Scale AI Valued at Nearly $14B With Amazon Funding

What does that tell me? That's very interesting, the sort of strategic backing you've got beyond just the size and scope of the round. First of all, thanks for having me. I think one of the key things of our role in the AI industry is that we really are an infrastructure provider. The three pillars of AI ultimately are data, compute and algorithms. Folks like Open AI solve the algorithmic piece, folks like NVIDIA solve the compute piece. And our role at scale is to solve the data pillar for AI. Our data foundry today powers nearly every leading large language model, including those from Open AI, Meta, Microsoft and NVIDIA. And so ultimately one of our goals with this financing ground was really to ensure that we can continue serving the entire AI ecosystem. So when you look at a lot of the strategic and corporate investors that we've brought on board, it really was to sort of bring together the entire ecosystem and multiple layers of the stack. So this includes other folks in the kind of infrastructure layer, folks like NVIDIA and the Intel as well as folks in the model layer, folks like Amazon or Meta and then lastly folks at the application layer like Cisco or ServiceNow. And so our our goal ultimately was to ensure that we can continue serving the entirety of the AI ecosystem as a little infrastructure provider by by sort of like bringing together that entire, that entire cohort. What about Google or Microsoft, are they interested? Is that for the later dates or can we only go with a certain number of each player on each system? You know, it's always, it's always like herding cats with these, with these, with these corporations. But we're, you know, our goal is to continue serving the entire ecosystem. We want to ensure that artificial intelligence on the whole is able to is able to accomplish the incredible potential, you know, our data engine. Our goal with our data engine is to generate all the frontier data needed to fuel us to AGI and pinch even beyond. You know our view. One of the ways we think about it is what are all the problems in data that need to be solved to get us from GPD 4 to GPD 10 and how do we ensure that we have the needs of production to do all that? Let's talk about the means of production, about the action of labeling such data. Because I know that you've really been thinking a lot about AI safety, the development, the, the biases, ensuring that that's something that's thought about within your business and those that you serve. But there was a lot of concern back in 2023 about who you employ to label data and a lot of that work being done in the so-called global S how you're paying them, at what rate you're paying them. How is that being solved now by technology, Alexander? Yeah. Ultimately, you know, we believe that the future of AI data rests on three principles, data abundance, frontier data and measurement and evaluation. In terms of abundance, I mean I think this is one of the the clearer areas we need to ensure that we are able to build a data foundry that utters in an era of data abundance. You know these models are becoming increasingly data hungry due to the scaling laws. Every successive generation of models requires exponentially more data and we need to ensure that we are able to build the the systems and the means of production that that allows to not resign ourselves to data scarcity. A lot of the key for us comes into the second bullet point though, Frontier Data. As we develop progressively more and more powerful AI systems, we need to be building Frontier data which is always pushing the boundaries of AI capabilities towards, you know, more advanced areas such as complex reasoning agents, multi modality, multi linguality and more. This from this production of frontier data requires human experts all around the world. And so you know, our view is that that humans and expertise are critical component of this production process. Alex, I remember August 2019, you were on a previous iteration of this show, you were 22 years old and you've just done a $100 million series C Fast forward to today, you've just announced the European HQ in London, you're valued at $14 billion. What's that like? How do you feel? You know, I think ultimately the most gratifying piece for me and I think for the entire company has really been how far AI has come in that time frame. You know, in 2019 it was the very primordial early days of of what's now called generative AI. You know even on that program we didn't talk at all about about the exciting things that Open AI or other companies were doing. We were talking about self driving cars. If you Fast forward to today we've gone from GPD 2 to GPD 4 and beyond or four and four O and beyond we have we have AI systems that are significantly more capable and I think we see a path to AI really improving everyone's lives in a way that was more of a pipe dream back in 2019. And so my hope is that you know five years later, if I'm back on the show that you know we can, we're looking back on the technology and seeing even further development in artificial intelligence, even further application of the technology, even further impact from from ultimately what we view as the the most exciting technology of our time.