Hi, my name is Tai-Danae Bradley. I'm a research scientist here at SandboxAQ, and today I am here with Jack Hidary, CEO of Sandbox, to talk about the next wave of AI. Jack, thanks for joining me today.
Tai-Danae, great to be here with you. It's an exciting moment.
It is an exciting moment. So I actually just wanna jump in and kind of get started on this topic because you were recently at an NVIDIA conference where Jensen, the CEO, talked about the next wave of physical AI. Now, Sandbox has anticipated this trend, and I just wanna jump into this a little bit. Can you tell us what is physical AI and how does Sandbox fit into that?
Well, it's a fundamental question because in the last two, three years, the world has been used to using LLMs, large language models, really fun, lot of exciting applications. They ingest all the worldwide web and you could ask it and prompt it for answers and essays and lots of great stuff, applications and customer service and marketing. But when we really talk about the next wave of AI, and Jensen hit it on the head with physical AI when he was on stage at his keynote. What does that mean? That really talks about the physical world. It talks about new medicines, new treatments for cancer, for Alzheimer's, for diabetes. It also means that we can think about new materials to lightweight our vehicles, lightweighted cars. So, the same amount of gas takes it further, lightweight our airplanes. It means that we can make materials stronger, lighter, and cheaper. And this is implications for 80% of the world's economy, great applications of large language models. But when we talk about bringing AI into the real world, you're talking about 80% of the global GDP, and that's a very exciting next wave of AI that Jensen identified and we at SandboxAQ have had as a vision from the very beginning as we bring all these new strands of AI together.
Now, Jack, in your answer, you mentioned large language models, which I know a lot of people know about these days.
Well, definitely Tai-Danae, large language models have exploded. There's so many companies doing it. Google, of course, with Google Gemini, OpenAI, and along with Microsoft working on ChatGPTs and then the reasoning models and there's so many others, Anthropic and also Mistral in France. And now, of course, the Chinese companies have a myriad of large language models as well, from DeepSeek to Alibaba's Qianwen. And also don't forget, Baidu's Ernie Bot. Ernie Bot is here now. So, really there's a huge number of large language models and really, they become commoditized because they're all using the same data, which is the data from the internet, the worldwide web, the same basic model, the transformer model initially conceived of and developed at Google and then put in a paper in 2017 and out there, as you well know, and now of course we see them proliferating everywhere. So an exciting moment for large language models. But yes, there's definitely, you know, an AI beyond that.
Well, Sandbox in particular focuses on something different—LQMs, large quantitative models. Can you just talk a little bit about that? What are the differences between the two and what does that mean for companies who are looking to use AI right now?
Yeah, no, great question. Quantitative models are really models that instead of focusing on being trained on huge corpuses of language, we really focus on first principles, physics and chemistry and pharmacology and engineering, the key drivers of the economy itself. When you wanna think about a new molecule, a new treatment, again, we want to think about chemistry, biology, pharmacology, we wanna think about a new battery chemistry that might help us for a satellite in space that might be part of the starlink system, as an example. We wanna think about battery chemistry and AI trained on that. We don't really want an AI trained in that case on social media. We want AI trained on the first principles of chemistry and physics, and that's where SandboxAQ is the global pacesetter, large quantitative models, LQMs, the data—completely different. We don't use the worldwide web, we don't use anything from the internet. In most cases, we're generating the data and that really helps us get beyond the data wall that is facing the large language models. Viewers here may know and may have heard of the data wall, which is that the language models have been so voracious, so hungry, they've now sucked in really the entire web and beyond, everything they can find in magazine articles and newspapers. They've licensed content from everywhere. And so really hitting that data wall, not much more to train on. In the LQM world, we're just getting started. We can generate a hundred million versions of a particular molecule that might be a great cancer drug, for brain cancer, for pancreatic cancer. We can develop 200 million or a billion different combinations of a new material that might lightweight a vehicle. And so there's a lot to be done in terms of generating data in the LQM world.
Well Jack, I wanna shift the topic a little bit because Sandbox recently had some exciting news. We've had another big capital raise, 450 million from top investors. Can you tell us a little bit, what are the implications of this and what are you gonna do with these new resources?
Well, Tai-Danae, it's very exciting to have closed this round of $450 million, not just in terms of the quantum of capital, but also the quality of the investors. It's great to see our previous investors from our first round re-up into this new round, always a great signal for a company. And then also wonderful to see new investors, the majority of the capital of this round coming from new investors to SandboxAQ and that includes financial investors, such as Alger Asset Management, the global investor based out of New York City, focused mainly on public equities, but also able to invest in a handful of private companies like SandboxAQ that are growing to global scale. Ray Dalio, well-known as one of the top investors on our planet today, the founder and creator of Bridgewater CEO and chief investment officer of that company for so many years. And now focus on the next wave. He's always ahead of the curve, Ray is, and so wonderful to have him as an investor and to seek his advice on key strategic issues. And then in addition to those financial investors, to have corporate strategics join in this round as well. And as two examples, we have Google and Nvidia who've joined in this round and great to have them in the round and working closely with them as well. So really very excited about this round and what it means for the future of SandboxAQ.
