Google Cloud COO Francis deSouza and WWT CEO Jim Kavanaugh on the Future of Enterprise AI

In this episode, Google Cloud COO Francis deSouza and WWT Co-Founder and CEO Jim Kavanaugh talk about how executives are turning artificial intelligence into a true competitive multiplier by building intelligent operations that run faster, cheaper and smarter; create predictive customer experiences that earn real loyalty; and accelerate innovation by embedding AI into every workflow.

Editor's Note: This special episode of the AI Proving Ground Podcast was recorded during WWT's Business Innovation Summit, which took place at the PGA TOUR's World Wide Technology Championship in November 2025. 

The AI Proving Ground Podcast leverages the deep AI technical and business expertise from within World Wide Technology's one-of-a-kind AI Proving Ground, which provides unrivaled access to the world's leading AI technologies. This unique lab environment accelerates your ability to learn about, test, train and implement AI solutions.

Learn more about WWT's AI Proving Ground.

The AI Proving Ground is a composable lab environment that features the latest high-performance infrastructure and reference architectures from the world's leading AI companies, such as NVIDIA, Cisco, Dell, F5, AMD, Intel and others.

Developed within our Advanced Technology Center (ATC), this one-of-a-kind lab environment empowers IT teams to evaluate and test AI infrastructure, software and solutions for efficacy, scalability and flexibility — all under one roof. The AI Proving Ground provides visibility into data flows across the entire development pipeline, enabling more informed decision-making while safeguarding production environments.

Chapters

• 0:00 Setting The Stakes For AI
• 1:03 Francis D’Souza’s Unlikely Origin Story
• 7:06 Why Leave A CEO Seat For Google
• 12:04 AI’s Scale Speed And Adoption Curve
• 18:36 Google’s Full AI Stack Explained
• 26:40 Open Choice Meets Hyperscale
• 32:00 Partners Not Pro Services
• 37:14 AI ROI Inside Google On Google
• 44:15 Bubble Talk And Real Value
• 49:15 Security As AI’s Trust Layer

Transcript

SPEAKER_00: 0:00

From Worldwide Technology, this is the AI Proven Ground Podcast. For years, we've talked about AI as potential, but that phase is emphatically over. Enterprises aren't just experimenting anymore, they're rebuilding their foundations, from data and infrastructure to security, to make intelligence part of the way they actually operate. And at the center of that transformation is Google Cloud's chief operating officer, Francis D'Souza, an engineer-turned executive who spent his career at the edge of every major wave, from decoding the human genome to redefining how we scale computing itself. Francis' message is clear. AI isn't another tool, it's a new operating system for business, one that demands fluency, speed, and a trust at every level of the enterprise. On today's episode, we're releasing an executive-level conversation between Francis and WWT co-founder and CEO Jim Kavanaugh about how Google is reshaping what it means to lead, and why the winners in this next era won't be the ones with the biggest models, but the ones who learn the fastest. This conversation took place just a few weeks ago at WWT's Business Innovation Summit, which curated some of the most engaging and influential minds in business and technology today, and equipped leaders with the clarity, conviction, and playbook to compete and win in an AI-defined future. So let's jump in.

SPEAKER_04: 1:33

Thank you for having me in that overly kind video.

SPEAKER_01: 1:35

Thank you. Well, we're gonna sit down, like I said, you know, kind of like yesterday. You know me, there's no script. I'll just kind of be bouncing around. So I was thinking uh, and as Joe's right in front here, and Joe, we're we're college roommates, is president of worldwide. We had dinner last night, and this Joe knows he's like, I have no idea where he's gonna go, meaning me. So he's correct. Uh so where where I wanted to start this was I thought we had a great dinner last night, and there's just so many things to talk about. But one of the things I think are really important to everybody here, very important to me, and I even brought it up yesterday, is that uh, you know, you you never know the background of different individuals that you're working with or that you meet. And it's amazing what you can learn when you ask a few questions. And I think in a world today, you know, that we're in, being inquisitive, be kick being curious, being a lifelong learner is so important. So asked a simple question to Francis, and he knows where I'm gonna go to kick this off. Uh so I'm like, so Francis, you know, and we haven't known each other for for long, and I feel like he's a lifelong friend now. So uh, you know, we were we were talking last night, uh and I was like, so where did you grow up? And you know, what is kind of your story? And you know, him and I had similar paths, I mean, academically. I mean, I was well, not really. Uh he kind of he can so let's walk he he grew up in Ethiopia, ended up migrating to Dubai, and I and I won't tell the whole story, but it's just fascinating. You're like, oh no, you grew up out in the valley, you were in tech the whole time. You know, so you really don't understand, I mean, moving because of political strife, right, in your original, you know, your country, uh the Ethiopia, moving to Dubai. And then maybe walk through a little bit would have a story that, oh, by the way, he's 16 years old in Dubai and he doesn't come from a wealthy family. Uh somehow, I don't know, I still don't know how you he submits you know uh his credentials or whatever to MIT and gets into MIT. Now, at that same point, he disappoints his dad uh because of this. So if he can just give a little perspective on that conversation that you had Yeah, it was not a conversation I was expecting.

SPEAKER_04: 4:04

So as as you heard, I I was born in Addis Ababa in Ethiopia. We moved to Dubai. Uh, you know, we weren't we were poor. So basically the highest aspiration from my high school was if you did really well, you could become a bank teller at the British Bank of the Middle East. That was like the peak of what you want what you could do. But I was always kind of a science nerd and I loved science, and I'd go to the library and I read about all these things you could do. And so, you know, I I study hard, I get grades, I apply to colleges, I get scholarships, and I come and tell my dad, okay, I I want to go to college and by the way, 16 years old. I mean, that's at this time I was 15 because I was still applying. But, you know, so I tell my dad this, and I said, Look, I'm trying to decide between Stanford and NMIT, but I want to go to MIT because I think I'm a science guy. And as you heard, my dad was really disappointed. And I was like, okay, I wasn't expecting that. Like, why are you disappointed? And he looks at the two and he's like, I always thought, now he dropped out of school at 13, he had to work, so he never went to college, right? So he's looking at this and he's like, I always thought you would be the kid that went to college, and and this one says university, and this one doesn't. Like, why don't you go to college? And I'm like, I think MIT is like a real college dad. And he's like, no, and then I realized like in Dubai, if you don't go, if you can't get into a college, you go to an institute to learn how to fix HVAC systems or be a mechanic or something. And he still today, I think, thinks I went to Boston to do that because I couldn't get into college. And he still asks, like, so can you fix the TV now? And I'm like, I can fix the TV now.

SPEAKER_01: 5:44

So I thought it was just, I mean, it's like one, a very entertaining story, but incredibly you know, insightful that you don't know, you know, people in their background until you ask some questions and you see, and and especially when you look at it globally, the political strife that forced you guys to move and your family and the disruption. So uh it's it's amazing. I I think I just find it fascinating to learn about different people in their backgrounds. And now your dad's in the States now.

SPEAKER_04: 6:14

He is, he lives in Redwood City in California.

SPEAKER_01: 6:16

Still asking you to fix the TV.

SPEAKER_04: 6:20

What did you do again?

SPEAKER_01: 6:23

So it kind of worked out. Yeah, no, a little bit. So, you know, you've you've had an amazing career. Uh, you know, even last night I was just looking through and thinking about some things and with Illumina and and you know, you've been in the healthcare space for a long time. I think that's a fascinating space around AI. I won't go down that road, but I think you know, AI is gonna change healthcare with all that we're doing. But let's maybe we'll shift to you know, why you moved to Google. Because I mean you had a uh CEO position, you you're you know, had great success. And you know what what intrigued you know about moving to Google at this point?

SPEAKER_04: 7:05

It's a great question. And look, uh there are a few threads that have played out consistently over my career across the different things I've done. One, at my heart, I'm an entrepreneur, right? And and so although I've I've done a couple of startups, one Microsoft bot, one Symantec bot, and I've worked at big companies. I worked at Microsoft, at Symantec, uh, at Illumina, and here, but in all of them, I felt like we were at the cusp of a big change in the industry. And that's what excited me, right? It was cybersecurity at Symantec, it was genomic sequencing at Illumina. I'm also an engineer, so I look for technology-driven innovations. And I can honestly say that, you know, um, AI is by far the biggest thing I've ever seen in my life, you know, if compared to the internet, compared to genomic sequencing. Um, and this transformation is different from anything I've seen before in a number of ways. You know, one, it's bigger than anything I've seen before. AI will truly touch every part of our lives, our personal lives, our healthcare lives, our work lives. And in a company, it'll touch every part of a company. So it's bigger than anything I've seen before. Two, it's going faster than anything I've seen before. 18 months ago, if we were talking, I would have been talking to you about all these pilots that companies are having, and people are trying to figure out what AI was. And today, we have many, many companies that are in production with AI in customer facing or internal facing or really important roles. And so it's gone from pilot to production faster than anything I've seen in my life. And then the third thing is the people who are adopting it are different than any technology wave I've seen before. You know, typically it would be the usuals, the financial services firms, you know, uh, you know, parts of the government that were really aggressive adopters of technology because of what they needed to. In this case, it's across the board. We're seeing healthcare companies, for example, adopt AI much more quickly than they've ever adopted technologies before. They're typically among the slower adopting industries for obvious reasons. And yet today, you know, we're seeing that. We're seeing manufacturing plans. So whether it's you know, healthcare companies like Seattle Children's Hospital or ASCO or Mercedes-Benz and Toyota, we're seeing such a broad adoption. And so AI is, I think, bigger than anything I've seen. And as I was looking at it, you know, this big AI revolution, at the heart of the heart of the revolution, in my opinion, is Google and the work that Google is doing, right? The seminal paper that came out in 2016 that kicked off the Transformer model and Gen AI, Attention is All You Need in 2016, that paper, that was written by Google. And if you look at a lot of the key technologies, whether it was Kubernetes enabling the movement to the cloud, you know, or today the A2A protocol, a lot of that is being developed at Google. And so for me, it was exciting to say, look, I can see how big this thing is, and I can also see the benefit of it. I think that finally with AI, we are going to be able to tackle some of the biggest problems humanity has. The truth is today, with all the advances we've made, we don't understand most diseases. We don't know how Alzheimer's works, or Parkinson's, or autism, or schizophrenia. We actually don't even know what they are. That's a blanket name for a lot of conditions that present similarly. We don't have anything today that can actually combat climate change because at this point we actually need to remove greenhouse gases from the atmosphere, and we have no technology that can do that in a way that makes sense from an energy perspective. AI gives us the toolkit to be able to attack those big problems. And so that's hugely exciting. And that's what brought me to Google.

SPEAKER_01: 10:37

Yeah, no, it's it's it's amazing. And you know, I'll give maybe a little bit of background, at least my perspective on Google and how uh I think instrumental Google has been to this whole move. Like, you know, you touched on Francis the I mean, if you think about it, I'll frame this up and then just ask for your thoughts. And part of the discussion that we had here is that you know, the overwhelming amount of opportunity that Google has, and probably the biggest decision that they have to make, and Francis has to make, you know, with his CEO Thomas is what do you go after and what don't you go after? But if you think about it, if you go back and just very quickly you look at, you know, Francis brought up Google basically created AI. You know, if you look at the Transformer model, that's where, you know, really AI, generative AI really came out. And then if you think about it with Google's acquisition and innovation internally, you know, with DeepMind and the folks that they had in the engineers, you know, open AI basically was incubated out of Google engineers around the transformer technology that was driving the generative AI. If you think about Anthropic, Dario, the CEO of Anthropic, where did he work? He worked, you know, at Google. And and if you think about uh Arthur, you know, uh the CEO of Mistral over in France, where did he work? He worked for Google. And so if you think about this, Google had this technology. When I even asked Francis and others, I'm like, why didn't you guys patent that you know and own it? And you know, so they they looked at that and it's kind of like Google has been very much an open source kind of company to drive some of these innovations. So if you think about they were really the innovator around generative AI as they're building out. And the other observation I would say that is very, very fascinating that we'll go into is that I don't think people truly understand that, you know, and I look at it, and what worldwide's trying to do, by the way, I'm jumping around a little bit, is skating to where the puck is going. I believe where Google is kind of guiding where that puck is going. And part of it is if you think about today, there I I think Google's the only company out there that has one of the preeminent leading large language models. So if you think about Google, you you know, think about Gemini. You look at, you know, the other ones, if you look at Anthropic, you look at open AI, or maybe perplexity. Google is one of the top, you know, two or three. One could be one debatable, you know, in capability. Uh, but also what Google has, and we'll talk a little bit about, we talked with NVIDIA yesterday, Google also, from the silicon, from the chip level up to a fully integrated system, that they have their tensor processing units, their TPUs that they are using to drive and power their large language model. And oh, by the way, Anthropic, the other largest model, just replaced an order internally, and they're not really a systems company at this point for only a million TPUs. A million TPUs. And so we talk about it, and it's like the largest market cap company today, and Google's very close, is NVIDIA. Well, one of the companies that literally has the potential to move into that infrastructure fully integrated and in debatable, you know, maybe the uh, you know, very efficient integration of a full rack potential processing unit to drive AI outcomes. They have the ability to go to market if they choose with a fully integrated system and they're doing some of this, uh and then they also have a large language model. I don't know anybody else that has that. I don't know that anybody else in the world that has it.

SPEAKER_04: 14:38

In fact, I think you touch on a good thing, which is uh a good point, which is I think while Google is well known, we are not known well today in terms of what we do in the enterprise. And the reality is if you knew us two or three years ago, we are a different Google Cloud today because the market is different and our offerings are different. And so it's probably worth sort of touching on maybe a couple of things to know about Google Cloud, right? So why are we seeing so much growth? If you looked at our numbers, so we just announced our Q3 numbers a few days ago. So in Q3 for the quarter, we did$15.2 billion in revenue in Google Cloud. That grew 34% year over year. So we've already talked about the fact that we're over at a$50 billion run rate in Google Cloud. And in Q3, we grew 34%. Now that growth is accelerating. The quarter before we grew 27%. And Google Cloud, if you look at even Q2, we're growing almost twice as fast as AWS. So what's driving that? And so let me maybe reintroduce Google Cloud a little bit and say when I talk to customers, you know, there are sort of three important themes that keep playing back to us. The first one is it's very clear that AI is resetting the cloud market completely. Every conversation I have is about AI. And then it'll be about the rest of the cloud and applications, but AI is a core part of it. And when companies talk about AI, you know, we are the only hyperscaler in the world that is not only a hyperscaler like Azure and like AWS, but has our own AI stack. You know, Microsoft Officer doesn't, and nor does AWS, and nor does uh Oracle. And we made that bet 10 years ago. So we just announced the next generation of our TPUs, our chips. This is our 10th year of doing chips. So 10 years ago, we made the bet to be AI first, up and down, and create our own stack. So we created our own chips uh with TPUs, and we started working on that 10 years ago. We create our own models and you know, released the paper in 2016 with attention is all you need. And today Gemini 2.5 is the leaderboard leader in almost every category and encoding it's as an anthropic, I'd say, at the top two. Uh, and then we have Gemini 3 coming out soon. And then we have agents on top of that. We've launched Gemini Enterprise, and so we've created a full stack. We're the only company in the world, not even the only hyperscaler, because there are other chip companies. Obviously, Nvidia is terrific, and we're deep partners with them and AMD. There are other model companies, OpenAI and Anthropic, and there are other agent companies, you know, Salesforce is doing its agents and SAP, but there's no other company in the world that combines that all into a stack. And there are benefits to customers from having a single stack because our engineers are building the models and the chips and the agents all together. So what that means is if you run agents on Gemini on TPUs, the performance is better than anything you can get anywhere else. Because our engineers knew where the models were going, right? So, for example, sparse code capability and chips, that showed up first in TPUs because they knew where the model was going. And so there are use cases we can talk to about this is the big benefit in terms of power, in terms of cost, and that really matters because AI is compute intensive, especially from an inference perspective. That hits you all the time. Anytime there's a query, you're paying that cost. So, one, we have our AI stack, and we've been on that journey now for over a decade. We're the only hyperscaler with that. Two is we're a fantastic hyperscaler. You know, we use our infrastructure to run Google. And so if you want a WAN or if you want a hyperscaler, and just look at the stats, the resiliency, uptime. You know, for us, it's existential. We have no other business. That our network has to run Google, search and ads and YouTube. And so seconds of downtime is a problem, right? And so we are super, super, you know, focused on we have two million miles of fiber optic cable that we've laid around the world. That's ours. We just announced four more points in Africa, north, south, east, and west. I mean, we really we have ships that are around going around the world right now laying cable for us, right? And we do that constantly. Um, so terrific hyperscale. And then the third thing they really appreciate is, and and you touched on this, it's our ethos has always been to be open. Like we, everything we do, we publish in the open source. When we created the Gen AI transformer model, we published that paper. When we created Kubernetes, we published and open sourced it. When we did A2A, we published and open source it. And so it's very deep to our ethos. Our security products are all multi-cloud, right? So, yes, they protect Google Cloud, but they also protect Azure and AWS and other networks. And customers really appreciate that. We want to in every part of your business, and we'll talk to you about the benefit of a full stack, but we allow you to mix and choose. So we are one of the largest partners in the world to NVIDIA. If you want to use Anthropics model, you can go to our Vertex platform. There you'll see over 200 models in our model garden. You can get DeepSeek, you can get Cloud. And so those are the three things to remember. Like we are, you know, the full AI stack company, the only company in the world that has it. We're a fantastic hyperscaler, and we're deeply committed to always being open and offering choice.

SPEAKER_01: 19:53

So I I I mean it really does require you to step back and think about what we're talking about here. Because if you look at, and again, great partners, you know, these are other great partners. Microsoft does not have its own large language model. You know, if you if you think about their partnering with Open AI. And you know, if you look at Amazon, you know, the partnering with Anthropic to use it. Google is the you know largest hyperscaler, one of the largest top three literally has their own proprietary purpose-built AI model. And the challenge, it seems, today is, and then I I do go back and then look at there literally could be one of the best. If you look at even a we have this discussion because of our great partnership, and you think about it, there's so much cooperation going on. Uh the great partnership we have with NVIDIA, talk to Francis, talk to TK about this. Uh you know, and the and the view is high tide floats all boats. That that the way Google's looking at it and the way Jensen's looking at it, NVIDIA is that we want companies, there's going to be this collaboration and the market's so big uh, you know, that if Google decides to go to market with a fully full stack integrated platform in a more, I would say, robust way, uh that will just elevate the entire market and could be good for all customers across the board. But this is one of the challenges I think that Francis and TK uh they have at Google to decide how aggressive are you going to go, because right now, similar to NVIDIA, their capacity is outstripped just by a few customers that you have today. So, you know, these are things that we're looking at in from a worldwide perspective, trying to figure out how do we work with an organization that is such an innovator, such a leader, and really trying to understand and collaborate on where the puck is going and how to drive these outcomes and this infrastructure that is that is going to be deployed. And I so I I would also pivot that maybe you could shed a little light on, you know, Google, you know, is is really thinking about how it's reinventing itself. And maybe just to be very direct, you know, a little bit of my thought, and you had a different perspective on this. So I was like, you had these tensor processing units, and I'm like, why didn't Google bring it to market? I'm like, well, my assessment was probably wrong, uh, was that you didn't want to cannibalize your search, you know, you know, your monopoly on search that you have, and why would you want to introduce this at that point? And Francis was like, a little bit of that could be, maybe, you know, but from you know, maybe you could answer is that you know, your perspective was Google likes to have things rock solid before you roll them out to the market. And a lot of the generative AI when it was first being brought to market was hallucinating a lot. So, you know, it's it's and right now, you know, a question going back is when you look at the growth of the cloud and what where Google is, you know, if you really looked at it and people are saying, oh, they're gonna start descending because of the search is gonna direct. Well, your Google, Google Cloud business is going up, and I'm assuming it's this combination of search and AI. And the last one I would throw out there would be do understand you made a small security acquisition on top of the things that you talked about, having fiber deployed across. So how are you looking at it from building out this these you know leading edge, you know, uh large language models? And oh, by the way, they're gonna be rolling out their next version, that there's gonna be all kinds of probably enhancements around code development, multimodal capabilities, inferencing, uh reasoning. So I'm throwing out a lot of questions, but maybe the the round.

SPEAKER_04: 23:52

So let me start by saying, first of all, I this is a very important conversation from my perspective because the reality is that you know at Google we have made a conscious decision that we don't create a big consulting group or pro-serve group, right? For us, it's less than 2% of our revenue. Um, and and AI is going to touch every part of every business. And so for us, it's very important that we build a big network of partners, a big network of system integrators and software developers on top of the capability we have. And so, you know, sort of explaining what we have and sort of articulating the opportunities, because this is a giant need in the market from customers, it's uh it'll be hugely valuable to customers. You know, one of the things I do too at Google too is I run this initiative called Google on Google, where we are cataloging all the use cases uh that we have run internally for AI. And you can imagine we've been on this journey for years, and so there are many, many, many use cases. And and and we are being very explicit about what's the ROI we're getting from running AI in these use cases. Um and we're cataloging in, and if anybody's interested, we'd be happy to share. Um, and and I do sessions where we take people through this. And it's across every part of our business. In one part of our business, for example, you know, we have a requirement that we want to test the disaster recovery plans of all our suppliers. That you can imagine we have tens of thousands of suppliers, right? That took weeks before. Now it can take only hours, right, because we use Gen AI to drive that process. In another part of our business in Treasury, we looked at optimizing the daily cash balances of our cash. And that generated tens of millions of dollars of benefit for us in the U.S. alone. You know, we have we use it in our SOC, our Security Operations Center. Now, we have a military-grade SOC. So you're thinking, okay, well, how much more juice can we get out of this? And we were able to reduce our response times by 96%. So we're that much faster now in a military-grade SOC, right? So we've cataloged customer-facing scenarios, internal use cases. Like there are probably hundreds now of use cases with hard ROIs associated with each of them. So we will need all of us to realize that benefit in the market. And so it's very important for us to articulate what we have. We are looking for engagement across every layer of the stack, right? We have an agent platform, Google Enterprise, uh Gemini Enterprise. But the reality is, while some companies are capable of generating their own agents, most aren't. And so they want help in doing that. You know, we have the models and the model garden, uh, and then you can access straight up to our infrastructure and our and our chips. The other change that customers will go through and they'll need help with is this isn't just a technology that they're adding. This is a transformation of their business process and their workforce, right? The reality is a successful workforce in the future is going to need to be bilingual, fluent in their domain, marketing, sales, development, and fluent in AI. That's a giant change. And so we need all of us, and that's why we agree, that's why we're partnered with in every layer of the stack, we're partnered with the Nvidia's and the anthropics, and the we all need to work together because this is a big change. And to when we make that change, we all do better, and we do better as humanity, but we need all of us sort of engaged. And so I think it's on us at Google Cloud to really be you know more clear and more vocal about what it is we have and how all of you can engage with us to drive the transformation that we're talking about. But it is a big challenge. We have a lot. I know what this year at Next, which is our big customer conference, we remarked that between last year's Next and this year's Next, we had over 3,000 innovations we brought to the market. That's a lot, right? And so we need to do a better job sort of really distilling it down and articulating how you can access it.

SPEAKER_01: 27:45

No, it's it's I mean, we're we're seeing so much of what you said. I mean, we we have a it's a lot of the projects that we've been pushing for you know over two years. Like I'll have my meetings with our teams, and it's literally like AI being AI first and literally driving it. And I look at it uh on a combination, it's really this collaboration between business and technology and really understanding the platforms that you're building, the AI models that you're building, but how is that aligning to the line of business owners? So the owners need to understand how IAs work, the technologists need to help, you know, the the line of business owners to understand. But where where do you where do you see us uh today where you know, talk about like a bubble? You know, people are talking about a bubble. My belief is and you know, that if you really look at and you look at like, you know, earnings and you look at multiples, valuations of companies, we really haven't started to see the benefit of true AI in regards to automation, scale, efficiency, differentiation. There's there's things that are happening as you just described, and we see the same thing at worldwide. But as that really starts to take hold, I believe there's going to be a flywheel effect that's going to completely transform those that have made the commitment to invest in that journey. Uh so how would you frame maybe at a macro level and just where do you see when people talk about a bubble? Sure.

SPEAKER_04: 29:17

Look, um I think we can all see the you know stock prices go up and the enthusiasm for AI stocks right now. And so it's natural to hear, okay, do you think we're in a bubble and we all remember previous bubbles, right? Um in my in my mind, here's how I think about it. This is not a financial bubble, and like this is not like the Dutch tulips or something where it's purely just financial markets sort of feeding on themselves. This is an in like a technology driven change that's happening that delivers a real benefit to customers, right? And we'll see real productivity, we will see real new businesses created. And so I think there's a reality here that's different from just a pure financial bubble. And I think that reality. Is going to create an enormous amount of value. And so I deeply believe that. I think we will look back, and I think some things that seem expensive today in retrospect will look cheap. Somebody told me once that the year Google was created, pets.com was also created. And so it doesn't mean everything that gets created now is going to win in the future. There are going to be companies that deliver real value and sort of, you know, and then grow. And then there are some that just don't make that change. And in fact, even some of the big companies today, if they don't become an AI-first company, I think we'll struggle with making this change, right? So even big companies today. So I think huge value will be created. And I think there'll be, in retrospect, some great even investments being made right now. But not everybody will make that change.

SPEAKER_01: 30:44

So maybe another one that we we spent some time talking about Francis, and you guys have made some big investments, and it happens to be under your realm of responsibility, security. And if we think about where you're going with building out, really driving innovation and massive global deployment around AI, and then thinking about the infrastructure side that you guys are working on that is just incredibly fascinating. But how do you look at security? You made a massive acquisition. It's under your realm of responsibility. So how does that fit into the overall scheme?

SPEAKER_04: 31:20

In addition to being the CEO of Google Cloud, I'm also president of our security business. And so I'm I was the one driving the Wiz acquisition. And uh that is the largest acquisition we've made in Google's history by a lot. Uh so it's a$32 billion acquisition. You know, we're still in the approval process, and we we expect to be through the process by the first half of next year. So it's still in process. Um cybersecurity is very important. It's very important to our customers, uh, and it's very important to us at Google Cloud. Even before Wiz, we have a big cybersecurity business. It's a multi-cloud security business. As you can imagine, you know, I talked about the fact that we run military-grade security for ourselves, and we also offer terrific security to our customers. We have, with our Mandi acquisition, we have the world's leading uh incident response team out there. You know, uh governments call us if they have a breach and sort of we jump in to help with the breach, but as do you know small businesses that are hit by ransomware and and and corporations. Um, we also have uh you know SOC capability, we have, you know, we have a whole powerful slew of cybersecurity offerings. This is really important because you know, trust is the fundamental enabler for this big transition we're talking about. And AI is also going to fundamentally change the security business, right? Every part of it. You know, one of the biggest ransomware attacks we saw in the last few months was by an organized outfit that was not actually very technical, but had used AI to create the ransomware technology that they then used to attack this enterprise. And so, you know, AI puts a very, very powerful set of tools in the hands of the attackers. And that's whether it's ransomware tools, deep fakes, you know. So we're now seeing a much more enabled uh threat, you know, adversary, you know, whether it's military, you know, paramilitary, organized crime, or even just hackers are now enabled at this very powerful tool. It also is giving us as defense very powerful weapons to be able to hunt the attackers. And so, you know, we have an awful lot of AI that we're using in our defense technologies, whether it's our SOC or the defense we provide our customers, to be able to identify threat actors much, much faster and have agentic SOC so that you can respond much more quickly than any human can. And so for us, we used to really focus on we wanted to make sure if a bad actor was in the Google environment, you know, we wanted to measure that time in minutes. And we wanted to shrink that time radically. And you can you have the power with AI to do things like that. In addition, AI creates a new attack surface in your company. Suddenly, you know, people can attack your models, and you need to think about what does that mean? And so we provide a model armor to protect you from things like prompt injection, for example. But you also have to watch your model to see is your model subtly drifting, right? Is it being trained and moving away from things you want it to do? So how do you detect drift in your model? And then if drift is happening, how do you retrain the model or rescue a model and get it back? In addition, you know, data is the fuel of AI. And so now you have a huge amount of data, too, that is unlocked by AI and represents a new attack surface. I was talking to one CIO who said, you know, all those SharePoint uh servers that I had in my company that were abandoned that I didn't really care about because it was security by obscurity, nobody knew where they were, we didn't really check all the access controls and all the files, but it wasn't my biggest problem. Well, the challenge is AI finds it, right? And AI and agents sort of crawl your network. When you have a SharePoint connector enabled, all that data gets discovered. And so now you have that whole surface to protect. And so it was very clear to us that we needed to see a step change in cybersecurity capabilities for enterprises. And so that's driving the big investment we're making to say, look, very few companies can provide that level of cybersecurity that's going to be essential. And so we need to be there and help drive that.

SPEAKER_01: 35:23

Yeah. So it's it's I mean, it's pretty overwhelming to think about all the things that Google is doing and the investments that are being made. And I and maybe I'll open it up to the audience for questions here. But just to frame it up again, to think about, you know, Google, you know, I think, you know, everybody known globally as the search engine. You know, you're gonna go Google this, you're gonna go Google that, you know, it's uh but if you think about where they're going and you think about what they have, and the the other perspective I'd say is you think about the the the largest of the large large language models, the A AI model, gener uh generative AI models, you know, open AI. You look at anthropic, you look at perplexity, you look at Google. There's only one today that is generating massive amounts of profit. And you talk about not only that, you know, Francis mentioned their growth, but they're incredibly profitable. So those other three organizations, you know, are investing massive amounts of dollars and eventually are gonna look to turn the corner to drive profit. But that's a very different thing. So when you think about Google is one that, you know, I think when I look at it from a worldwide, we want to, you know, establish and integrate this partnership because they have such a massive platform. And for the most part, they were going to consumers now doing things in the enterprise. But I think Francis would probably agree, you know, that they they don't necessarily have the Microsoft Salesforce to go out and sell at that point, and you know, that but they have so much to offer here. So my my only uh point is you know, this is something the market is moving very, very, very quickly and it's completely transformational. Uh Google is a company that uh you know everybody needs to be being paying very, very close attention for so many different reasons. So let me maybe open it up to you for anybody here for questions uh to Francis.

SPEAKER_03: 37:36

Hey Francis, just um just with the amount of innovation happening across Google on a global scale at all times, um, any kind of lessons or insights you can share on how your your team goes about kind of connecting the dots, keeping the collaboration going across you know such a large organization, and and realizing when you've really got something and we should double down on this, just really how you manage the innovation process across Google?

SPEAKER_04: 38:03

It's a great question. Um, and one, you know, I think we're three decades old, so I mean, one that's been sort of you know in front of us all the time, which is how do we keep uh pushing innovation, how do we keep pushing the cutting edge, even as we get larger and larger. And and so there are a few things that I know are sort of guiding principles for us. You know, um one is you know, we really focus on sort of uh a long-term view on where things are going. Um and and so we have an an opinionated position on how things are gonna evolve. So we believed, you know, 10, 15 years ago that AI was gonna be a big transformative force, and we invested and and we were, you know, we were underappreciated for well over a decade about why are you building chips and you know, what is this AI models you're building? But we we believed because you know we are deep in the technology, we could see where the trend lines were going. And so one is we have a uh point of view-driven investment strategy around where we believe the world is going. Now, not all of them work out, and so we have gates along the way that says, okay, we wanted to explore this technology and it's a dead end. So, for example, right now, we're also investing in quantum computing, and we have been for a while. In fact, uh a couple of weeks ago, one of our uh the guy who runs our quantum team just won the Nobel Prize for the work on quantum. Um today, again, there's no real commercial benefit for quantum, but we believe you know where it's going, and it was a big milestone also we hit this week, where we hit the first quantum verifiability. So uh Waymo was a bet like that too. You know, we've been working on Waymo for over a decade. It was unclear that, and and we didn't do the hype approach of it's coming this year, it's coming this year. We just kept grinding away at it. And you know, we use LAR or LIDAR and we believe that was the right technology, and and now you know, Waymo is doing terrific. So one is a point of view, take the big bet, you know, stick with it with milestones, you know, along the way. And fortunately, as you said, we have a profitable business model that allows us to do those things. Two is it's you know, people just we are we we want to get the best people in the world because we're taking on some of the hardest problems. And so we spend a disproportionate amount of time recruiting the top talent, getting them early, you know, giving them the freedom to run. We do give our tech teams a lot of freedom to run, you know. So you know, Go explore uh Sundar announced this week we are exploring data centers in space. Um and so you know, part of that is you get better access to solar power as you go up into space. Sounds crazy. Who knows if it's going to work, but we have a team that believes in it and we're gonna give them the freedom to run. Um and then the other thing I'd say is we we focus it on big problems. So we're starting with the big problem and thinking about all the ways to get there. So we also have a uh a focus area on um biology and and medical sciences, and that's a an important area for us. So with that, you know, we've invested in alpha fold, which is the protein folding you know capability, uh, alpha genome, which is our genome model. We've uh alpha folding. How is it coming with alpha fold? Uh fantastic, honestly. It's it's crazy how big a step forward that was in the industry to be able to predict the structure of hundreds of millions of proteins and therefore create new druggable targets in those molecules. But we have alpha evolved for math, we have the weather first model. So um and so those are some of the core elements around it. Thank you for the question.

SPEAKER_01: 41:30

Did do you see out of I'm curious, just Alpha Fold, do you see uh line of sight to any breakthroughs on any I mean, especially where you came from knowing this space? Uh how do you see that progressing?

SPEAKER_04: 41:41

Aaron Powell Yeah, I think that'll be a big deal. In fact, we spun out a company Isomorphic Labs, you know, based on the line of sight we have around, okay, if you know the structures of these proteins, you know, so much of uh protein binding is three-dimensional. It's you need to know what it looks like to figure out if it's going to bind. And so, you know, uh we're very excited about the potential for drugs. Now that is a still a long process to get a drug to market. And so early indications are good and we're gonna keep you know sort of pushing it forward. Wow.

SPEAKER_01: 42:11

No, absolutely fast. Any any other questions? I'm so good morning.

SPEAKER_02: 42:18

Um I'm in the transportation space, and you talked about how Google is really advancing, right, from a hyperscaler perspective. What what would you say from a strategy perspective for leaders that have already invested in the Microsoft and the Azures and the AWS arenas from a cloud perspective, right? Because you're we've heard these neo clouds are popping up and everything, Google's advancement. Um what would you say would be the best strategic direction for a leader and why we should invest in Google if we have alternatives that we could use?

SPEAKER_04: 42:51

Yeah, it's a great question. And I'll start by saying, look, I um we would never ask you to be single cloud. So we'll never ask you to, and even on us, we would say it's not that it's not a good strategy to be single cloud. So you need to be multi-cloud. And the things we do, whether it's our data approach, so you know, whether it's BigQuery, we are multi-cloud and other technologies we provide you. So fundamentally, I'd start by saying, you know, we would never ask you to be multi-cloud. Now, we want you to look at Google as one of your clouds. And the reason for that are the two things I talked about. One, uh, we have, in my opinion, the strongest AI stack that exists out there today. And for capabilities that you want to launch that leverage AI, you should take a hard look at us. Because if you build applications, whether it's at the GenticLair or apps, using our models on our stack, the price performance, power benefit you get is very significant. And you know, power is a constraint in the industry. So being able to be more efficient from a power perspective, that's important. You know, price is also going to be very important and performance. So I'd say if you have applications that you know you want uh AI built in or AI-driven applications, you should look at our stack. And and it doesn't mean you have to pick the full stack from us. You know, you pick the hyperscaler, you can choose the model you want. If you want coding, for example, I'd say we're great and anthropic is great with Cloud, so you should look at that. But Cloud also runs on the Google stack. And so you can go layer by layer and say, okay, I want to pick and choose this, but uh it's hard to imagine a better stack for AI than Google Cloud. And then the third thing I'd say is we are the most open cloud. So one, we give you the most performant AI stack, you know, the no-compromises cutting-edge AI stack, and we give you full choice. Like, and that is also unique. And so you are always, you're never locked in, you're always flexible on where you want to go, and that's where I'd look at it. Ultimately, too, I'd say, look, we're early in the AI evolution, right? And so you're betting on not just where a company is today, but you're betting on their roadmap and and who do you think can go into the future. And you know, certainly even things like models, it we're spending almost 100 billion a year to keep the models going, right? Not many companies can do that and will do that again and again and again. And so let's say, you know, that's another consideration to say look at the roadmaps and look at the history and see who who do you think is going to be the right partner for you, not just for today, but a year from now, five years from now, plus. So those are things to consider.

SPEAKER_01: 45:24

Any last one last question? So uh happy to kind of wrap it up. Um my view is just incredible, you know, uh innovation coming out of Google. And just, you know, personally from uh uh a CEO perspective, a leadership perspective, uh the opportunity they have in front of them is is absolutely amazing. Um I I think you really do have to step back to really understand what we talked about, you know, one of the only like profitable large language models at this point, an organization that has spent 10 plus years building out full stack integration of AI infrastructure, which they're deciding on what they're going to do with that and how they go to market. Um we are spending some time and there's questions I'd like to ask Francis that he doesn't want me to ask right now, so I'm not going to ask those. Uh maybe I'll say, Jim.

SPEAKER_04: 46:23

And I'll say this to everyone. But uh we really value our partnership. Um and you know, it has been an important partnership for us. And I think we're honestly just at the very beginning. If you listen to everything I talked about and where you know the the our customers are gonna go, we are going to need you know ever more of this partnership than we've ever had before. So what a terrific foundation that we have built together. And I'm hugely excited about where this is gonna go.

SPEAKER_01: 46:48

Yeah, I well th thank you. And and we are the same. And I I will say that Francis uh you know, to his one point, and I'm not we're not gonna go into the detail on any of this, but he basically we had a great conversation last night. And he's one of these like, okay, go back and put a business plan together, and I want you to walk me. I'm not we're not gonna go through what that may bean, but uh it's it's it's just amazing if you truly go back and look at it and if you enjoy learning and looking at and trying to connect the dots about you know what Google did to create generative AI, the transformer technology, what they're sitting on, the engineering that they have, literally designing from a chip level to a full stack, you know, of tensure processing units, you know, that would be very similar. And in some cases, because it is an integrated stack, power, cooling, capacity, inferencing, training, these are things that this integrated stack is optimized to drive those outcomes that is very, very unique in the market. And so depending on how they decide to go to market with some of these things, uh, and then just building on the large language models and the coding, you know, that's coming out. That's uh to me, it's it's it's amazing. And if if I was in France's shoes, it'd be the biggest challenge would be prudently, thoughtfully picking and choosing where do you double and triple down your efforts because there's so much. I mean, like we were talking, they they you could carve out so many different pieces of Google, and it could be a massive company for others. And even I look at like Alpha Fold. I've been tracking and watching and listening and you know, paying attention. I think it's just amazing what is going to happen because of AI, generative AI, and especially with your background and and healthcare and you know genomics. Uh so anyway, I'm rambling on. I just want to say I I realize that. And my executive team is like, my God, somebody get them off. Uh uh, but I guess if you don't find this stuff fascinating, I I don't know what to tell you. So uh I really appreciate Francis you coming up here. Uh I do want to just say, circling all the way back to your development, and it's just fascinating when you see somebody, you know, growing up Ethiopia, moving because of political plight and you know, uh challenges and you know, going, you know, to Dubai and you know, moving your family and then to the states and all that he's doing. Uh just absolutely amazing. So thank you very much for being here.

SPEAKER_00: 49:25

Thank you. Okay, if there's one takeaway from this conversation, it's this AI leadership isn't about chasing every breakthrough, it's about choosing the right ones and building the discipline to operationalize them. Google's story reminds us that the real advantage comes from integration, the alignment of data, infrastructure, security, and people into a single intelligence system. This episode of the AI Proving Ground Podcast was co-produced by Nas Baker, Kara Kuhn, and Diane Swank. Our audio and video engineers John Knoblock. My name is Brian Felt. Thanks for listening, and we'll see you next time.