It's a pleasure now to welcome Jack Hidary, CEO of SandboxAQ, to the Cipher Brief Studio. Jack is joining us from the Abu Dhabi International Petroleum Exhibition and Conference, which I don't think a lot of people in our audience know. It's the world's largest energy exhibition and conference. And Jack, I want to get into a few of the things that you're doing there, but first, I just kind of want to get a sense from you of what's the buzz like in the room there. I mean, it's an incredible amount of people, an incredible amount of exhibitors, all focused on energy.
Suzanne, great to be with you here from Abu Dhabi, from ADIPEC. This is the world's largest energy conference. Energy has always been important, of course, in our society, but now it takes on a new significance because of the deep connection between AI and power. Without more energy, we cannot power up the data centers for AI. And AI now is also helping us think about new forms of energy. So this interconnect between these two worlds colliding, AI and power, it makes this doubly exciting this year and last year compared to previous years.
I think, you know, you just hit on the buzzword there, which is the ability to generate the power that is needed to run the AI that is already changing the world, whether you like it or not. And there's a lot of talk as well about sort of China and competition in this sphere. Are people talking about that on the sidelines there, about where sort of China is in relation to building these energy centers and where sort of the rest of the world is at the moment?
Well, right now, actually, it's really about the total needed addition to power up world data centers. You know, there is some geopolitical discussion, but I would say the overwhelming discussion is how are we going to meet this need? Where is the AI going? What are the use cases of AI? All the different industries that it's going to change, and how do we meet this demand through an all-the-above strategy, all kinds of energy. The second topic, I would say, Suzanne, is really about how does AI itself help us have breakthroughs for energy itself. I'll give you an example: batteries. We know that batteries are critical not just for EVs, but actually the bigger market for battery storage is stationary batteries to store electrons generated but not needed at that moment. Batteries serve as a key buffer that can make base load power from intermittent power and deliver that when needed on demand. Now, batteries today are still too expensive to really go worldwide in terms of a commonplace means of storage. And so we want to use AI to actually bring down the cost of batteries. And that's a new kind of AI, Suzanne. That's AI for the physical world, not AI for the digital world. Most of the AI that all of our viewers here are familiar with is AI for the digital world. ChatGPT, Google Gemini, wonderful tools changing the world about how we do customer service and marketing and summarizing documents, creating deep research documents. But AI for the physical world, for new forms of energy, for storing that energy for batteries, that's one of the key topics here today as well.
That's fascinating. I know that you have been talking a lot with folks there as well, Jack, about AI and cyber, kind of the impact that's going to have. I just spoke recently for a State Secrets podcast interview with General Timothy Haugh and General Paul Nakasone, both who led NSA, who were commanders of US Cyber Command, and this is really top and center as well. How are you thinking about this relationship between AI and cyber, both kind of pros and cons?
Suzanne, this is a fundamental topic right now. AI is unfortunately the dream tool of all hackers. Whether they're a state-sponsored adversary or individual or group hackers, AI is their dream tool because with AI you can create phishing letters. You can look for patterns and find vulnerabilities in ways that no human could do before. On the other hand, AI is now also going to help us protect those critical national infrastructures and national defense and security. It's mandatory that we now use AI to defend our key infrastructure and our national defense. So AI both has a negative intent in terms of the adversary, but also can help protect us. And building on that, we also now have to consider the quantum threat. Quantum right now is not yet able to crack the length of keys that we use today in national security. When we think about the RSA keys, elliptic curve, those kind of security protocols that we've had for about 40 years, Suzanne, they're not able to crack those yet, but that time is coming soon. The quantum companies, be it the large companies such as Google, IBM, Amazon and others, Microsoft, as well as the upstart companies such as IonQ and others, they're making really good progress. And here China is making significant progress in their quantum initiative led by Pan Jianwei. And so in the next few years we're going to see the kinds of quantum computers that crack today's encryption protocols. The good news is that NIST, the National Institute of Standards and Technology of the Department of Commerce, has worked with partners worldwide, more than three dozen countries, to come up with the antidote, the defense against these kinds of quantum attacks. The challenge, Suzanne, is are we going to be able to switch over to those protocols in time before those quantum computers come?
Yeah, I think time is obviously really of the essence when it comes to all of these issues. When you look at the rate at which technology is advancing, I want to dig in just a little bit more with you, Jack, on the now versus hype question when it comes to quantum. It's often described as five years away. I think you said a few years away. What is actually deployable today in SandboxAQ's energy work, and what still really belongs in the research lab?
Well, great question. So, SandboxAQ, the A's for AI, the Q's for quantum, these are the twin engines of transformation of our economy. When you look at the energy sector that we're discussing here today at ADIPEC right here in Abu Dhabi, when we think about how can we get more yield when we get hydrocarbons out of the ground, what products can we make out of that? Not just some low-grade fuels, but maybe new kinds of products, maybe carbon fiber, maybe structural products that can help lightweight vehicles, that can help lightweight planes to make them more fuel efficient, even maybe become structural components of buildings. And so when we think about all the use cases of the energy products and the feedstock, we really can start opening that up if we have AI for chemistry. So again, this takes us beyond AI of language to AI for the real world, AI for the physical world. So that's one thing we're talking about. Now, those use quantum techniques on GPUs, the kind of chips that Nvidia makes. Nvidia, of course, is one of our investors, one of our partners, and it's great to be able to work with them. Just last week, in fact, we announced at the Nvidia conference in Washington, DC, a new AI model for catalysts. It's the world's most advanced model to make these catalysts. What is a catalyst? Of course, it's a compound that is used to accelerate a reaction, reduce the energy needed for a reaction, but does not get used up in that reaction. Catalysts are fundamental to the energy sector. Every major oil company here, every energy company represented, is a major user of catalysts. But it's been very hard, Suzanne, to develop novel catalysts. This AI developed with Nvidia and SandboxAQ now opens up that territory to innovation for all the energy companies.
Jack, you know, energy infrastructure is very messy, though. We're talking about legacy systems, siloed data, remote assets. I wonder what's the hardest technical challenge, do you think, when deploying AI in these environments, and how are you thinking about solving it?
Well, it's a great question. Fundamentally, first we have to think about numbers and equations, not words. LLMs, of course, have good application to summarize documents, customer service. If you've got a downstream business with filling stations, things like that, you want to use LLMs for customer service, marketing, great stuff there. But in terms of the heart of an energy company, the heart of it is the upstream, the midstream, the downstream. There you want to use quantitative tools, tools trained on chemistry, physics, the kind of things that create new products. What we're seeing in the energy space because of now the global surge in demand and also of more available supply, we're seeing the pressure on the energy companies to move up the value chain and not just offer the normal feedstocks as output, but actually offer value addition and become more integrated companies. We're seeing that happen both here in the Gulf region as well as back in the US.
I know you do a lot of work as well, Jack, your company does with government. When you're thinking about the future of national security, what are you thinking? I mean, I hate to ask this question, but I do anyway. You have to have some top targets. I mean, we already talked about the fact that time is really not on your side. The competition is fierce. What are you thinking the top three things are that need to be almost addressed or somehow resourced as quickly as possible in order to sort of ensure national security? It's a tough question.
