Please help me welcome our next guest, the co-founder and CEO of Cohere, Aidan Gomez.
Great to see you again. So you can help me with what I think is the question on the mind of a lot of folks here in Davos, which is, you know, all businesses are being told from the board down, you got to do AI. AI is big. They spent, you know, a year doing experiments. They spent, you know, I think most of last year trying to move those into production and a lot of them are still feeling that they're getting stuck. So, help us understand, you work obviously exclusively on enterprise AI. Where are companies finding success and what when they are struggling, what tends to be why they're struggling?
Yeah. Well, in particular, Cohere works in what we describe as critical industries. So the ones that are the most security sensitive: public sector, financial services, telco, etc. And so for them we have seen broad, large-scale deployments. We've seen customers 100x the scale of, you know, an initial pilot pool of tens or hundreds of users to tens of thousands, hundreds of thousands of users within the organization. What we're seeing as the major blockers are largely compliance related. So whether it's data security or checks and balances, monitoring of these agents making decisions around what should humans always be in the loop for versus what are we comfortable completely outsourcing to an agent. So our platform is built to tolerate that and be very controllable in terms of setting those autonomy policies. But it's still a broad barrier across industries.
And you know, I'm asking everyone this, so you're not the only one I'm asking the bubble question to, but I feel like that's another big thing that's going to come up constantly is this idea that there may be a bubble in AI. And again, I think when we're talking about a bubble, that doesn't necessarily mean AI isn't good, AI isn't real, AI isn't big. It's really a mathematical economics question of are we investing the right amount for the returns we can get over a reasonable amount of time. And I'm curious in that lens, do you think there is a bubble and if so, where in the ecosystem does it lie?
You know, Cohere takes a different strategy than some of the other players in the space. We don't actually build out the infrastructure for serving our models, our customers do that. So we give them the software, we give them the models, the platform around those models, but we don't build the data centers to deploy it. And so we're not as close to that side of the picture as some of the other guys are. You're completely right. It's not a question of whether AI is effective. It's incredibly effective. It's only going to increase from here. The question is about utilization. How much compute do we need, how fast, and whether the projects that are underway are timed appropriately to the realities of the market. I don't know, that is the short answer to the question. I don't know whether it's a bubble in terms of AI infrastructure. I can say there still needs to be a lot more from where we are today. But the numbers that you and I see in the press about what's being signed up for, I don't know if those are reasonable in terms of like the next two, three years.
But you're kind of happy you're not building them.
We'll talk about where you're taking Cohere because I feel like in the early days of AI, you know, there weren't that many big names, you know, and so we would constantly hear about you guys. And I don't feel like I hear the name Cohere come up as much, but as you mentioned, your customers aren't necessarily the ones that are getting talked about. They're definitely not consumers and they tend to be not just enterprises, but enterprises in, you know, kind of state-regulated industries. Where are you taking Cohere? What's been happening behind the scenes, where is your growth headed?
Yeah, we've been squarely focused on the enterprise and I think because we don't do these massive, you know, tens of billions, hundreds of billions compute deals, maybe we stay out of some of the headlines. I would say that's a good thing about our business. That's a feature, not a bug. But our focus has been on security, ensuring that companies have full control over their AI systems. So they deploy the models in their environment. Cohere has no ability to observe data flowing through it or shut it off. And so in the new realities of the geopolitical landscape, that's increasingly important is this notion of sovereignty and resiliency. So Cohere's deployment method and the fact that we're Canadian has been a huge, you know, boon to that.
Well, that's literally where I was going next. Today is in this environment which, you know, I've heard is a little turbulent on the geopolitical front. You know, AI is obviously more strategic, enterprises and governments care a ton about sovereignty, data residency. How is that affecting buying decisions and is it a benefit? Are you seeing specific business that you're winning because you're not a US company or maybe not even a US or Chinese company?
Yeah, I think resilience is a huge priority now globally. The past year has sort of woken everyone up to the reality that we're going to have to carry our weight. We can't all depend on a single country, a single liberal democracy. And there does need to be, you know, more than just one country carrying the weight in AI. And so in that sense, Cohere has become a focal point of one of the few companies outside the United States which is able to deliver the services and has developed the capability to build these models. I think it's only going to become even more critical and important and that's why we focused on these critical sectors like the sort of pillars of the economy including public sector. It's because those customers, it is a national security concern, right? Like telecoms, the communication of your public, all of their text back and forth, in financial services, every single financial transaction, interaction, in energy, the grid and being able to protect those systems, their architectures, etc. All of that is a national security concern and Cohere is one of the only companies on the planet that can deploy securely enough to meet those organizations' needs.
And how global is the business? I think so we sat down I think it was two years ago in Davos. You had this tiny apartment and we met and then we chatted on stage last year. How global is the business? Give us a sense for your scale and also I think you were expanding internationally maybe to I feel like Asia for the first time one of the times we chatted. Where are things at?
Yeah, we have fantastic partners like Fujitsu in Japan, LG in Korea. And we've grown a lot into Europe. So I think this time last year our revenue was something like 80% US concentrated and at this point that's come down to less than half, while also more than doubling in size. And so there's been massive rest of world growth for us in Canada, in Europe, in the Middle East, in Asia. Yeah, we have offices all over the place. I think the latest we opened up was in Seoul. And before that it was Paris. So very rapidly expanding. We expect to like double headcount this year. And I think that is driven by these changes and the broad awareness that each country cannot depend on a single other country. They need to spread their bets and make sure that they're resilient in the event that one of those gets switched off.
