Palo Alto Networks CEO Nikesh Arora on Where AI Goes Next
How long can any conviction about AI really hold when the technology keeps changing underneath it?
At ICONIQ Frontier, Nikesh Arora, chairman and CEO of Palo Alto Networks, joined ICONIQ’s Divesh Makan to open the day with a conversation about what it means to build and lead through a period of unusually rapid change. As Nikesh put it, “Every three, four months you get a curveball.”
Drawing on a career spent navigating major shifts in technology, Nikesh shared how his own thinking has evolved as agents, open-weight models, and new approaches to inference have emerged. For founders and operators, the implication is bigger than deciding where to add AI. The more fundamental question is whether today’s products were designed for the world now taking shape, or whether some need to be rebuilt from first principles.
That idea carried into a broader discussion about why startups can have an advantage when new categories emerge, how the economics of AI may evolve, and what Nikesh looks for in founders pursuing ideas that initially sit outside the consensus.
The conversation also turned to cybersecurity, where AI is changing both the speed and nature of the threat. As models become better at finding and connecting vulnerabilities, Nikesh argued that companies will need to focus just as intensely on shortening the time it takes to detect and remove bad actors once they are inside.
Divesh Makan (00:00):
Thank you, everyone, for coming today. I'm super excited about this. On behalf of all of us at ICONIQ, this has been something that I've been looking forward to and I'm so grateful you're here and have chosen to spend that time with us today. So thank you.
(00:14):
It's been quite a year, as all of you have experienced, but this technology this last 12 months has been something that we've never experienced before. It's not like the internet. It's not like mobile or the cloud. It's evolving at a pace that none of us truly understand and can take hold of. So the leaps between these models we're seeing, stuff we can do today that we couldn't do a year ago is happening in a month, going to a week. And this is a real opportunity and a question for us.
(00:49):
When we gathered last year, the question we talked about in this venue was the first and foremost question around AI, how do you move from experimentation to real enterprise impact? So now a year later, what's our set of questions we got to think about? We have a much clearer picture of what this all looks like, but as the technology has advanced, the fundamental question has come together, which is how do you balance near-term ROI with long-term? Where does the LLM end versus the application layer that kicks in? What is simply an API core versus a real business? And who's going to own it?
(01:28):
I'm not sure we have all the answers, but how this plays out is going to have enormous implications for the companies getting built, where we put capital in. We have an incredible group today, as you spend time getting to know each other, that would value perspective on these questions and actually have a strong opinion on this. The people building the systems, bringing the technology, and some of the world's leading companies that invest in the capital behind it are sitting in this room. So we may not leave today with all the answers, but the idea is hopefully we'll walk away a lot more informed than we were at the beginning of this day.
(02:06):
This unprecedented pace of AI has helped us navigate technologies we've seen before, which is the good news. We've had this pace in other things. But very few people in the world have seen this type of torrid AI evolution through the lens of normal technology. And one of them is Nikesh Arora, who's going to be here today.
(02:27):
I've known Nikesh for 15 years and he's been a very good friend and mentor and partner of ours at ICONIQ. He's been in the center of some of the most transformation companies of our decades. He's been thinking of telecom, then the internet, mobile at Google, SoftBank, cloud, and now AI at Palo Alto Networks. And he isn't an observer. This guy is in the middle of everything. He's making massive decisions and massive changes around where this needs to go. So I can think of no better person to have join us today and kick off the day.
(02:58):
So Nikesh, welcome.
Nikesh Arora (02:59):
[inaudible 00:02:59].
Divesh Makan (02:59):
Nikesh, we've had lots of conversations, one of them around this topic of if there was something you believed about AI 12 months ago that you no longer think is true today, what does that sound like? What does it look like?
Nikesh Arora (03:22):
Well, it's hard to have beliefs about AI because every three, four months, you get a curveball. And the curveball we all got two, two and a half years ago was this notion of ChatGPT and everybody was, "Holy shit, this thing can do stuff about 70% accurately." And that's when you saw Google got caught unawares because Google was busy trying to perfect it to make it 100% right and never let it out of the research labs until it was 100% right. And suddenly, we found an exciting consumer use case and people out there trying to do 70, 80% of the stuff, which seems to be above average intelligence. Most of us don't get 80% stuff right, so this seems like pretty cool.
(04:02):
We thought we had it all figured out, we're going to have LLMs, and suddenly you saw this OpenClaw execution and suddenly you had agents that were running around. That opened all of our brains up to saying, "Holy shit, we actually don't have to be physically present to get something done. We can enable a swarm of agents and have them do shit." And that was cool.
(04:19):
And then I'm sure you have Sam later today I saw and then you have Dario and we are all trying to figure out who would be king. Is it Sam, Dario, or Dennis? And now in the last 60 days or so, the conversation has shifted to open weight models. I won't say open source, open source has always been around, but open weight is a new thing. It's like enterprises can possibly train their own models so that we don't have to be captive to the weights of the frontier LLMs. And suddenly, we all feel somehow the idea, not that most enterprises know what to do with it, but the idea that I can control my destiny seems much more exciting than the fact that I have to be captive to Sam or Dario.
(04:58):
I don't think we're done. I think this is going to go on. Like this morning, I was listening to this guy who's trying to build an inference stack, which is the cheapest stack to serve as opposed to fastest stack. And his thesis is we all got head faked into a fast low latency compute stack because we were coding and prompting, which means you want a response to come back quicker. He's saying if you want long-running agents overnight, you don't need low latency. You need cheap inference.
(05:26):
So my point is every three, four months, as people anticipate our behaviors, the new idea changes. And it's too soon to call where things are going to settle, but clearly, some things are emerging.
Divesh Makan (05:40):
Interesting. So Nikesh, if you had $1 billion to invest into-
Nikesh Arora (05:45):
You can make that happen, right? I hear you have a lot of money to invest. Why make it hypothetical? Just give me $1 billion and I'll deal with it.
Divesh Makan (05:56):
So you now have $1 billion.
Nikesh Arora (05:56):
Okay, fantastic. Wonderful.
Divesh Makan (05:57):
What would you not-
Nikesh Arora (05:58):
I'll be right back.
Divesh Makan (06:00):
Expensive coffee down the road.
Nikesh Arora (06:01):
Exactly, right.
Divesh Makan (06:02):
But what would you not invest in because you think economics are going to fall apart?
Nikesh Arora (06:07):
It's fascinating to see where the economics are showing up, right? To understand economics, you have to start where the profit pools are because we're not creating new profit pools. The consumer profit pools forever have been in three areas. They've been in subscription, advertising, or transaction. Whether you look at Apple, you look at Google, you look at Amazon. No company has built all three by the way. Usually you have a dominant economic engine. Amazon was a transaction engine. You bought shit in Amazon and they limped along in advertising and subscription. Well, now they're bigger, but still, their core business is transactions. Google on the other hand is a large advertising business, but it's not that great. It's zero transaction business and they started building a subscription business.
(06:55):
Look at it pretty much across the world. If you look at consumer business, only three sources of revenue. And there is some version of that that ends up becoming a profit pool in the consumer app side.
(07:07):
We haven't seen AI tap those profit pools aggressively yet. Yet. You can see, at the end of the day, we're competing for finite amount of consumer attention and finite in a very large sense, amount of consumer dollars and they're going to show up at some place at some point in time. I believe the next five to 10 years is more than likely that advertising and transaction will get much more efficient.
(07:35):
Today, you spend a dollar on advertising, you end up with a two to 5% conversion rate. It's highly likely with AI, your conversion rates goes to 20%, which means you get much more efficient in advertising. Conceptually, advertising profit pools could shrink because the money is going to move to transactions. So that's where the consumer stuff. Doesn't answer your question.
(07:54):
The enterprise side is, in the end, businesses spend money on technology, which eventually shows up as a profit pool for anybody serving the enterprise stack. You have this huge layer in the middle of data center businesses. I coined a new phrase last week called neoscalers because I said neoclouds are going to get cheap so people gave me shit. So I said, "Okay, neoscalers will survive." So the two or three people who are fans of Nebius and CoreWeave got placated, so they didn't kill me on Twitter yet. So we'll see.
(08:23):
But there's a whole layer in the middle which are the data center providers. I don't know how that lands. I feel like you're building inefficient compute because you have a license for land and power. And some people will make it and some people won't. I'm pretty sure that Google doesn't pay the same price to buy chips and memory that somebody building 50 megawatts does. Has to be different, right? It's a scale business. If that's true, finally somebody's building compute at one and a half times the cost of what hyperscaler is building it at.
(09:00):
Let's assume, at some point in time, the world normalizes and you can pick your horizon. I said two or three years and people beat me up, so let's say five or seven years so you don't beat me up. But in five to seven years, if compute normalizes, then the price has to go down. If the top line goes down and your cost can't go down because you already put the money in the ground, then your margins shrink. So at some point, I think the data center business becomes a regular IR business and I see company after company.
(09:29):
Company after company says, "I'm going to raise equity and I'm going to give it to the data center providers. I'm going to build a data center." That's the worst way to fund a business. The worst way to fund CapEx is equity. We're busy giving money to people building data centers as equity because somehow we believe that's going to... Equity's the lowest on the liability table when you're a high CapEx intensive business. The first guy who comes to take the chips is not the equity guy. The guy takes the chips back is Blackstone who lent you the money, or Jensen will take it back if you lent your money, anyway. I already lost my train of thought.
Divesh Makan (10:03):
No, no.
Nikesh Arora (10:05):
Where do I put my billion dollars?
Divesh Makan (10:06):
Yeah.
Nikesh Arora (10:07):
I would put my billion dollars in people helping companies do AI transformations because every CIA has no idea how to make that transformation happen, and I'm sure a bunch of people are trying to do that. And I'd put my billion dollars in somebody who was going away for three years, say, "I'm not going to have a product for three years, Nikesh, give me $1 billion." "What are you going to do?" "I'm going to take a product that is done in a certain way today. I'm going to rip it out and build it from scratch."
(10:36):
And I apologize, but I'll say one more thing. Every one of us, including people who run large businesses or products, has this existential question. The existential question is, in five years from now, do I want to be Waymo, Tesla, or Mercedes? The reason I use that analogy is Waymo was not drivable for the first nine years, but today there's no human being sitting in the car. [inaudible 00:11:04].
(11:05):
A Tesla has all of us driving them. I can do 50, 60% of FSD or 80% of FSD, but I still see people with a steering wheel and it keeps saying, "Grab your wheel," to you, but it's more AI-enabled than not.
(11:19):
And Mercedes probably AI washed. There's like a Grok agent you can talk to saying, "How's it going? How's your day? I'm AI. I'm going to help you understand your car." Okay, great.
(11:29):
Now, if you take that analogy, there are people taking software companies and they're putting little AI agents. So hey, how would you like to extract your data today? Okay, great. How can you help me? I'm going to just call my Claude AI agent and help you answer the question. So the question is which one do you want to be? I don't want to buy software which has been AI washed five years from now. I at least want to buy software that is AI-enabled or re-imagined with AI.
(11:58):
And if I look at people in the startup business, they have this drug where they've been told, get your first beta customer in six months, show me ARR in 12 months, raise more money, show me more ARR in 24 months, raise more money. And all you end up building is Frankenstein.
Divesh Makan (12:14):
That's helpful, right? So then if I push on economics, Nikesh, in this angle for a second, you've argued that Frontier Labs may be caught in this AI business model trap. Help me understand more about that. Consumers generate the data but don't pay enough, and enterprises, who can pay but demand reliability for these models, but don't yet provide that level that they need. So five years from now, who actually captures the value?
Nikesh Arora (12:40):
The good news is once you get escape velocity, you become too big to fail. And you become too big to fail... So yes, I did say there's a business model trap. The business model trap in my mind, it just follows. OpenAI, Anthropic both allow us to use pretty much their product for free if you're a consumer, at least on the chat side, not as much as the coding side, but there's a lot of free. The value I guess to them is that free allows them to train so that they get a better frontier LLM over time. So if it's free, you're the product. So we're the product. That's fine.
(13:15):
The challenge I suspect right now is 50, 60% compute is going towards serving free. As a result, when they look for money to go build their next data center from CapEx perspective, they end up charging enterprises twice as much, which means we're all paying twice as much.
(13:31):
Now, interestingly, going back to what we just talked about, that twice as much is pretty non-discriminatory. I could be running long-running persistent agents, I still pay twice as much. As opposed to I could be running low latency compute, I still pay twice as much. So I think over time, you're seeing that routers are being touted as a people who will allow us to arbitrage from a commercial pricing model. I'm not a huge fan of routers conceptually because I think at some point in time, we barely understand one model. Trying to parse your intelligence across seven different models and trying to figure out what's the right answer is holy shit.
(14:07):
Anyway, my view is over time, and you're already seeing that, over time, what's going to happen is I suspect both Dario and Sam, and even Google, should build their own routers to allow us to move between their own models which could be cheaper, because they're the most logical people to hold context and harness and memory in one place and abstract the model layer below that. I think third parties trying to abstract harnesses and context and intelligence and then trying to route them across the board to 12 different models is going to be a task from hell.
(14:41):
So I think price will come down for the enterprise world. I think we'll all pay a lot less for tokens, but I suspect we're going to end up using a lot more tokens. So I think the math shouldn't change for our friends in the LLM business. I think the maths will head fake a bunch of people who are trying to build this middle layer. Because if you're building a product purely for economic reasons, generally, it's an unstable place. It's called an arbitrage model. This industry is not stable enough to arbitrage.
Divesh Makan (15:11):
That's interesting. Pricing power, given this two-sided problem, one is that token prices need to come down-
Nikesh Arora (15:20):
They are.
Divesh Makan (15:21):
... but enterprises need to actually efficiently work and benefit from this. Where does the power sit? Do you think it sits with the LLMs ultimately or the enterprises that say, "I'm going to only pay you X"?
Nikesh Arora (15:34):
Right now, we're not seeing too many must-have use cases. The only must-have use case enterprises is coding. And the coding use case is part, I want to be really good at coding, and part is paranoia. If I can't stand up as CEO and say, "My guys are doing amazing coding and very efficient," we must not be AI first. So everybody's desperate to make sure their teams are using a lot of coding and getting it done right.
(15:59):
The pricing power is too early an argument. I don't seem to imply it's instable. We don't have a mature industry and ecosystem yet in the AI space, so everybody's trying a different set of things. The problem is you show up with one pricing model, the next startup shows a different pricing model, they're fundamentally different. The good news is they both have wonderful VCs funding them, so they'll fund their losses, we might get a better deal in the short term. So we're fine.
Divesh Makan (16:24):
You've talked about this idea of startups, since we have a lot here in the room, where you believe that startups regularly beat Palo Alto Network in new categories and you'd rather buy them than fight them. So for the founders here in the room, where are the startups beating you right now? And what does that tell us about the opportunity that you believe that they exist?
Nikesh Arora (16:42):
We're serving 80,000 customers. We have to be consistent with our customers, we can't show up every day with a new flavor for product and we can't show up with constant experimentation. If you just take a look at what's happening in AI and ignore everything else we do for now, the moment agents come out, the first thing my customer says, "Okay, wait, how am I going to find my agents? I got a bunch of them around. They're sitting in SaaS. They're sitting in coding engines. They're sitting in Bedrock and Foundry. When I find them, how am I going to be able to track them? And how am I going to figure out who they are?" That sounds like a pretty rational set of problems.
(17:15):
And if you go to Foundry, say, "Well, I can tell you my agents." They go to Marc Benioff and say, "I can tell you my agents." You go to Claude Code, "I can tell you mine." Dude, I don't need 17 different people telling me about my agents. I need to figure out how this works together. Then they come to us.
(17:27):
Now, we're not smart enough to figure out how the world's going to evolve and how this stuff is going to all line up. And something you'll find is that about 40 companies get funded in the space of agentic security. You'll fund some and a bunch of Israeli companies get funded. A bunch of your peers in the industry will fund them. You guys just afford us a very large resource lab. So I get to see all 40 companies within reason and see who's got the right idea. And I can wait for six months or 12 months or 18 months or 24 months to see good team, good execution. Have they got a cohesive end-to-end thesis and how this gets figured out?
(18:01):
I still have internal efforts, but remember, I'm one of 41 efforts. You have to have humility to believe that I will not get it right. More than likely, I won't get it right nine out of 10 times, but somebody will. But I have something they don't. I have cash. I have working cap. So I can go buy them and bring them, make them my team. That's how we do it. So you just have to have enough humility.
(18:19):
I went to Google. I remember when we told Salar Kamangar and [inaudible 00:18:24], "Let's go buy Netflix." Say, "Ah, 50 engineers and $10 million, we can replicate it. No problem." "Let's go buy WhatsApp." "That's a messaging app. I mean, that's no genius. Messaging, we can just build WhatsApp, just give me 50 more engineers, $10 million." So I haven't seen a problem that 50 engineers, $10 million can't solve in a large company. I don't know why we just don't give them 50 engineers. Actually we do a lot of times and [inaudible 00:18:44] don't solve the problem because the 50 working outside are much smarter, much more dedicated, much more resourceful in attacking problems of first principles. So just got to respect that.
Divesh Makan (18:52):
Yeah. All right, so let's talk about AI and the impact on cybersecurity for a second. This is probably something where certainly a group of us believe the next world war will be fought in some version of cyber. So let's go back for a second.
(19:07):
AI has the greatest impact on two ends of the spectrum. One is of course healthcare and changing our lives, and the other is the nature of war at the other extreme. And cyber, broadly speaking, is becoming one of the areas that are going to define AI. And if that's right, what does cybersecurity look like over the next few years, Nikesh? Maybe you can give us that landscape broadly.
Nikesh Arora (19:28):
Cybersecurity is not very complicated, since I've only spent a few years, and I'm sure there's people in the room. Rex is in here. I saw him and a bunch of other people, probably spend more time than I have. Cyber is pretty straightforward. If it's a known bad, we stop it at the perimeter. Somebody's going to attack you, we find out, we'll stop them. In our industry, the perimeter defense is what is stopping the known bad, what we call SOC or SIM analytics is analyzing a lot of data with AI and stopping the unknown bad before it becomes a problem.
(20:00):
The average time to solve the unknown bad is four days in the world. Perimeter obviously happens real time. The problem is the average time that AI is now showing us that they can concatenate vulnerabilities and attack someone is 12 minutes. So we have a bit of a gap between 12 minutes and four days, I think.
(20:21):
So this is going to be great. Everybody has to spend a lot of money to go from four days to 12 minutes. I've been trying for eight years to tell CEOs and CIOs this, but it seems like Dario is a much better communicator than I am. When he brings out something like Mythos, everybody's shit scared. CEOs want to talk about Mythos. We're an insurance company. When you're scared, we do well.
Divesh Makan (20:44):
Fair point. So tell us about Mythos a little bit. And I-
Nikesh Arora (20:47):
Who? Sorry?
Divesh Makan (20:49):
Mythos.
Nikesh Arora (20:50):
Okay, sorry.
Divesh Makan (20:51):
Go back a little bit. I remember when you had that phone call with our friends in the White House with the CEO group and Dario. Tell us a bit about what that felt like, what you learned and where things now today on that front. And how worried should we be?
Nikesh Arora (21:08):
Very worried. Always be very worried about cyber. Mythos, look, it's kind of obvious. We've spent a lot of time training models to code. It's obvious. And we've been pumping training data into Claude, into Codex and all these things and teaching how to code. Well, guess what? When you teach somebody how to do something really well, it also knows what bad looks like. Models got really good at finding bad code and it's so surprised that humans have written bad code in 20 years, many times over. AI is not lazy.
(21:40):
So now you can point Mythos at your code and it's going to find all the bad code or vulnerabilities as quickly as you want it. And now with persistence, it can keep understanding vulnerabilities and go persist and go break into somebody's infrastructure. That was the proof of concept that Mythos showed us. As a result, now there's a flurry everywhere saying, "I got to go fix all my vulnerabilities and I got to make sure that we patch them so nobody can get in."
(22:07):
You can take the paranoid view or you can take the relaxed view. If you see it, you can't unsee it. If I tell you, "Hey, CEO, you've got 300 vulnerabilities. You don't know anything about it, you didn't do your fiduciary responsibility." So I'm trying to tell people how many vulnerabilities they have so they patch them. But that's what Mythos is doing. People want to use Mythos to find all their vulnerabilities and patch them. That's the paranoid version.
(22:31):
The sort of thoughtful version says, "Well, guess what? Whether it's a vulnerability that I find today or tomorrow, I will have vulnerabilities and somebody will find them. I may not find all them and patch them. Somebody's going to find them." What happens then? Well, what happens then is they get into my infrastructure, which they can by doing other things too. I can steal your credentials. I can hold employee hostage. I can get your info. So actually, the real focus should be on take my four days to one minute because if I can make sure I find it in my infrastructure, I find bad actors in my infrastructure fast enough and I can eradicate them, then I shouldn't worry that much about the vulnerability side of it. So that's where the focus needs to be and that's where people are beginning to understand.
(23:11):
Unfortunately, that's a six-month thought process, six months purchase process, 12 month implementation. So if you're generous, if you're [inaudible 00:23:22], people will take two years to get it done. But it's surprising how the company still haven't caught onto the idea they need to go do something about it because remember, there is so much competition for technology budget in a company today. Competition in technology budget has never been higher in a company than it is today. If you look at the largest cost pools of enterprises, that's what I want efficiency at. That's where I have money, which I can then re-appropriate to different places. So you should go look at my largest cost pool and say, "How can I fix it?"
(23:50):
Now, the one large cost pool, which, contrary to popular opinion, is not going to diminish is my engineering pool. I have 6,000 engineers. Even if they get twice as effective and efficient, I'm not firing any one of them because my competition's going to be innovating faster. I want to do more things. I've never met an engineering team that doesn't have a product roadmap and things that they need to get into the next 12 months and they still don't get it done. So I have enough backlog. I'm not releasing my technical organization. I might need different kinds of people who understand AI. I'm not reducing them, but all other functions, they're fair game.
Divesh Makan (24:26):
Interesting. So then Nikesh, in the same light, when you're thinking about that cyber risk you brought up a little earlier, how does every CEO in the room think about it for their companies and what they need to do about it on a practical sense?
Nikesh Arora (24:39):
I think they should not worry about it.
Divesh Makan (24:40):
[inaudible 00:24:41].
Nikesh Arora (24:41):
They should call me. I have people. I can sort that out. They should focus on their day job because there's so much to win in their category. And this is a serious point. We all want competent people to take care of stuff that we shouldn't have to pay attention to because as CEOs, we spend our time fixing stuff that's not working great. Now, yes, and seriously, you should pay attention to cyber and make sure that your IT team and chief security officer, if you have one, are making sure that bad guys will get caught in your infrastructure fast. Don't complicate it, then hope that never happens to you.
Divesh Makan (25:19):
I like that. Okay. And then now full circle on this topic, when you're having conversations with the White House and our side of these dialogues, do they understand what's going on-
Nikesh Arora (25:32):
They understand everything.
Divesh Makan (25:33):
Okay.
Nikesh Arora (25:33):
That's how we have such a great country, so you're asking the wrong question. Do they understand what?
Divesh Makan (25:38):
What are they telling you about cyber and what they're going to do from the government side to help us manage where we are today?
Nikesh Arora (25:45):
To give them credit, they're being very thoughtful about not... Look, you can have multiple reactions. You can say, "Holy shit, these things are dangerous, stop." Well, they understand if you stop them, then we're going to lose our competitive edge vis-a-vis the rest of the world. So they're trying to figure out how to let us all keep moving at the pace we'd like to move at without creating tremendous amounts of risk or embarrassment for us and risk for companies. So I think they're striking the balance. I think your friend Sam, your friend Dario, they're all helping and there are researchers sitting in DC helping them test these models, train these models.
(26:19):
Look, AI is very powerful. I think we're all focused on cyber because that's the Mythos moment, but once you teach AI to do great drug discovery and build great drugs, that same AI will become a bioweapon-enabling mechanism. Then you're going to have a huge furor, "Oh, my God, how do you make sure there are enough guardrails that people can't use AI to build bioweapons and chemical weapons?" So for everything that you teach AI, there's a dark side. So that's a broader question. How do you make sure that you stop AI from flipping over to the dark side? I think you should save that question for the last speaker of the day.
Divesh Makan (27:00):
Okay. All right. Sam will have a fun one. All right, so let me, one final question here. Nikesh, as you... And I was half kidding, but other half not kidding. This topic of you being a serial investor in your spare time, what's exciting you right now? And how do you think about the landscape since you are actively investing as much as anyone else is?
Nikesh Arora (27:21):
I just want to get into the... I just heard there's a pre-inception round. I heard about an inception round. I asked somebody what's an inception round? Well, it's a pre-seed. There's a pre-inception round which I'd like to invest in because then you guys can keep marking me up and then I'll just sell into your series A or series B.
(27:37):
But on a more serious note, look, I said this to you. If you go back historically and think about the biggest companies that we've all seen in our lifetime, these are people who did crazy [inaudible 00:27:51] where there are not a lot of people thinking about the same thing. Do I want to invest in yet another customer transformation startup? No. Even though that's where people are going because, yes, will there be IR there? Yes, there'll be IR there.
(28:03):
I met this guy the other day who's trying to build a first bio prime of the United States. I met another guy who's trying to build a space prime for the United States. This is [inaudible 00:28:11] we don't understand. And he's saying, "Well, when you start putting data centers in space and you have other countries trying to bring down your things, carrying those things, I'm building a space prime. I can analyze all the space data." Sounds kooky, but we met Elon early days. He sounded kooky too, right? When Sam was sounding kooky when he said we're going to do $500 billion of data centers when he came three years ago to you and me talking about [inaudible 00:28:34] a $10 million company talking about a $500 billion data center. Sounded pretty radical.
(28:38):
Now, there are many founders sound like that. Some of them are kooky, so be careful. And that's where the art is. You got to figure out which ones are not kooky, which ones are real. But I think more interested in somebody who wants to do something fundamentally different from first principles as opposed to yet another version of the same out there.
Divesh Makan (28:57):
Perfect. Thank you.
Nikesh Arora (28:58):
Thank you.
Divesh Makan (28:58):
So Nikesh, thanks-
Nikesh Arora (28:59):
Have a great event, folks.
Divesh Makan (29:00):
Thank you again for spending time with us.
Disclaimer:
This presentation does not constitute investment advice or an offer to sell or a solicitation of an offer to buy any securities (including, without limitation, any investment fund or investment product referenced herein) which will only be made pursuant to definitive offering documents and subscription agreements.
Certain statements provided in the presentation are made by Nikesh Arora, CEO of Palo Alto Networks.
Mr. Arora is a client of ICONIQ. ICONIQ did not compensate Palo Alto Networks or Mr. Arora directly or indirectly for participation in this presentation, but given his role as a client of ICONIQ, Mr. Arora may be subject to certain conflicts of interest. As a client of ICONIQ, Mr. Arora’s interests may be aligned with the performance, reputation, and fundraising prospects of such funds and the adviser. Positive statements about ICONIQ could indirectly benefit Mr. Arora in his capacity as a client.
This presentation may contain forward-looking statements based on current plans, estimates and projections. You are cautioned that a number of important factors could cause actual results and outcomes to differ materially from those expressed in, or implied by, the forward-looking statements. The numbers, figures, and case studies contained in this presentation have been included for purposes of illustration only. No information is contained herein with respect to conflicts of interest, which may be significant. The portfolio companies and executives highlighted in this presentation are not representative of the entire portfolio or investing history of ICONIQ Venture and Growth.
Certain of the economic and market information contained herein may have been obtained from published sources and/or prepared by other parties. While such sources are believed to be reliable, none of ICONIQ or any of its affiliates, partners, employees, and representatives assume any responsibility for the accuracy of such information.
All of the information in this presentation is presented as of September 1, 2026 (the “Production Date”) (except as otherwise specified), is subject to change without notice, and may not be current or may have changed (possibly materially) between the Production Date and the date actually viewed by you. ICONIQ assumes no obligation to update or otherwise revise any information, projections, forecasts or estimates contained in this presentation, including any revisions to reflect changes in economic or market conditions or other circumstances arising after the date the items were made available to you or to reflect the occurrence of unanticipated events.
ICONIQ is not acting as an adviser or fiduciary in any respect in connection with providing this presentation and no relationship shall arise between you and ICONIQ as a result of this video being viewed by you.
These materials are provided for general information and discussion purposes only and may not be relied upon.

Explore ICONIQ’s proprietary research, benchmarking, and analytics shaping how leading companies scale and perform.

Get ICONIQ’s latest research, insights, and founder conversations delivered to your inbox.


