Exploring the Advancements and Challenges in AI Agent Deployment
John Nay, founder and Chief Executive Officer of Norm Ai, joins theCUBE's special presentation with NYSE Wired, focusing on the upcoming Artificial Intelligence Agent Conference 2025. Hosted by John Furrier, co-founder and co-Chief Executive Officer of SiliconANGLE Media, this insightful discussion covers the pivotal developments in AI infrastructure and the regulatory complexities faced by enterprises.
In this episode, Nay shares their expertise in regulatory AI infrastructure, particularly as it pertains to AI agent deployment in highly regulated sectors. The conversation, hosted by Furrier, delves into the evolving landscape of AI technology, compliance challenges, and the strategic initiatives underway at Norm Ai to address the pressing issues surrounding AI deployment. The discussion provides valuable insights for both technology and policy influencers.
Key takeaways from the discussion include the emphasis on the need for dynamic, real-time compliance frameworks that align with regulatory standards, as emphasized by Nay. Furthermore, the episode highlights how enterprises can leverage existing compliance structures to integrate AI technologies more effectively, offering a glimpse into the future of AI agent scalability and regulation. The conversation underscores the importance of bridging the gap between engineering, policy, and technology for sustainable AI innovation.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: AI Agent Conference. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: AI Agent Conference.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Register For theCUBE + NYSE Wired: AI Agent Conference
Please fill out the information below. You will recieve an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for theCUBE + NYSE Wired: AI Agent Conference.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: AI Agent Conference. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: AI Agent Conference.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Sign in to gain access to theCUBE + NYSE Wired: AI Agent Conference
Please sign in with LinkedIn to continue to theCUBE + NYSE Wired: AI Agent Conference. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
AGNT Podcast Ep. 3 with Gemma Allen & Raphaëlle d'Ornano
Exploring the Advancements and Challenges in AI Agent Deployment
John Nay, founder and Chief Executive Officer of Norm Ai, joins theCUBE's special presentation with NYSE Wired, focusing on the upcoming Artificial Intelligence Agent Conference 2025. Hosted by John Furrier, co-founder and co-Chief Executive Officer of SiliconANGLE Media, this insightful discussion covers the pivotal developments in AI infrastructure and the regulatory complexities faced by enterprises.
In this episode, Nay shares their expertise in regulatory AI infrastructure, particularly as it pertains to AI agent deployment in highly regulated sectors. The conversation, hosted by Furrier, delves into the evolving landscape of AI technology, compliance challenges, and the strategic initiatives underway at Norm Ai to address the pressing issues surrounding AI deployment. The discussion provides valuable insights for both technology and policy influencers.
Key takeaways from the discussion include the emphasis on the need for dynamic, real-time compliance frameworks that align with regulatory standards, as emphasized by Nay. Furthermore, the episode highlights how enterprises can leverage existing compliance structures to integrate AI technologies more effectively, offering a glimpse into the future of AI agent scalability and regulation. The conversation underscores the importance of bridging the gap between engineering, policy, and technology for sustainable AI innovation.
AGNT Podcast Ep. 3 with Gemma Allen & Raphaëlle d'Ornano
Gemma Allen
Host, theCUBE + NYSE WiredtheCUBE
HOST
Raphaelle d'Ornano
Founder & CEODecoding Discontinuity
search
Gemma Allen
>> Welcome to AGNT, the podcast where enterprise tech meets the agentic era. I'm Gemma Allen, joined by my co-host, Raphaelle d'Ornano, broadcasting from the New York Stock Exchange. On every episode, we unpack how intelligent systems are reshaping companies, markets and the way real work gets done. From Fortune 500 boardrooms to breakout upstarts, we're digging into the strategies, technologies and people to find the next chapter of AI. Let's get into it. Raphaelle, welcome back.
Raphaelle d'Ornano
>> Thank you. Great to be back.
Gemma Allen
>> What another crazy couple of weeks in the world of AI and agentic. I think for today's episode, I really want to chat to you about, I guess a question that has some uncomfortable implications, right?
Raphaelle d'Ornano
>> Okay.
Gemma Allen
>> In the agentic era, who actually owns the rent? What does that mean for some of these really high-profile, super successful companies that have had such an incredible win or decades of wins and where we're at right now? So, let's start with a very controversial company in of its own Palantir.
Raphaelle d'Ornano
>> Right.
Gemma Allen
>> You wrote a really interesting piece on this week. You have some very interesting theories and some also, I think, controversial views on Palantir's place and the moat surrounding the company. Tell me, break it down for me.
Raphaelle d'Ornano
>> So, I think, indeed, the whole software space, and when I say software like applications, infrastructure, this is all being completely redefined in terms of the rules of value by AI, and specifically agentic AI. I think what has been really interesting in the past weeks is we've seen a very clear acceleration, in case no one has noticed, for me it's been very clear, on number one, OpenAI with Frontier, which has had less attention than it should have had in my view. And of course, the fabulous success of Claude Code, with people realizing the extent of the numbers with Anthropic's recent round. So, in that you're seeing that the LLMs are for sure moving up the stack because that's how they actually build an orchestration layer and a form of defensibility as models commoditize. And then the other way, you have the traditional SaaS incumbents that are trying to own this orchestration layer to defend themselves in front of what is clearly like a new war. And so, I've tried to frame that in terms of how do you capture value in that moving parts game by looking at who owns intent, who owns the context on which your agents are grounded and who owns the workflow moats? And I've applied that to Palantir because out of all of the companies that are very well-valued, I think Palantir stands out in quite a crazy way. The stock has been extremely popular, has made a lot of Americans very fortunate with what has happened. And the stock today, despite being down almost 35% since the end of last year, which is still consequential, it's still trading at 71 times revenue.
Gemma Allen
>> Insane.
Raphaelle d'Ornano
>> You heard me correctly, 71 times revenue. And Anthropic where people are thinking that, "Oh, valuations in LLMs are huge," it's trading at 27X, And companies that are in the regular domain, like ServiceNow, which has been hit a lot, is trading at around eight times. So, just to have a sense of the magnitude. So, I wanted to apply my lenses and to say, Palantir today is priced as if they own that orchestration layer in the theory that I built, and clearly, they could only be the only winner being priced at 71 times earnings. So, my goal was to go and not criticize the company for the sake of criticizing the company, I don't think that has any interest. And by the way, the numbers are exceptional. So, maybe we start by saying that. The Palantir numbers that we've seen in their latest earnings, that's a lesson for enterprise software. This company, this is the 10th quarter by which the company has continued to grow. We have growth rates at around 70%. We have net retention, which is one of the holy grails in SaaS, at 138%. Most importantly, we have a company that is actually having strong traction in the business world, not just in the government sector, but their commercial revenue has hit over $500 million. So, let's start by this. The numbers are good, and on that I give them a good point. But if you go back to this orchestration parts and you look at what have they built, which is the ontology, the ontology is kind of the digital twin of the enterprise, which allows to make sense of the data and to translate the data into the right object to actually ground the decisions that you're making. That is more of a context substrate then where you're actually expressing user intent and making decisions. So, I can clarify that for sure. But the point is, as you think about the moving pieces today, Palantir is a very strong context layer, but as OpenAI and Anthropic or the others are going to build that orchestration layer, I think that Palantir is going to be called upon by other company's agents, like a Microsoft agent, a Salesforce agent, an Anthropic agent. And so, they're going to find themselves at the bottom, and above them, this orchestration layer which is going to capture the value. I'm not sure that they will take that position. And because I'm not sure what I'm saying is that the valuation they have today reflects a no-doubt scenario.
Gemma Allen
>> So, let's talk a second about net retention because that opens up a couple of interesting conversations for Palantir. First of all, what they do, I think, as a company is somewhat largely misunderstood. It's a company that has a very controversial focus, and I would argue for many, many reasons. But they don't collect data, they don't actually own the data. They provide a tool, like an insight tool, an ontology for other companies to basically help triangulate data and make sense of something that can seem very, very chaotic and that has many, many use cases. The one we all know of is the government use case, obviously. And right now, I mean there's some very alarming things happening in society that has Palantir suddenly become a headline name. A year or two ago, a lot of folks didn't know who Alex Karp was. They didn't really know what Palantir did. Even some of the stories that are featuring around the Palantir technology potentially brought down Bin Laden, nobody really knew that side of the story up until quite recently. But their net retention has largely been driven by their government contracts. They also have mass use cases in other parts of industry. They're a company that really affects all parts of society in terms of how data is used to drive efficiency or intel. So, when we think about the defensibility of a company like that, first of all, folks might say, "Well, if you have government contracts, the one thing we know about government is once you're in, it's very hard for them to ever get you out." Which five years ago would've been a moat factor in and of itself. And also, as a company, they seem to have really broadened their commercialability into many different parts of life and society, outside of just defense. So, what do you say to that conversation around how it impacts their position? And also, competitively, who could really compete with them?
Raphaelle d'Ornano
>> Right. So, I think to start with the part on defense, Palantir has grown a lot from its base of, again, having the right contracts with government, being a key actor in national security and I don't have a problem with that. And that has given them a lot of stickiness. It's a key factor of their moat, for sure, and it's also a factor of why the company sometimes is not loved at all, but I would say that's a separate discussion. I think that in the commercial part, which is growing very fast, they're also profoundly misunderstood because Palantir today, they just signed a big contract with Airbus, for example, French company.
Gemma Allen
>> Saw that.
Raphaelle d'Ornano
>> So, they're actually having use cases that today, we can say, for sure, are much broader than in governments though that is really not just their watch. This is where they have a core presence. And I think that what we want to do is go behind the narrative because Palantir is maybe the company that is the best positioned in terms of having a great narrative, and maybe we talk about how Palantir is actually applying the ontology to themselves. Because if there's one company that has mastered the narrative, it's Palantir. What they are saying is that, "You need Palantir in an indispensable way to make sense of your data." And you said that very correctly, they're not selling data, they're allowing companies which have these huge masses of data to actually translate that data into something that allows for a business outcome. So, it's like you go from raw data, you turn it through the ontology mechanism into a context that makes sense, and you're then able to build your workflows on top and make sense of that data because it's connected to the right objects, the right definitions in the enterprise. So, that is indispensable. And so, a lot of companies are using Palantir today because they're like, well, we know data is critical, we know context is the moat, so we need to make sense of that. So, Palantir is going to help us do that, which I think they're doing a very good job at. And some industries are very dependent on Palantir, in particular in the defense, aerospace industry. There are some sector specifics, I would say. But is that the moat? Not necessarily. Is it because you have a strong government presence that you should be valued? Again, let's go back to the levels that we're talking about. I'm not saying, "Oh, ServiceNow is valued eight times, they're valued 20 times," which would already be something that we would be debating on the show.
Gemma Allen
>> Yeah, that's true.
Raphaelle d'Ornano
>> I'm saying this company is valued 71 times, and that is after a 35% decrease because there is starting to be rationality in the market. So, it's a company that has been so successful at building this narrative. Again, it's a great company. Their numbers are great. I am not contesting any of that, and I don't want to go into all the debates that have had on the accounting manipulations. That's not what I want to do. I want to go and look at the architecture in this new stack that is being built out with the LLMs coming full speed, with the incumbents that are playing the fence because they need to transform their own positions in this new game. What is Palantir's role? And in the end, who is at the top of the ladder? Who is at the bottom of the ladder? It's exactly who wins, who... It's actually not who wins, who loses. It's who is up in the hierarchy, who is down, who is in the middle? Why? Beyond any nonsense or, "Oh, the numbers are wrong." But in terms of from an architectural position, this is really defensible. Or this may be, again, as agenetic AI builds out, be called upon other orchestrators that will use the indispensable ontology that Palantir provides, but as a component of a bigger orchestration ensemble, that's how I'm thinking about it.
Gemma Allen
>> How do you envision a scenario where they are acting as an orchestrator? What does that look like? Does that look like they are controlling workflows? They are mission control within an agentic workflow for some of their largest strategic customers? Bring it to life for me.
Raphaelle d'Ornano
>> That's exactly what they're doing. So, they have a customer base that is, of course, much more narrow and a distribution base that is narrow than the traditional SaaS incumbents, for sure. But in the companies and in the sectors in which they're heavily penetrated, and again, defense being one of them, for sure, they are that orchestrator end-to-end and I'm not contesting that in any kind of way. I think in the defense sector, I don't think it's going to be, and maybe I'm wrong, I don't think it's going to be OpenAI or Anthropic that is going to be that orchestrator that has access to your intent that then translates that into the right set of workflows, building on super-sensitive context. So, I think they're well positioned there and they are effectively that orchestrator. But again, from a valuation standpoint, this is a company worth like $218 billion. The government part is not the only component. They have this whole commercial business. And that commercial business, which is critical in the global valuation we're talking about, I do not think that that's where they will be the orchestrator. And again, I do not have a crystal ball, and maybe I will be wrong and I'm very happy to be wrong. My whole point is in their valuation today, it's assuming that this is already the case and there's no debate, like we shouldn't even be having this discussion. The valuation is saying Palantir is the orchestrator in all of its domains, that's how it is and that's what is encapsulated in that kind of valuation, which again, Anthropic doesn't have. And Anthropic, it's very clear with Claude Code and what they're doing that through coding being the wedge and coding being the entry into the enterprise, they're moving up the stack and starting to claim that orchestration position, which they have been doing over the past year in quite a successful way. And it's one example amongst others. And again, Frontier from OpenAI, I mean, I was looking at some notes and let me find a quote... If I can find the quote, which is not the case... Okay-
Gemma Allen
>> Well, let's talk for a second about the threat factor of Anthropic and OpenAI. Let's talk about those separately and then let's compare them a little bit, right? Because I guess one could argue that if somebody is coming to eat anyone's lunch in any industry right now, including Palantir, it might be frontier LLMs, like Anthropic and OpenAI. They've both had very interesting weeks. The OpenAI suspected raise. Anthropic have just had a huge series two. But just this morning, I mean, there's a lot of belief in the market that that OpenAI are also about to raise a huge round that's going to value the company at $750 billion. Last round of funding valued them at $500 billion, which is just-
Raphaelle d'Ornano
>> So cheap....
Gemma Allen
>> insane. It's hard to even comprehend valuations like that. We had Altman and Dario on stage in India this week. I don't know if you saw that video where-
Raphaelle d'Ornano
>> Of course. Of course. They don't like each other....
Gemma Allen
>> where they refused to hold hands. It was actually almost, in a way... I thought it was, honestly, AI-generated. No, it was real. It was very, very real. The BBC reported it, but things are really, really heating up. OpenAI had a very interesting acqui-hire. They hired Peter Steinberger. Again, what a just fascinating case of a guy who really just considered himself a builder. His last company was rendering for PDF documents, was an SDK for PDFs, then built OpenClaw, which obviously became... We had a whole debate around IP and then suddenly Altman hires him. Now, he's valued at God knows how much money. I mean it wasn't disclosed, but he actually said in an interview, "I just always really wanted to join a big company. I always wanted to work at a big company." And he said, "I want to create AI that my mom can use." I just thought those two things were so interesting for somebody who's now become, overnight, one of the most prolific names in tech. But when you look at the war, the race between OpenAI and Anthropic, was that a very good call by Altman? Possibly. Was the conversation that happened around how Anthropic approached him from an IP perspective, obviously making it quite adversarial? Again, was that the right call? Maybe not, but who knows? It feels as though everything is so fickle and moving so fast that what's right today is wrong tomorrow anyway. So, what are your thoughts on this race between these two players?
Raphaelle d'Ornano
>> So, first I would say that I think up to a certain point, we've been putting the AI labs kind of in the same box, and we're talking about Anthropic and we're talking about OpenAI, but we could talk about xAI, which is also valued at hugely big numbers. We could talk about Google with Gemini, et cetera, and all of the others that are Microsoft, which is now building its own models, et cetera. So, these business models are profoundly different. And the way that they're building moats, and I like the notion of moats and how do you build competitive advantages that are durable in this super fast-moving. They both have a very different approach to building a moat and people are putting them in the same basket, which is wrong, I think. So, OpenAI, their moat is not that clear, and I've said that openly many times. It's a great company, they're growing very fast, but how are they going to build durable revenue and durable competitive advantages? Now, we have advertising as a monetization engine. I mean, that was quite criticized, but I mean that's going on from a consumer perspective. They've really penetrated the enterprise and are starting to penetrate the enterprise, which is not at all a surprise to me, in the sense that, of course, that is their ambition. Let's see if they're actually able to penetrate the enterprise because penetrating the enterprise is actually very hard. And I think that the incumbents, like the Salesforces, ServiceNow, Microsoft have a durable advantage. They need to accelerate, but I think they're there. So, it's not clear what OpenAI really is doing. Clearly, getting OpenClaw is going to be an accelerating factor. I mean, we are seeing people that are now able to build their personal assistants seven or eight... I've seen someone who's built seven to eight personal assistants in one weekend. He bought a separate computer to avoid any security risks. We need to talk about the security risks. It is not without risk. So, he literally bought a new computer and built, I think eight personal assistants, one to help him with PR, one to help him with appointments, blah, blah, blah, all of that through OpenClaw. And he's very happy up until now. So, that's more on the personal side. Is that of how you're going to be running the enterprise? I don't know. Is everyone going to be building 8 or 10 agents to do everything they have to do in their day? Maybe, maybe not. Again, I do not have the answer, but back to Anthropic. Anthropic, they have their race to orchestration. They are very clear about the enterprise, they are very clear around security, et cetera, and that is giving them problems with what is happening with the Pentagon right now. So-
Gemma Allen
>> Yeah, exactly. There's a lot off-
Raphaelle d'Ornano
>> They have their own challenges right now and sticking with their approach, so let's see where that goes. And I'm not saying that to defend Dario in any way, but again, OpenClaw is a great acquisition. Is it critical in Anthropic building its moat? And again, maybe I'm wrong. I don't think that it's . While saying that, OpenClaw is great and that it's going to deliver tremendous business value for the people that are using it. So, again, it's not like OpenAI versus Anthropic, and every time Anthropic does something it's against OpenAI. We should stop opposing them. These companies are having two very different strategies, and it's like if we're saying like, "Oh, Salesforce and Microsoft, they're exactly the same companies." No, they have things in common, they're competitors on many aspects, but they have their different strategies and there's lot of nuance in what is being built.
Gemma Allen
>> And some of the potential criticism of Anthropic... I'm a big fan of Anthropic, right? I think it's a phenomenal product, and Claude is a phenomenal dev tool, but is that they're a brilliant brain, right? It's a brilliant brain for enterprise, but it's a component. Whereas OpenAI, and I agree with you, it's not really clear what the strategy is because they're all things to all people. And maybe a year or two ago, people would be more skeptical of that than it seems they are right now. But when you think about Anthropic and the Anthropic defensibility, let's just take the example of the DoD scenario with Pete Hegsworth potentially unleashing a whole tirade of accusations against Anthropic. If it was to be the case that they were to say to any of their FedRAMP clients or customers, "If you have Anthropic in your workflow, you're no longer FedRAMP-approved." The impacts of that on a company like Anthropic cannot be underestimated. So, in that component model where they are acting like a very, very incredible cog within a wheel, what are your thoughts on the longer term strategic defensibility?
Raphaelle d'Ornano
>> Look, I think that it was interesting to see the latest round and how it was oversubscribed, which I can understand for sure. These companies are very risky business models. It doesn't mean that it's good or that it's bad. I'm not giving any judgment here. These business models are profoundly different than the SaaS 2.0 wave that characterized the previous era of technology. Why? Because of trading computes. You have companies that are being built that are spending billions of dollars according to the scaling laws that we think are holding for now, but which could be challenged at any point. Remember that DeepSeek V4 is supposed to happen at any point now. And again, maybe it might not have any implication, but in any case, as Dario said in a very recent podcast, there is a scenario by which you're buying a lot of compute, it's costing you a ton, your revenue is not where you want it to be and everything blows up. I mean, the first risk, it's that you have a mismatch that is a very costly mismatch between the actual revenue you're generating, be it enterprise consumer of any revenue and your costs. I would say that's the number one thing. Those companies are growing at a rate that we have never seen in history. I was telling that Palantir was growing 70% per year, which is huge by any standards. Anthropic is now $14 billion in ARR. It was $1 billion not even two years ago. So, think about the scale that is happening. So, those companies have risks. Again, that is not good or bad. Let's take away that judgment. There are specific animals that need to be understood in the current paradigm and those are risky companies. It may turn out to be amazing, and again, you do not have this kind of growth rates very often, but it's not for everyone. Not everyone should be investing in OpenAI or Anthropic. And it's exactly like we had talked about MiniMax two weeks ago. MiniMax is just doing great. You have an open-source model. You're not sure about the monetization. You have commoditization that could be happening from your own national competitors at any point, what is the sustainability of the valuation? MiniMax has gone up, up. We don't know if it's going to hold. So, those are very exciting companies that I love exploring and understanding their moats because it's generally different and interesting and something happens almost every day, I think it's fair to say that, but buckle up, right? Buckle up.
Gemma Allen
>> Buckle up. Well, to finish, so you said that not everyone should invest in Anthropic and OpenAI, it's not for the faint-hearted for sure. Do you think we're going to get to a stage this year where we're going to be able? Do you think those companies are going to float this year? What are your thoughts?
Raphaelle d'Ornano
>> I mean, I'm pretty sure they're both going to be doing two major rounds, very likely. And I think that we should have at least one of those two listings in H2, maybe both. I believe more in one. Let's see. And again, we have xAI in June, Elon wants to do that around his birthday, which is in June. So, H1 we have Elon. H2 we have, what, maybe Anthropic first and then OpenAI January, February, 2027 maybe Q4. I don't know. I think there's going to be a private round. I think both private rounds are going to be oversubscribed. We're going to hear the same thing every time. "Oh, they're having trouble fundraising. It's not easy. Oh, but the fundraising did go through."
So it's going to be the same story. We didn't hear that for Anthropic, but we're going to hear it for OpenAI. It's going to be difficult. Then, you're going to have a huge fundraising where people are going to be like, "Oh, my God. I didn't know that that was possible in the private markets." And then, people are going to go back to normal and then we'll have an IPO with many questions that are going to be very interesting to address. Because remember that in these companies, just looking at the numbers and putting that in a spreadsheet is not going to help you in any way to analyze all of the single points of failures that could make the model completely blow up.
Gemma Allen
>> Wow.
Raphaelle d'Ornano
>> So, buckle up once again for the public investors that are going to look at these companies when it actually... A bumpy year ahead. But again, it's interesting and it's what we like in doing this.
Gemma Allen
>> Well, buckle up. Let's leave it with that, Raphaelle. Can't wait to catch up with you in two weeks.
AGNT Podcast Ep. 3 with Gemma Allen & Raphaëlle d'Ornano
search
Gemma Allen
>> Welcome to AGNT, the podcast where enterprise tech meets the agentic era. I'm Gemma Allen, joined by my co-host, Raphaelle d'Ornano, broadcasting from the New York Stock Exchange. On every episode, we unpack how intelligent systems are reshaping companies, markets and the way real work gets done. From Fortune 500 boardrooms to breakout upstarts, we're digging into the strategies, technologies and people to find the next chapter of AI. Let's get into it. Raphaelle, welcome back.
Raphaelle d'Ornano
>> Thank you. Great to be back.
Gemma Allen
>> What another crazy couple of weeks in the world of AI and agentic. I think for today's episode, I really want to chat to you about, I guess a question that has some uncomfortable implications, right?
Raphaelle d'Ornano
>> Okay.
Gemma Allen
>> In the agentic era, who actually owns the rent? What does that mean for some of these really high-profile, super successful companies that have had such an incredible win or decades of wins and where we're at right now? So, let's start with a very controversial company in of its own Palantir.
Raphaelle d'Ornano
>> Right.
Gemma Allen
>> You wrote a really interesting piece on this week. You have some very interesting theories and some also, I think, controversial views on Palantir's place and the moat surrounding the company. Tell me, break it down for me.
Raphaelle d'Ornano
>> So, I think, indeed, the whole software space, and when I say software like applications, infrastructure, this is all being completely redefined in terms of the rules of value by AI, and specifically agentic AI. I think what has been really interesting in the past weeks is we've seen a very clear acceleration, in case no one has noticed, for me it's been very clear, on number one, OpenAI with Frontier, which has had less attention than it should have had in my view. And of course, the fabulous success of Claude Code, with people realizing the extent of the numbers with Anthropic's recent round. So, in that you're seeing that the LLMs are for sure moving up the stack because that's how they actually build an orchestration layer and a form of defensibility as models commoditize. And then the other way, you have the traditional SaaS incumbents that are trying to own this orchestration layer to defend themselves in front of what is clearly like a new war. And so, I've tried to frame that in terms of how do you capture value in that moving parts game by looking at who owns intent, who owns the context on which your agents are grounded and who owns the workflow moats? And I've applied that to Palantir because out of all of the companies that are very well-valued, I think Palantir stands out in quite a crazy way. The stock has been extremely popular, has made a lot of Americans very fortunate with what has happened. And the stock today, despite being down almost 35% since the end of last year, which is still consequential, it's still trading at 71 times revenue.
Gemma Allen
>> Insane.
Raphaelle d'Ornano
>> You heard me correctly, 71 times revenue. And Anthropic where people are thinking that, "Oh, valuations in LLMs are huge," it's trading at 27X, And companies that are in the regular domain, like ServiceNow, which has been hit a lot, is trading at around eight times. So, just to have a sense of the magnitude. So, I wanted to apply my lenses and to say, Palantir today is priced as if they own that orchestration layer in the theory that I built, and clearly, they could only be the only winner being priced at 71 times earnings. So, my goal was to go and not criticize the company for the sake of criticizing the company, I don't think that has any interest. And by the way, the numbers are exceptional. So, maybe we start by saying that. The Palantir numbers that we've seen in their latest earnings, that's a lesson for enterprise software. This company, this is the 10th quarter by which the company has continued to grow. We have growth rates at around 70%. We have net retention, which is one of the holy grails in SaaS, at 138%. Most importantly, we have a company that is actually having strong traction in the business world, not just in the government sector, but their commercial revenue has hit over $500 million. So, let's start by this. The numbers are good, and on that I give them a good point. But if you go back to this orchestration parts and you look at what have they built, which is the ontology, the ontology is kind of the digital twin of the enterprise, which allows to make sense of the data and to translate the data into the right object to actually ground the decisions that you're making. That is more of a context substrate then where you're actually expressing user intent and making decisions. So, I can clarify that for sure. But the point is, as you think about the moving pieces today, Palantir is a very strong context layer, but as OpenAI and Anthropic or the others are going to build that orchestration layer, I think that Palantir is going to be called upon by other company's agents, like a Microsoft agent, a Salesforce agent, an Anthropic agent. And so, they're going to find themselves at the bottom, and above them, this orchestration layer which is going to capture the value. I'm not sure that they will take that position. And because I'm not sure what I'm saying is that the valuation they have today reflects a no-doubt scenario.
Gemma Allen
>> So, let's talk a second about net retention because that opens up a couple of interesting conversations for Palantir. First of all, what they do, I think, as a company is somewhat largely misunderstood. It's a company that has a very controversial focus, and I would argue for many, many reasons. But they don't collect data, they don't actually own the data. They provide a tool, like an insight tool, an ontology for other companies to basically help triangulate data and make sense of something that can seem very, very chaotic and that has many, many use cases. The one we all know of is the government use case, obviously. And right now, I mean there's some very alarming things happening in society that has Palantir suddenly become a headline name. A year or two ago, a lot of folks didn't know who Alex Karp was. They didn't really know what Palantir did. Even some of the stories that are featuring around the Palantir technology potentially brought down Bin Laden, nobody really knew that side of the story up until quite recently. But their net retention has largely been driven by their government contracts. They also have mass use cases in other parts of industry. They're a company that really affects all parts of society in terms of how data is used to drive efficiency or intel. So, when we think about the defensibility of a company like that, first of all, folks might say, "Well, if you have government contracts, the one thing we know about government is once you're in, it's very hard for them to ever get you out." Which five years ago would've been a moat factor in and of itself. And also, as a company, they seem to have really broadened their commercialability into many different parts of life and society, outside of just defense. So, what do you say to that conversation around how it impacts their position? And also, competitively, who could really compete with them?
Raphaelle d'Ornano
>> Right. So, I think to start with the part on defense, Palantir has grown a lot from its base of, again, having the right contracts with government, being a key actor in national security and I don't have a problem with that. And that has given them a lot of stickiness. It's a key factor of their moat, for sure, and it's also a factor of why the company sometimes is not loved at all, but I would say that's a separate discussion. I think that in the commercial part, which is growing very fast, they're also profoundly misunderstood because Palantir today, they just signed a big contract with Airbus, for example, French company.
Gemma Allen
>> Saw that.
Raphaelle d'Ornano
>> So, they're actually having use cases that today, we can say, for sure, are much broader than in governments though that is really not just their watch. This is where they have a core presence. And I think that what we want to do is go behind the narrative because Palantir is maybe the company that is the best positioned in terms of having a great narrative, and maybe we talk about how Palantir is actually applying the ontology to themselves. Because if there's one company that has mastered the narrative, it's Palantir. What they are saying is that, "You need Palantir in an indispensable way to make sense of your data." And you said that very correctly, they're not selling data, they're allowing companies which have these huge masses of data to actually translate that data into something that allows for a business outcome. So, it's like you go from raw data, you turn it through the ontology mechanism into a context that makes sense, and you're then able to build your workflows on top and make sense of that data because it's connected to the right objects, the right definitions in the enterprise. So, that is indispensable. And so, a lot of companies are using Palantir today because they're like, well, we know data is critical, we know context is the moat, so we need to make sense of that. So, Palantir is going to help us do that, which I think they're doing a very good job at. And some industries are very dependent on Palantir, in particular in the defense, aerospace industry. There are some sector specifics, I would say. But is that the moat? Not necessarily. Is it because you have a strong government presence that you should be valued? Again, let's go back to the levels that we're talking about. I'm not saying, "Oh, ServiceNow is valued eight times, they're valued 20 times," which would already be something that we would be debating on the show.
Gemma Allen
>> Yeah, that's true.
Raphaelle d'Ornano
>> I'm saying this company is valued 71 times, and that is after a 35% decrease because there is starting to be rationality in the market. So, it's a company that has been so successful at building this narrative. Again, it's a great company. Their numbers are great. I am not contesting any of that, and I don't want to go into all the debates that have had on the accounting manipulations. That's not what I want to do. I want to go and look at the architecture in this new stack that is being built out with the LLMs coming full speed, with the incumbents that are playing the fence because they need to transform their own positions in this new game. What is Palantir's role? And in the end, who is at the top of the ladder? Who is at the bottom of the ladder? It's exactly who wins, who... It's actually not who wins, who loses. It's who is up in the hierarchy, who is down, who is in the middle? Why? Beyond any nonsense or, "Oh, the numbers are wrong." But in terms of from an architectural position, this is really defensible. Or this may be, again, as agenetic AI builds out, be called upon other orchestrators that will use the indispensable ontology that Palantir provides, but as a component of a bigger orchestration ensemble, that's how I'm thinking about it.
Gemma Allen
>> How do you envision a scenario where they are acting as an orchestrator? What does that look like? Does that look like they are controlling workflows? They are mission control within an agentic workflow for some of their largest strategic customers? Bring it to life for me.
Raphaelle d'Ornano
>> That's exactly what they're doing. So, they have a customer base that is, of course, much more narrow and a distribution base that is narrow than the traditional SaaS incumbents, for sure. But in the companies and in the sectors in which they're heavily penetrated, and again, defense being one of them, for sure, they are that orchestrator end-to-end and I'm not contesting that in any kind of way. I think in the defense sector, I don't think it's going to be, and maybe I'm wrong, I don't think it's going to be OpenAI or Anthropic that is going to be that orchestrator that has access to your intent that then translates that into the right set of workflows, building on super-sensitive context. So, I think they're well positioned there and they are effectively that orchestrator. But again, from a valuation standpoint, this is a company worth like $218 billion. The government part is not the only component. They have this whole commercial business. And that commercial business, which is critical in the global valuation we're talking about, I do not think that that's where they will be the orchestrator. And again, I do not have a crystal ball, and maybe I will be wrong and I'm very happy to be wrong. My whole point is in their valuation today, it's assuming that this is already the case and there's no debate, like we shouldn't even be having this discussion. The valuation is saying Palantir is the orchestrator in all of its domains, that's how it is and that's what is encapsulated in that kind of valuation, which again, Anthropic doesn't have. And Anthropic, it's very clear with Claude Code and what they're doing that through coding being the wedge and coding being the entry into the enterprise, they're moving up the stack and starting to claim that orchestration position, which they have been doing over the past year in quite a successful way. And it's one example amongst others. And again, Frontier from OpenAI, I mean, I was looking at some notes and let me find a quote... If I can find the quote, which is not the case... Okay-
Gemma Allen
>> Well, let's talk for a second about the threat factor of Anthropic and OpenAI. Let's talk about those separately and then let's compare them a little bit, right? Because I guess one could argue that if somebody is coming to eat anyone's lunch in any industry right now, including Palantir, it might be frontier LLMs, like Anthropic and OpenAI. They've both had very interesting weeks. The OpenAI suspected raise. Anthropic have just had a huge series two. But just this morning, I mean, there's a lot of belief in the market that that OpenAI are also about to raise a huge round that's going to value the company at $750 billion. Last round of funding valued them at $500 billion, which is just-
Raphaelle d'Ornano
>> So cheap....
Gemma Allen
>> insane. It's hard to even comprehend valuations like that. We had Altman and Dario on stage in India this week. I don't know if you saw that video where-
Raphaelle d'Ornano
>> Of course. Of course. They don't like each other....
Gemma Allen
>> where they refused to hold hands. It was actually almost, in a way... I thought it was, honestly, AI-generated. No, it was real. It was very, very real. The BBC reported it, but things are really, really heating up. OpenAI had a very interesting acqui-hire. They hired Peter Steinberger. Again, what a just fascinating case of a guy who really just considered himself a builder. His last company was rendering for PDF documents, was an SDK for PDFs, then built OpenClaw, which obviously became... We had a whole debate around IP and then suddenly Altman hires him. Now, he's valued at God knows how much money. I mean it wasn't disclosed, but he actually said in an interview, "I just always really wanted to join a big company. I always wanted to work at a big company." And he said, "I want to create AI that my mom can use." I just thought those two things were so interesting for somebody who's now become, overnight, one of the most prolific names in tech. But when you look at the war, the race between OpenAI and Anthropic, was that a very good call by Altman? Possibly. Was the conversation that happened around how Anthropic approached him from an IP perspective, obviously making it quite adversarial? Again, was that the right call? Maybe not, but who knows? It feels as though everything is so fickle and moving so fast that what's right today is wrong tomorrow anyway. So, what are your thoughts on this race between these two players?
Raphaelle d'Ornano
>> So, first I would say that I think up to a certain point, we've been putting the AI labs kind of in the same box, and we're talking about Anthropic and we're talking about OpenAI, but we could talk about xAI, which is also valued at hugely big numbers. We could talk about Google with Gemini, et cetera, and all of the others that are Microsoft, which is now building its own models, et cetera. So, these business models are profoundly different. And the way that they're building moats, and I like the notion of moats and how do you build competitive advantages that are durable in this super fast-moving. They both have a very different approach to building a moat and people are putting them in the same basket, which is wrong, I think. So, OpenAI, their moat is not that clear, and I've said that openly many times. It's a great company, they're growing very fast, but how are they going to build durable revenue and durable competitive advantages? Now, we have advertising as a monetization engine. I mean, that was quite criticized, but I mean that's going on from a consumer perspective. They've really penetrated the enterprise and are starting to penetrate the enterprise, which is not at all a surprise to me, in the sense that, of course, that is their ambition. Let's see if they're actually able to penetrate the enterprise because penetrating the enterprise is actually very hard. And I think that the incumbents, like the Salesforces, ServiceNow, Microsoft have a durable advantage. They need to accelerate, but I think they're there. So, it's not clear what OpenAI really is doing. Clearly, getting OpenClaw is going to be an accelerating factor. I mean, we are seeing people that are now able to build their personal assistants seven or eight... I've seen someone who's built seven to eight personal assistants in one weekend. He bought a separate computer to avoid any security risks. We need to talk about the security risks. It is not without risk. So, he literally bought a new computer and built, I think eight personal assistants, one to help him with PR, one to help him with appointments, blah, blah, blah, all of that through OpenClaw. And he's very happy up until now. So, that's more on the personal side. Is that of how you're going to be running the enterprise? I don't know. Is everyone going to be building 8 or 10 agents to do everything they have to do in their day? Maybe, maybe not. Again, I do not have the answer, but back to Anthropic. Anthropic, they have their race to orchestration. They are very clear about the enterprise, they are very clear around security, et cetera, and that is giving them problems with what is happening with the Pentagon right now. So-
Gemma Allen
>> Yeah, exactly. There's a lot off-
Raphaelle d'Ornano
>> They have their own challenges right now and sticking with their approach, so let's see where that goes. And I'm not saying that to defend Dario in any way, but again, OpenClaw is a great acquisition. Is it critical in Anthropic building its moat? And again, maybe I'm wrong. I don't think that it's . While saying that, OpenClaw is great and that it's going to deliver tremendous business value for the people that are using it. So, again, it's not like OpenAI versus Anthropic, and every time Anthropic does something it's against OpenAI. We should stop opposing them. These companies are having two very different strategies, and it's like if we're saying like, "Oh, Salesforce and Microsoft, they're exactly the same companies." No, they have things in common, they're competitors on many aspects, but they have their different strategies and there's lot of nuance in what is being built.
Gemma Allen
>> And some of the potential criticism of Anthropic... I'm a big fan of Anthropic, right? I think it's a phenomenal product, and Claude is a phenomenal dev tool, but is that they're a brilliant brain, right? It's a brilliant brain for enterprise, but it's a component. Whereas OpenAI, and I agree with you, it's not really clear what the strategy is because they're all things to all people. And maybe a year or two ago, people would be more skeptical of that than it seems they are right now. But when you think about Anthropic and the Anthropic defensibility, let's just take the example of the DoD scenario with Pete Hegsworth potentially unleashing a whole tirade of accusations against Anthropic. If it was to be the case that they were to say to any of their FedRAMP clients or customers, "If you have Anthropic in your workflow, you're no longer FedRAMP-approved." The impacts of that on a company like Anthropic cannot be underestimated. So, in that component model where they are acting like a very, very incredible cog within a wheel, what are your thoughts on the longer term strategic defensibility?
Raphaelle d'Ornano
>> Look, I think that it was interesting to see the latest round and how it was oversubscribed, which I can understand for sure. These companies are very risky business models. It doesn't mean that it's good or that it's bad. I'm not giving any judgment here. These business models are profoundly different than the SaaS 2.0 wave that characterized the previous era of technology. Why? Because of trading computes. You have companies that are being built that are spending billions of dollars according to the scaling laws that we think are holding for now, but which could be challenged at any point. Remember that DeepSeek V4 is supposed to happen at any point now. And again, maybe it might not have any implication, but in any case, as Dario said in a very recent podcast, there is a scenario by which you're buying a lot of compute, it's costing you a ton, your revenue is not where you want it to be and everything blows up. I mean, the first risk, it's that you have a mismatch that is a very costly mismatch between the actual revenue you're generating, be it enterprise consumer of any revenue and your costs. I would say that's the number one thing. Those companies are growing at a rate that we have never seen in history. I was telling that Palantir was growing 70% per year, which is huge by any standards. Anthropic is now $14 billion in ARR. It was $1 billion not even two years ago. So, think about the scale that is happening. So, those companies have risks. Again, that is not good or bad. Let's take away that judgment. There are specific animals that need to be understood in the current paradigm and those are risky companies. It may turn out to be amazing, and again, you do not have this kind of growth rates very often, but it's not for everyone. Not everyone should be investing in OpenAI or Anthropic. And it's exactly like we had talked about MiniMax two weeks ago. MiniMax is just doing great. You have an open-source model. You're not sure about the monetization. You have commoditization that could be happening from your own national competitors at any point, what is the sustainability of the valuation? MiniMax has gone up, up. We don't know if it's going to hold. So, those are very exciting companies that I love exploring and understanding their moats because it's generally different and interesting and something happens almost every day, I think it's fair to say that, but buckle up, right? Buckle up.
Gemma Allen
>> Buckle up. Well, to finish, so you said that not everyone should invest in Anthropic and OpenAI, it's not for the faint-hearted for sure. Do you think we're going to get to a stage this year where we're going to be able? Do you think those companies are going to float this year? What are your thoughts?
Raphaelle d'Ornano
>> I mean, I'm pretty sure they're both going to be doing two major rounds, very likely. And I think that we should have at least one of those two listings in H2, maybe both. I believe more in one. Let's see. And again, we have xAI in June, Elon wants to do that around his birthday, which is in June. So, H1 we have Elon. H2 we have, what, maybe Anthropic first and then OpenAI January, February, 2027 maybe Q4. I don't know. I think there's going to be a private round. I think both private rounds are going to be oversubscribed. We're going to hear the same thing every time. "Oh, they're having trouble fundraising. It's not easy. Oh, but the fundraising did go through."
So it's going to be the same story. We didn't hear that for Anthropic, but we're going to hear it for OpenAI. It's going to be difficult. Then, you're going to have a huge fundraising where people are going to be like, "Oh, my God. I didn't know that that was possible in the private markets." And then, people are going to go back to normal and then we'll have an IPO with many questions that are going to be very interesting to address. Because remember that in these companies, just looking at the numbers and putting that in a spreadsheet is not going to help you in any way to analyze all of the single points of failures that could make the model completely blow up.
Gemma Allen
>> Wow.
Raphaelle d'Ornano
>> So, buckle up once again for the public investors that are going to look at these companies when it actually... A bumpy year ahead. But again, it's interesting and it's what we like in doing this.
Gemma Allen
>> Well, buckle up. Let's leave it with that, Raphaelle. Can't wait to catch up with you in two weeks.
