Name: Saudi Arabia's Big Bet on AI
Uploaded: 2025-10-28
Duration: 19 min 31 s

Saudi Arabia's investment in AI is matched only by tech giants like Microsoft and Meta. The country is singularly focused on becoming a trusted global AI player by leveraging its geography and low energy costs to build massive data centers and partnering with industry leaders on cutting-edge R&D facilities. Can the country continue its progress by drawing world-class talent and strengthening alignment with international standards of trust and security, to compete on the world stage? 

Saudi Arabia's Big Bet on AI

Conferences

Thanks for joining us today. Um, this is gonna be, I think, a very interesting conversation. Um, Tarik, I want to start with you. Um, for people who don't really know Humane, you're a relatively new company. Um, can you just very quickly explain a little bit about what Humane is and what it's trying to do? Absolutely. Good morning, everyone. So, uh, and, uh, you know, it's the story actually started in August of last year. I mean, I mentioned to quite a bit of people. I, I, uh, remember it was my birthday. I was flying to see colleagues and friends. I got a call from the royal court saying, you know, His Royal Highness wants to meet you. And uh my first impression, oh my God, what I do, you know, so uh but it it's an idea that started from what I observed when I moved to the kingdom, which is plenty of amazing vision, um, great opportunity, but I also saw significant fragmentation in this market. If you really try to do something really um big ambitious, we have to pull our resources together, especially. When we talk about what you mean was started uh uh uh and organized around, it's a company that is focused on the AI total value chain. We're not just a model company, so we're building gigawatt capacity data center, not just talking about it, it's about 6 gigawatts of capacity that we have to really um build. This is massive, um, a diversified supply of chipset vendors, um, including what we're doing with Tiano and the team on both Edge and data center. A cloud platform that allows the customer to select the best chip to serve their model, a foundation model that was completely trained from the scratch. This is not a distilled from an open source, a model that understands our culture, our value, our religion, and then today I will talk briefly later on about true value realization. This is the part I really think. That humane um set itself apart on what we created around value realization, especially around enterprise and the vertical use cases that we want to drive. So it's a It is no doubt an audacious mission, and it feels we've been doing this for quite a bit of time, right? And there's a lot of talk these days about AI sovereignty. A lot of people countries are very concerned about maintaining AI sovereignty, but I think there's a question for a lot of countries of how much of that AI stack they can own. And when I say AI AI stack, you mention it's everything. From the energy to the chips to the data centers to the models that are then trained then and then the applications on top of that, but you're in this sort of fortunate position here in Saudi and with Humane. You're sort of trying to own pretty much the whole stack, is that right? Yeah, I mean, well, there's a reason for that. Look, if you look across the world, the main obstacle I think now people are facing is lack of energy. It's really simple as that. So, um, every company that is in the model space, every company that is building AI native applications, they are chasing capacity. So if Saudi Arabia had led the world through energy via oil exports, you know, I have no doubt in my mind we should lead the world through energy via token exports. So that's what we went after. Let's take advantage of the abundance of resources, land, power, connectivity that we have. And you have the ability, I understand, to potentially run inferences for a lot less here because of that abundant. Well, two reasons. So it is true that the cost of energy and the abundance of land and connectivity plays a huge role and factor in driving costs and economics, but we also do something very different. I don't see a single company um that that is always encouraging uh new entrant into the space to emerge, so the cost of chips are not cheap. So the partnership that I'm doing with Q Qualcomm really is addressing the technology disruption. So obviously the the relationship that we have with Qualcomm is not just about please Cristiano, give me the best price, hardly that. Actually more about let's talk about how do we innovate. At the chip architecture and the design to deliver lower costs on economics, especially for inferencing interesting, and I want to come to you, Cristiano. When I've talked to Qualcomm in the past, there's been a lot of emphasis on the idea of AI at the edge that you know, one of the ways we're going to address a lot of the concerns people have about data privacy, about personalization, about latency, and about energy usage will really be to push. This technology to edge devices as much as possible and Q Qualcomm has, you know, great chips that run on edge devices. But recently, you know, you've also started to talk about designing chips as Terek says with new architectures for servers, for data centers. Can you talk a little bit about this and where the emphasis is now in Qualalcomm? How are you, you know, is this going to be a balanced approach? Are you still mostly focused on the edge with maybe just some server design? No, look, I, I love this question. So maybe let, I won't unpack this because there's a lot of things that you just talked about it. So first, I want to zoom out a little bit. Uh, when we talk about computing and processing in general. it has always been no exception, cloud and edge. So it's interesting, like for example, for us, when we think about AI by definition in the very beginning we have clarity. AI is going to be hybrid because that's how computers are. Computers are everywhere and I think that you have a computer in the in the palm of your hand, you have your your laptop, you have your car and and because AI is the future of computing. And it's changing, you know, the architecture of processing, the architecture of software, it is going to affect every computer, including the computer at the edge. So I think from that, we have the premise that we need to take AI into everything and, and there's a huge opportunity. To actually make all of the devices that exist at the edge ability to process AI, and that's one of the things we're doing with Humane and and to be fair, I think one of the few companies on the AI space that actually saw this first. The second part is our DNA. So we come. We come from the mobile industry and I think one of the best thing about the mobile industry, it is a very, very difficult engineering environment besides being in a very difficult neighborhood too, uh, and in the engineering environment you have to pack a massive amount of computing, but you cannot compromise battery life form factor thermals. So what we decided to do is why don't we take some of the innovations we've been. Doing and take it to all of the computers at the edge all the way to the data center and that feeds into the reason we got into the data center and why Humane was an incredible partner for us is because of this view. As AI is going to scale and it's going to basically go from creation of AI to put into production of AI tokens for what matter? Tokens per dollar matter and We have the ability to bring some innovative architecture to really allow AI to scale because that's what happens at the edge and I think that's that's how we're thinking about the market right now and and I'm just going to say we're probably very blessed because we're one of the few companies today that can build a leading technology in 5 watts all the way to 500 watts. Very it's difficult to find. semiconductor company that is going from 5 watts to 500 watts and that's what we're doing right now, right? That's interesting again it's sort of this whole gamut approach and I think you've talked about the idea of hybrid hybrid AI, the idea that you know you're going to have some bits of intelligence that you can have at the edge, and other, you know, maybe if it's more reasoning, maybe if it's something that where latency is less critical, you're going to do that in the cloud. So you have this kind of high. approach and I know you've been talking to Humane about this, and I think you have something to show off even on this so I will tell you just to emphasize, by the way, I mean Christiane and I go since the CDMA days, 1996, so we've been working together for a very, very long time. You're showing our age. I know, very unfortunate. So but the reality is I'll give you my hypothesis how humane can succeed. When this uh direction about the consolidation and bringing various entities to form humane together, my immediate discussion with the team, the new team that came together from the various entities, but some are government, some are private. If we're gonna claim we are an AI company, two things are required. Well, number one, you need to have depth. And I'm blessed with that, by the way, I have an unbelievable talent that is really unexpected in a region where always been told you cannot do this. So my model team and my researchers are like world class actually, you know, what they've done is pretty impressive. Second thing, and I shared this with the Qualcomm and their Snapdragon summit. We said we will build all the systems inhumane with AI agents that are running the entire operations. That is to me the best sales pitch rather than giving anybody a PPT. So HR, finance, legal today, and including fully autonomous payroll integrated with Saudi uh banking system, we run it. So an idea came to my mind almost 14 months ago, um, what would be the future compute user interface. So we build, by the way, this is our, we call this Humane Horizon Pro, our first AI PC. This is a complete reference design built in the kingdom. So this is not a white label product because it's very different. White label is simple, by the way, just put your logo and put it there. But we actually crafted together with quantum engineering design, a reference design that allows us to build a premium segment product. At an affordable low effective cost, but I will tell you the objective was not to build a PC. We don't want to be a PC builder, but we reimagined how the new operating system of tomorrow looks like. So every humane employee today carries this laptop in the in the in their environment today. The operating system that is running this, and this is I'll give you a prelude to this, this week is a really, really big announcement for us. We are launching Humane One. This is the world's first AI operating system. So rather than looking at icons where you click for separate application, discrete application, now you either speak, your intent, whether I want to go on vacation, book a business trip, run my payroll. What's my P&L? So this has been a great partnership, by the way, we launched this in the US last month and this is just the beginning, and this is again not about making PCs. It's actually far from it. It is about the intelligence that we have put on the edge. What is running inside this is Q Qualcomm MPU GPU CPU. This is running a 7B model on device. Think about the privacy that you could now achieve with this device. So and this is one of many things that we're doing to drive edge applications. If the model doesn't have the right knowledge or the right data sets, then we go to the data center to extract. Additional information, right? That's fascinating, and I think you're going to see more and more devices like this. You're you're early with this. I mean this is the way a lot of people think, think we're going to go. But Christiane, I want to come to you, to you actually, and to you both, Turk, with a quick question because we don't have too much time in the session. But you know there's a lot of concern out there about an AI bubble. You keep hearing all this talk about valuations are too high. There's a lot of about what happens is there a bubble? And if this bubble bursts, what effect is that going to have on the industry? Cristiano, is there a bubble and how worried are you about, you know, possible impact for what you're doing at Qualcomm from from perhaps overvaluation? Yeah, nice to see you too. That's great thank you for asking the question, but let me, let me say it a different way. Look, and I know this is going to sound terrible, but When the internet started in uh in 1999, right, and uh and I'm I'm not saying this to talk about whether what happened in the dotcom or 2000, but when the internet started, there was this big vision for the internet and if you look at what happened today, right, from 25 years later, the internet is much bigger today than people even dream was the internet at that time. And it happened in different places, you know, who would expect it to be the winners in the early days, it changed, it happened at different places in different industries, but it was much bigger than people thought in 1999, and I feel this way about AI. When you think about AI in the long term. AI is going to be massive and it's probably underestimated. It's the new is the fundamental change how we think about computers and democratize so many things is how we process information and and we tend to think, we look at large language models, but there's a lot going on. There's besides what you see large language models, visual, everything that is based on image, physical AI, which is going to drive robotics, so it's a massive opportunity. So there's really no question. About how big this opportunity is and how important it's gonna be a part of the economy. Now the interesting thing is I'm gonna now answer uh your loaded question in a in a different way. Um there is has been this mad rush and uh to build uh capacity, uh, to, you know, everybody's playing to win. And there has been this incredible focus on we just gonna, we're just gonna build, uh, it doesn't matter the energy, it doesn't matter the cost. Um, as somebody that is now uh entering this place and, and coming from the other mature industries that we have been to. I think what the thing that Tarek said is it's very important. There will be competition and especially when I go to scale, and then I think the people that had invested in architectures that are able to compete and be able to deal with the efficiency that is going to be required are going to be very well positioned for that. So the answer to your question is, um, there, everybody's playing to win. It's hard right now for you to declare who the winners are. Uh, the opportunity is probably bigger than people think, but you know, it's, it's going to go through a competitive environment very soon, and, and I think that's what actually I look at this as the opportunity for Qualcomm, right? So not not everybody is going to win, but there's going to, you know, it's going to be transformative in the long term and 100% got it. Tarik, you know, I want to come to you because. As you well know, Humane's only about a year old, you know, a little bit over a year old company. You weren't the first in the region to kind of get out there. You have, you have a big competitor in G42 in the region. How do you sort of see yourself in positioning to kind of catch up and pull ahead in this kind of game to be the key regional player? So, so I think after the end of this week, um, I'd encourage everybody. When you see what you mean has done in 9 months, maybe you will ask how did that happen. So even though we are a um a young startup, but we're, you know, I, I will tell you I'm always thankful for having access to the depth and the talent that we have across all of our business segments. It might sound very strange. I don't necessarily um look at competition the same way that others look at it. I really love the idea. Of, um, distributed knowledge, um, and I don't like centralized knowledge, especially when it comes to AI. I think it's good for humanity to see different parts of the world participate into um um into this evolution. So, um, uh, I'd love to see this region whether it is coming from Saudi, UAE, uh, different GCC areas. Um, but as far as Hue is concerned, the mission at hand is really very simple. We have to execute. We have to be results obsessed, and I am not MOU obsessed, right? So the results is really what matters to me. So at the end of the day, a product like this did not happen just by coincidence. Humane one doesn't happen by coincidence. Humane search, humane vision, all of them are coming this week. All of this came together just in 9 months. I want to say that we have created truly an AI, Silicon Valley now in Riyadh, and we are just, as you know, just started, right? We're almost out of time, but I want to ask one other question, which is, you know, the geopolitics of this increasingly the noise from Washington is, you know, you either have to be on our side and build completely on an American stack or or forget about it, you know, if you want to, if you want to use Chinese. Components or equipment that's a problem. How are you both sort of seeing this shape out? Uh, yeah, maybe go to you first because you, you're a brand new company. You had a choice of, of what equipment to use and to build on. Look, it's, it is really, uh, it's a reality that we have to face, and it is unfortunate because I find tremendous opportunity, by the way, to leverage a large startup on east and west. Um, I consider Saudi Arabia to be, you know, at Switzerland in a way, but we also have to be cautious and really understand that, um, uh, you know, that the right partnerships do matter. So today for Humane, if you look at our technology stack, whether it is from data center, chips, software, cloud, we, we are really working with a significant amount of US partners and there is a depth of relationships that has been built to really advance this. Um, uh, some of these topics are completely out of hand. I hope one day that this, um, ability to find a way to work, um, between east and west would happen, um, but for the time being we really are partnering together today with, um, especially on the chips and the infrastructure with US suppliers and Christiane, just quickly, I mean, you're a US company, but you want to sell all over the world. I mean, how do you deal with this current environment? Look, I, you know, especially the semiconductor industry. Been navigate into this environment I think it's a reality we're doing the best we can, I think, to, uh, continue to be a global provider of technology but but maybe I'll, I'll I'll try to answer this question a little bit different because maybe I'm just gonna use the example of semiconductor manufacturing which one to get often, uh, you know, in the middle of those discussions. Um, it's, it's good to have, uh, geographical diversification of ecosystems, and that's true in general. I think that's true even about what Tarek just said about having the ability to democratize technology and get developed in different places, and you just need to know that in different markets there's going to be different requirements. You need to be prepared. to navigate with that, and I think that's the, that's the reality right now. I think you have to tailor business to each market and it is a good thing to actually have multiple ecosystems develop in parallel. Right? I know we're out of time, Turk. I know you have a helicopter to catch, so thank you very much for listening. Thank you, gentlemen, for joining us. Thank you very much. Thank you. Thank you.