But we do not use quantum computers. Everything we do is quantum-related, but we use high-performance computing. We partner with Nvidia, use GPUs [graphics processing units, editor’s note], and work through Google Cloud and other cloud providers. We don’t use quantum computers.
But beyond quantum computers, there are many technologies that are linked to quantum and don’t require quantum computers. For example, sensors. In the medical space, we have a medical device for cardiologists--it measures the electromagnetic field of your heart. And then we use AI and quantitative models. I can explain that later. We use AI to denoise the signals from the sensors.
Can you give a concrete example of how your technology is used in real-world applications, perhaps even one that could be relevant to a country like Luxembourg? GPS?
That’s a perfect example. We use the same approach for navigation--this would be a very good use case for Luxembourg, for example. We put it on airplanes. We work with the Air Force, Airbus and Boeing. We have a sensor that tells you where you are without GPS--because GPS is very easy to jam and spoof. Forty-one thousand flights were impacted last year only between July and August, but of course, when that happens your pilot doesn’t announce it, can you imagine? To make that work, you use AI models to denoise all the signal data the sensor receives in the air.
What we use are large quantitative models (LQMs). They’re the next evolution of AI--physical AI, that applies to the real world, the real economy, that brings real value.
Is your technology agnostic--in the sense that it could be useful in any sector of economic development? Or are you focused only on defence, communications, space?
What we use are large quantitative models (LQMs). They’re the next evolution of AI--physical AI, that applies to the real world, the real economy, that brings real value. Large language models (LLMs) are great, but they focus on language and content generation--great for creating content or reducing cost.
But if you want to develop a new drug, for example--we work with a pharma companies on problems like Alzheimer or Parkinson’s--you’re not going to use an LLM. If you ask an LLM: “Give me a new drug composition to cure Alzheimer’s or cancer,” it can search the internet, but the solution isn’t there.
So, you need quantitative models based on physical equations--physics, chemistry, biology--to generate compounds that really solve problems, whether it’s for cancer, new materials or batteries. Those are large quantitative models--not large language models.
In terms of the data points you develop, how could we compare your large quantitative model with a large language model? Do you use larger amounts of data? How do you work with data?
You don’t really compare them--you use them together. An LLM can be part of an LQM. Large quantitative models are not just single models--they’re systems. You have an LLM inside, and then you add the equations, a neural network, a knowledge graph, other AI models... You combine them. The data you use depends on what you have available, but then you can also generate new data.
Let’s say you combine two molecules--and those two have never been combined before. That information doesn’t exist on the internet. But you can run physics models to simulate the interaction and generate new data--digitally, not in the lab.
What are the consequences of developing this kind of technology, in terms of data processing? Do you need bigger data centers?
We do require high-performance computing--that’s why we work with partners like Nvidia, Google and others, who just recently became also our investors. When you simulate molecules, you run a lot of compute. For sensors, you don’t need that much. For cybersecurity, less compute is required, but you still need GPUs to run the AI models.
In Europe right now, there’s a proliferation of AI models--machines talking to machines, AI agents creating credentials and leaving them behind. It creates huge vulnerabilities. Do you work on this?
You need real security models that track what AI agents are doing--to give access, manage credentials. It’s a big challenge for companies.
We want to have a positive impact--at scale.
More generally, what is the mission of the company? What do you want to achieve?
We want to have a positive impact--at scale. That’s why we work on the biggest human problems and with global partners--like Deloitte, Accenture, EY and others. We work on big problems: lack of GPS, Parkinson’s, batteries, PFAS, cancer, new materials for the US Army…
For example, in any conflict area without GPS--like the Middle East--it’s a huge issue for aviation and other services. In the US alone, GPS outages cost $1bn a day.
Some countries, like Russia or China, don’t have the same GPS technology as in the U.S. or Europe. Do you have geographic restrictions?
Right now, you need the sensor installed on an airplane. We’re working with the US Air Force, Boeing, Airbus and other allies. We haven’t had time to look into China in detail, but I know they are working on similar solutions. China is advanced in AI and quantum as well.
Could you give me some figures about the company? Revenues, employees?
We are almost 300 people--30% based in Europe. Over 70-75% of the company has a PhD. We hire technical talent: computational biologists, chemists, cybersecurity experts, mathematicians…
We’ve raised $950m so far, from many investors--including the Department of Defense through INQTEL, and other investors like Eric Schmidt, Jim Breyer, T. Rowe Price, Marc Benioff and others.
Yes, we have revenue across all areas--but we don’t disclose amounts yet. We’re a private company.
Did you know Luxembourg before coming to Nexus? How can a small country like Luxembourg remain on the global stage in this world of giants?
That’s a good question--I hope to contribute to those conversations when I’m there. Luxembourg is small, but powerful and dynamic. It has real advantages--in finance, space , communications. Its history shows it can lead without having the biggest muscles or biggest investments.
One of its strengths is agility. You can pivot an entire industry. For example, right now, most global banks are working on post-quantum cryptography. It’s hard for big countries to move fast. But Luxembourg already has a security entity serving all its banks and is building a quantum-resilient infrastructure. It could become the first nation to be quantum-safe for financial services. That’s less challenging here than elsewhere.
They did it before--and now they’re leading the only global quantum-secure network, using QKD technology with ESA EvolutionQ and Rhea, supported by European funds. This network will connect cities like Singapore and Brussels. They always picked the right bets.
Who is Fernando Dominguez Pinuaga?
Professional background: Fernando Dominguez Pinuaga is vice president of global outreach at SandboxAQ, where he leads global expansion across AI and quantum solutions. He joined SandboxAQ after six years at Alphabet’s innovation lab Google X, where he helped launch breakthrough ventures. Earlier, he led strategic partnerships at Google, notably in Latin America.
Academic profile: Dominguez Pinuaga holds engineering degrees from Universitat Politècnica de València and CentraleSupélec, a master’s in energy markets from IFP School, and completed executive programmes at MIT, the New York Institute of Finance and Harvard Business School (AMP 203).
Rugby career: Before his tech career, Dominguez Pinuaga was a professional rugby player and represented Spain in two Rugby World Cups--a testament to his resilience, discipline and team spirit, which he now applies to building frontier technologies.
Fernando Dominguez Pinuaga will be speaking at Nexus Luxembourg on 17 June. Check out the .
This article was written for the issue of Paperjam magazine, published on 12 June. The content is produced exclusively for the magazine. It is published on the site to contribute to the full Paperjam archive. .
Is your company a member of Paperjam Club? You can request a subscription in your name. Let us know via