Raphaëlle d'Ornano of Decoding Discontinuity, founder and CEO, joins Gemma Allen of theCUBE and NYSE Wired, host, for a live discussion from the New York Stock Exchange. The conversation examines trends at the intersection of enterprise software and agentic artificial intelligence and analyzes orchestration, inference compute and market dynamics.
d'Ornano provides expertise in strategic AI adoption and market frameworks and they analyze topics such as Figma's Anthropic partnership, Salesforce's Agentforce, Nvidia and the inference economy, GPU supply dynamics and evolving SaaS defensibility. Allen delivers live reporting and highlights growth and the market complexities around Nvidia and inference-driven compute demand.
Key takeaways include d'Ornano's view that first-mover advantage and developer adoption are critical to securing an orchestration role, with coding serving as a wedge into enterprise platforms. theCUBE Research analysts recommend mapping systems of record, evaluating integration paths into the agentic stack and monitoring AI lab roadmaps and financing dynamics to inform enterprise strategy and vendor selection.
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. 5 with Gemma Allen & Raphaëlle d'Ornano
Raphaëlle d'Ornano of Decoding Discontinuity, founder and CEO, joins Gemma Allen of theCUBE and NYSE Wired, host, for a live discussion from the New York Stock Exchange. The conversation examines trends at the intersection of enterprise software and agentic artificial intelligence and analyzes orchestration, inference compute and market dynamics.
d'Ornano provides expertise in strategic AI adoption and market frameworks and they analyze topics such as Figma's Anthropic partnership, Salesforce's Agentforce, Nvidia and the inference economy, GPU supply dynamics and evolving SaaS defensibility. Allen delivers live reporting and highlights growth and the market complexities around Nvidia and inference-driven compute demand.
Key takeaways include d'Ornano's view that first-mover advantage and developer adoption are critical to securing an orchestration role, with coding serving as a wedge into enterprise platforms. theCUBE Research analysts recommend mapping systems of record, evaluating integration paths into the agentic stack and monitoring AI lab roadmaps and financing dynamics to inform enterprise strategy and vendor selection.
AGNT Podcast Ep. 5 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. And 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 defining the next chapter of AI. Let's get into it. Raphaelle, welcome back to the NYSE and to theCUBE and NYSE Wired. We were at San Jose and Nvidia GTC last week. It was a wild week, and I know you had a lot to say about it on Substack and on your socials this weekend. So maybe let's start there. Let's start with just the absolute madness of this moment. Like what are your thoughts on some of the learnings and conversations from Nvidia?
Raphaelle d'Ornano
>> Well, I mean, I think this conference every year, it's kind of like the moment when everyone is trying to understand, is this really real? Is this going to continue at this speed? So I think that it has taken a position in this space that goes way beyond Nvidia. I mean, Nvidia goes beyond itself. And so I think GTC is a very key moment in like, okay, where are we? And Jensen has been very clear over the past month on inference is the big deal. Everything is going to be compute, like everything, everything, everything. And I think that what we saw at that conference was that there's going to be way more chips than we thought there were going to be. I mean, I'm not surprised. I don't think he would be surprised. Many people are not surprised. But it's like there's going to be way more as AI gets adopted as agents proliferate with the mold book continuing and reverberating. And one announcement that I really liked that he made at the conference, which was during the second hour of this monster keynote, where he announced that they were going to be kind of the hyperscaler of agentic AI. He didn't say that. He doesn't say things that way. But he said, "We're going to launch the agent toolkit. This is what it's going to be. It's going to be the ground on which enterprises can build their agents, hence more and more tokens. We're going to provide all of the security. We're going to let all of the agents communicate in the right way. We're going to connect this to the right data. We're going to make agents function with one another, like everything that is needed in agentic AI as an architectural construction." And in doing so, Nvidia is going to be selling chips, of course, but now they're going to be selling even more chips because if they become like the foundation, again, the AWS of agentic AI, that's a really smart move. Even though they're not selling the software, for me, this is like act two of the CUDA playbook which made Nvidia what it is. When everyone was ignoring Nvidia back 2016 and 2006, sorry, when he announced the CUDA and then it was very big. And even though there's differences, I think that this is what is happening. And his goal is not to monetize the software. His goal is to sell more and more chips and to be the indispensable player. And we can talk about the competitive dynamics of this.
Gemma Allen
>> I mean, he's so incredibly smart in so many ways. He's the ultimate orchestrator himself.
Raphaelle d'Ornano
>> Right.
Gemma Allen
>> Can you think of anyone in tech who has done it so almost like subliminally as well as the way Jensen operates? It's unbelievable. But there's a couple of things that I saw were interesting from that announcement. Especially it was also 20 years of CUDA, so they were celebrating that last week too.
Raphaelle d'Ornano
>> Right.
Gemma Allen
>> In the meantime, they're creating the next generation of how they're going to own the industry for the next wave in such a, like you said, almost like quietly-
Raphaelle d'Ornano
>> Quietly.
Gemma Allen
>> A quiet way. But if you think about it, and I know you've talked a lot about the SaaSpocalypse and we're going to talk about that today too, but he has hedged his bets both ways. Because if you look at what's bundled in in terms of what he's offering enterprises, it gives them the opportunity to create their own agentic workflows using their own policies, their own procedures like OpenShell. Some of that technology takes some of the governance layer that has always been so key to needing a SaaS operator from the perspective of an enterprise buyer and allows you to do it yourself. It also arms though the SaaS ecosystem somewhat with these tools too, and they're also using them. So the battleground is drawn and it's like Nvidia are now on both sides of it. They cannot lose.
Raphaelle d'Ornano
>> Well, what I think was really interesting was indeed what is at play from a strategic point of view? So I've been writing a lot about the fact that the LLMs and especially Anthropic and OpenAI playing catch up are winning the battle for the orchestration graph. Like imagining where is the human user, the LLM, the systems of records and any context that you fetch to make the agents grounded and like, who is going to win? Is it going to be the LLMs who are at the center close to human or is it going to be the enterprise SaaS incumbents that have many other attributes to play that role? And so who's going to win? Is it going to be the LLMs themselves who do not have the governance, the security, the enterprise, accumulated data over decades and CRM and like workforce management, ITSM, et cetera, or is it going to be the enterprise incumbents that have those attributes but that do not have the MCP, the LLMs, the intelligence and everything that is being built at very high speed? And so Nvidia is giving the SaaS companies the ability to say, "We're providing you with the ground. You're going to be able to build the SAP agents, the Salesforce agents, the Palantir agents," which is ironic because Palantir does connect the data to... I mean, allows enterprises to use their data and AI-Q is kind of like a comparable, if I must say. So I do think it's interesting from a Palantir perspective. But they're giving the SAS incumbents that opportunity, yet they're providing, and this is of course very subtle, the governance themselves. So they're giving it to them, they're giving them the opportunity, but at the same time, they're actually taking that. So look, I think that in both cases, it's a smart move. In both cases, there's going to be more chips. Jensen doesn't really care who wins. If it's the LLMs, he sells the chips, if it's the SaaS companies, he's still selling the chips. So where he doesn't win is if it's a vertical integration where Google sells its chips, Google has the orchestration ground, because this is a competitive move to the hyperscalers. And I've heard a lot of commentary by which people are saying, "Oh, but no, no, no, this is like they're not competing at all." But wait, the hyperscalers are also building all of this orchestration ground because everyone wants to build a playground in which all of these agents, the kids are going to play because this is where you're going to have all the inference and where you are going to have this wave of software playing out. So everyone wants to own the gates of the playground, the actual ground, and have everyone play on their grounds, and allow for that, the security, the trust, the compute, everything, of course the cloud, et cetera, et cetera. So there's one scenario in which Nvidia doesn't win is if we have vertical integration. Again, this is not going to be one move and not the others, but in any case, it's very smart. And I do think that versus the CUDA, which indeed is 20 years old, I think this is going to be way faster.
Gemma Allen
>> For sure.
Raphaelle d'Ornano
>> And this is not going to take 10 years.
Gemma Allen
>> I mean, one thing that's fundamentally going to change is the cost and licensing models for SaaS. Because if you are getting access to certain parts of your value add in an open source format, then like OpenShell or whatever it is, then that is going to fundamentally change how you think about pricing, how you think about competitive advantage. Kramer said to Jensen, "Please stop talking in code, stop talking in like fluffy kind of language. I don't understand this language. I need to know how what you're saying should change how I think about investing." I think that's so interesting because we know that we're living in this time whereby it's evident that a lot of folks that have to make these decisions day in and day out don't necessarily understand the spiel and the narrative and the place of signal around tech. So I thought it was like an interesting point by Kramer to make so publicly. He also made it to theCUBE, to us as well directly, that the language of the industry is shifting so much. And then the other thing Jensen said when he was asked by Kramer as well about job losses, he was asked, "What does it mean from the perspective of headcount?" And he made the point that really these mass job losses to him indicate a sign of like a lack of innovative thinking by leadership teams. Because it shouldn't be about doing more with less staff. It should be doing more with more and more should be coming in the way of this agentic and AI additional opportunity. Don't fire somebody or replace them with an agent, give them a band of agents and see what they can do. And I wish the industry would maybe run with that more because I think at this time in tech, for such uncertainty, it's an important message. This should not necessarily be a race to the bottom for human thinking.
Raphaelle d'Ornano
>> Well, I'll start with the second point indeed on the impact of jobs. So the first statement I would make is that when you look at like investor days, when you look at like all of the communication that is done by public companies, tech or non-tech companies, it's the same thing. You see like two opposite directions. You see, number one, the CEOs who, I don't know, are ignoring that what is happening. They're like, "My headcount is going to be the same. We don't anticipate any changes in headcounts." That's a very strange declaration to make because I understand that you don't want to have 50% less of your headcount, but to say that you don't think there's going to be an impact is a form of hypocrisy and either hypocrisy or like, okay, are we really living in the same world? That's number one. The others that are like axing jobs like crazy and getting sometimes compensated by the stock market, sometimes not, a lot of them are doing that because they had bloated headcounts from the past.
Gemma Allen
>> Like Block?
Raphaelle d'Ornano
>> Maybe, exactly. I mean, to take an example, but there are many others. And it's like AI is the perfect excuse. So as you say, there is like a clear middle where for the companies that have the ability to augment their throughput, who are able to say, "My employee is going to be doing more with AI. We're going to have more product releases. We're going to have more material or content material. We're going to have more X, Y, Z." Because the company is in a competitive position by which it can actually do more. Because in some industries, there's just nothing more to do. You don't have any more business, you're bounded by competitive limits, you have pricing constraints. And so even if you wanted to have more throughput, there is no end market for that throughput. So the companies, once again, that are going to win this whole AI wave are the companies that are going to, again, augment that, price that correctly, and be able to like have maybe a bit less employees. And I don't know if that's a quarter, one third, one fifth, I think no one knows, so let's not give statistics, but with less employees and reasonable proportions, getting more throughput, for sure. Now to go back to your first point on, well, what does this mean for what software is becoming, I think we're right now in an interesting position in the sense that first, the SaaSpocalypse is not gone. And that's hypocrisy number two. It's not because we had a war, it's not because we had other key geopolitical and macro environment moments that diverted from this problem that AI is hurting SaaS is gone. And some investors were quite like, I don't know, they saw that because there's a war in Iran, because there's macro decisions that are being made or not made, well, okay, we can go and invest back into SaaS. So I think the problem is still here. The problem is starting to take a different form where we are understanding that we're not going to vibe code all the applications. We're not going to vibe code your payroll system, your CRM, your Shopify system, whatever. That is naive and that is not even what the labs are saying. No one is saying we're going to vibe code away Workday or ServiceNow. That's not the point. What we are understanding is that software is going to be about providing value through agents that again are in this playground that I mentioned. So it's about controlling the playground of agents that are providing value, and the business model is going to have to change. I have been saying for a long time that we are moving away from a price per seat to consumption in one form of another. But again, it's going to be what is the value that you bring to your customer through your solution? How do you price that based on what the agents are actually accomplishing for you in that workflow? And so it's a question of redefining the superpowers that these platforms have, again, from the trust, from the context. You know I like the context mode a lot. I think of course trust, governance is critical of course, but I think what's really valuable is the context in which decisions are grounded, like the years of pattern matching this works, this doesn't work, this is the right way to do this decision, this matters a lot. And for agents to actually work with the intelligence that is available, the context is what will make software work.
Gemma Allen
>> So let's talk a little bit about context, especially from the perspective of the SaaSpocalypse. You've written a lot about how it's the wrong type of panic. But there is also a very real event happening right now whereby the story, the response, the reply from these SaaS companies isn't exactly succinct. It's either it can come across somewhat defensive, somewhat naive. If you think about the messaging on the other end, if somebody was to say... I heard a CEO yesterday and he gave the example, took Netflix 20 years to make Blockbuster bankrupt. Is that an answer right now? You know what I mean? We cannot compare speed and scale 20 years ago to today, for so many reasons. And as a CEO of a SaaS company, you of all people should know this. So I think it's important that the messaging becomes a little bit clearer. And I think to Jim Kramer's point to Jensen, it is a crazy time. There's a lot of noise, a lot of signal, a lot of patterns that just don't seem to be consistent as well in terms of how Wall Street's responding to what's happening in the tech market. So we see more and more kind of almost like the same sounding replies from these firms on their earnings calls, et cetera, where they're asked to talk about their response to the AI era, talk about how ready they are for the agentic world we're kind of very, very rapidly facing into. And it seems to me as well, Raphaelle, that a lot of their messaging sounds a little bit the same. I've had moments as a journalist where I've thought, do LLMs write these messages? Am I crazy to ask that? But what are your thoughts? What do you think they need to be clearer on?
Raphaelle d'Ornano
>> Well, so my thoughts are the fact that there are two things that should not be underestimated and that are lethal if that is the case. Number one, it's speed. When you look at the speed of reasoning, intelligence, everything that these LLMs are producing, in over the past three months, okay, from November to, the past five months, from November to now, like the pace of progress has been like astonishing. When you see what's happening in coding, which is really the application case for everything, this is very impressive. So first, I think that taking comparison points from, oh, this happened in 2005 and in 2010, this is what happened. Okay, well, things are going to happen 10 times faster, if not 20 times or 100 times faster. So that's the speed is not something that the human brain is accustomed to. Again, you need to be thinking about this and obsessing about this all the time. And even when you're doing that, it's hard to grapple with. So I think number one. Second, I think that we are now at a point where CEOs and investors are recognizing that this is really profound in terms of a redefinition of what is a given business model. Like what does it mean to be a software platform? I don't think at all that software is dead. I think there's a lot of value in software, like a lot of value. I think right now we're at a point where the market is not making any difference and it's kind of like throwing software as soon as it's written software because everyone is panicking, which creates a lot of opportunities in doing that. But there is this need to understand the profoundness of the impact. So speed, profoundness. So I would say that a lot of CEOs are like, what do you want to say when you're getting hammered every day? Because I mean, what it was, yesterday, Amazon didn't... So they didn't even release a specific agent like Claude releases their Claude Cowork, et cetera. They said that they were building agents to replace a part of the jobs that they had eliminated in the part of their last job cuts. And we lost billions of dollars of enterprise value in software just yesterday. I mean, it was like a crazy day yesterday. People were like, "Oh, really? Something's happening." Yes, software is down massively. That was just yesterday because of something that is a very random announcement like Amazon is doing that and so are thousands of companies that are all building agents. But you see, it's enough to wipe away billions of enterprise value because people are in panic mode. So to not panic, you have to acknowledge like, okay, this thing is going to hit me. It's going to be a violent hit." Now is the time to start thinking about moats, what makes a durable moat in this new paradigm.
Gemma Allen
>> So if you were advising the CEO of one of those companies who had a terrible day yesterday, it was a bloodbath, and you were to say to them, "Here's five things you need or three to five things you need to be super, super clear on. This messaging needs to land in every audience from investor to customer to employee," what do you think those things are?
Raphaelle d'Ornano
>> Well, I would say what is your proximity to the human person? What is your capability to capture the intent and to translate that into an agentic workflow? Are you able to sit next to the human and understand what he or she wants to get accomplished to allow agentic AI to get to that business outcome, yes or no? If you're far from that human intent, you're in bad shape. I can tell you that right now. Number two, what is your context that is unique, that is defensible, and is it architectured correctly with the right memory systems that allow agents to go and query that context and to actually make the decisions now? So one and two. One and two are sufficient right now. Don't go and talk about governance, security, and domain expertise because everyone is saying that and now Nvidia is saying that too.
Gemma Allen
>> Exactly.
Raphaelle d'Ornano
>> So that's not the right argument.
Gemma Allen
>> It's almost pointless talking features. They mean so little in this current moment.
Raphaelle d'Ornano
>> I mean, they mean a lot, but it's like if you don't have that, then, well, you should not even be speaking. I mean, because then you have a much bigger problem. It's kind of like it's baked in that enterprise grade software has everything that protects the user, that protects the data, that protects the global enterprise systems that are using it. If you're not able to comply with that, then you cannot be playing in enterprise software. So of course it's a critical component if you don't have it, but it's like it's not the cherry on the cake, it's the flour in the cake. Sorry for my expression.
Gemma Allen
>> For sure. I've been thinking about this and I've been thinking, I grew up in a world of like mission statements and value statements in tech. You wanted to really like sell your product to the market, to the audience. I think now what a lot of these companies need is an agentic through line.
Raphaelle d'Ornano
>> Yes.
Gemma Allen
>> They just need a short descriptor of exactly how it is that they will play, defend, survive, and add value 15 years from now.
Raphaelle d'Ornano
>> And play offense.
Gemma Allen
>> Exactly.
Raphaelle d'Ornano
>> I would even say play offense.
Gemma Allen
>> Exactly.
Raphaelle d'Ornano
>> And that is not, if I may, oh, I have an AI feature or I have an agentic feature. The market doesn't want to hear that. The market wants to understand that you have acknowledged that this is a paradigm.
Gemma Allen
>> .
Raphaelle d'Ornano
>> Who are you in this paradigm and not just, "Oh, I did that. I did that." That's not going to bring any value, like any money on the table. So I'm actually very confident that there's a path to getting there and that we're getting there. There's so much damage that has been done. Some companies are going to emerge. The public markets right now are still really in expectation mode and it's difficult, so it's going to take a bit of time. But I'm confident that in the next weeks, in the next months, some companies are going to emerge very strong with all of this in mind.
Gemma Allen
>> I hope so. Well, we're up on time for today, but tell me what's ahead for you for this next week?
Raphaelle d'Ornano
>> Well, a lot of work, a lot of research. I'm releasing a very exciting manifesto in a couple of days, so I will be happy to talk to that. So that has taken a lot of time and it's like the most fascinating work I've ever done.
Gemma Allen
>> Well, if ever there was a moment that an industry needed one, Raphaelle, it's right now. So I'm so excited to read it and to see you at the cusp of this. And that's a wrap for this week's episode of AGNT. Tune in next week where Raphaelle and I will break down what's happening in the industry and what's really happening behind these moats.
AGNT Podcast Ep. 5 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. And 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 defining the next chapter of AI. Let's get into it. Raphaelle, welcome back to the NYSE and to theCUBE and NYSE Wired. We were at San Jose and Nvidia GTC last week. It was a wild week, and I know you had a lot to say about it on Substack and on your socials this weekend. So maybe let's start there. Let's start with just the absolute madness of this moment. Like what are your thoughts on some of the learnings and conversations from Nvidia?
Raphaelle d'Ornano
>> Well, I mean, I think this conference every year, it's kind of like the moment when everyone is trying to understand, is this really real? Is this going to continue at this speed? So I think that it has taken a position in this space that goes way beyond Nvidia. I mean, Nvidia goes beyond itself. And so I think GTC is a very key moment in like, okay, where are we? And Jensen has been very clear over the past month on inference is the big deal. Everything is going to be compute, like everything, everything, everything. And I think that what we saw at that conference was that there's going to be way more chips than we thought there were going to be. I mean, I'm not surprised. I don't think he would be surprised. Many people are not surprised. But it's like there's going to be way more as AI gets adopted as agents proliferate with the mold book continuing and reverberating. And one announcement that I really liked that he made at the conference, which was during the second hour of this monster keynote, where he announced that they were going to be kind of the hyperscaler of agentic AI. He didn't say that. He doesn't say things that way. But he said, "We're going to launch the agent toolkit. This is what it's going to be. It's going to be the ground on which enterprises can build their agents, hence more and more tokens. We're going to provide all of the security. We're going to let all of the agents communicate in the right way. We're going to connect this to the right data. We're going to make agents function with one another, like everything that is needed in agentic AI as an architectural construction." And in doing so, Nvidia is going to be selling chips, of course, but now they're going to be selling even more chips because if they become like the foundation, again, the AWS of agentic AI, that's a really smart move. Even though they're not selling the software, for me, this is like act two of the CUDA playbook which made Nvidia what it is. When everyone was ignoring Nvidia back 2016 and 2006, sorry, when he announced the CUDA and then it was very big. And even though there's differences, I think that this is what is happening. And his goal is not to monetize the software. His goal is to sell more and more chips and to be the indispensable player. And we can talk about the competitive dynamics of this.
Gemma Allen
>> I mean, he's so incredibly smart in so many ways. He's the ultimate orchestrator himself.
Raphaelle d'Ornano
>> Right.
Gemma Allen
>> Can you think of anyone in tech who has done it so almost like subliminally as well as the way Jensen operates? It's unbelievable. But there's a couple of things that I saw were interesting from that announcement. Especially it was also 20 years of CUDA, so they were celebrating that last week too.
Raphaelle d'Ornano
>> Right.
Gemma Allen
>> In the meantime, they're creating the next generation of how they're going to own the industry for the next wave in such a, like you said, almost like quietly-
Raphaelle d'Ornano
>> Quietly.
Gemma Allen
>> A quiet way. But if you think about it, and I know you've talked a lot about the SaaSpocalypse and we're going to talk about that today too, but he has hedged his bets both ways. Because if you look at what's bundled in in terms of what he's offering enterprises, it gives them the opportunity to create their own agentic workflows using their own policies, their own procedures like OpenShell. Some of that technology takes some of the governance layer that has always been so key to needing a SaaS operator from the perspective of an enterprise buyer and allows you to do it yourself. It also arms though the SaaS ecosystem somewhat with these tools too, and they're also using them. So the battleground is drawn and it's like Nvidia are now on both sides of it. They cannot lose.
Raphaelle d'Ornano
>> Well, what I think was really interesting was indeed what is at play from a strategic point of view? So I've been writing a lot about the fact that the LLMs and especially Anthropic and OpenAI playing catch up are winning the battle for the orchestration graph. Like imagining where is the human user, the LLM, the systems of records and any context that you fetch to make the agents grounded and like, who is going to win? Is it going to be the LLMs who are at the center close to human or is it going to be the enterprise SaaS incumbents that have many other attributes to play that role? And so who's going to win? Is it going to be the LLMs themselves who do not have the governance, the security, the enterprise, accumulated data over decades and CRM and like workforce management, ITSM, et cetera, or is it going to be the enterprise incumbents that have those attributes but that do not have the MCP, the LLMs, the intelligence and everything that is being built at very high speed? And so Nvidia is giving the SaaS companies the ability to say, "We're providing you with the ground. You're going to be able to build the SAP agents, the Salesforce agents, the Palantir agents," which is ironic because Palantir does connect the data to... I mean, allows enterprises to use their data and AI-Q is kind of like a comparable, if I must say. So I do think it's interesting from a Palantir perspective. But they're giving the SAS incumbents that opportunity, yet they're providing, and this is of course very subtle, the governance themselves. So they're giving it to them, they're giving them the opportunity, but at the same time, they're actually taking that. So look, I think that in both cases, it's a smart move. In both cases, there's going to be more chips. Jensen doesn't really care who wins. If it's the LLMs, he sells the chips, if it's the SaaS companies, he's still selling the chips. So where he doesn't win is if it's a vertical integration where Google sells its chips, Google has the orchestration ground, because this is a competitive move to the hyperscalers. And I've heard a lot of commentary by which people are saying, "Oh, but no, no, no, this is like they're not competing at all." But wait, the hyperscalers are also building all of this orchestration ground because everyone wants to build a playground in which all of these agents, the kids are going to play because this is where you're going to have all the inference and where you are going to have this wave of software playing out. So everyone wants to own the gates of the playground, the actual ground, and have everyone play on their grounds, and allow for that, the security, the trust, the compute, everything, of course the cloud, et cetera, et cetera. So there's one scenario in which Nvidia doesn't win is if we have vertical integration. Again, this is not going to be one move and not the others, but in any case, it's very smart. And I do think that versus the CUDA, which indeed is 20 years old, I think this is going to be way faster.
Gemma Allen
>> For sure.
Raphaelle d'Ornano
>> And this is not going to take 10 years.
Gemma Allen
>> I mean, one thing that's fundamentally going to change is the cost and licensing models for SaaS. Because if you are getting access to certain parts of your value add in an open source format, then like OpenShell or whatever it is, then that is going to fundamentally change how you think about pricing, how you think about competitive advantage. Kramer said to Jensen, "Please stop talking in code, stop talking in like fluffy kind of language. I don't understand this language. I need to know how what you're saying should change how I think about investing." I think that's so interesting because we know that we're living in this time whereby it's evident that a lot of folks that have to make these decisions day in and day out don't necessarily understand the spiel and the narrative and the place of signal around tech. So I thought it was like an interesting point by Kramer to make so publicly. He also made it to theCUBE, to us as well directly, that the language of the industry is shifting so much. And then the other thing Jensen said when he was asked by Kramer as well about job losses, he was asked, "What does it mean from the perspective of headcount?" And he made the point that really these mass job losses to him indicate a sign of like a lack of innovative thinking by leadership teams. Because it shouldn't be about doing more with less staff. It should be doing more with more and more should be coming in the way of this agentic and AI additional opportunity. Don't fire somebody or replace them with an agent, give them a band of agents and see what they can do. And I wish the industry would maybe run with that more because I think at this time in tech, for such uncertainty, it's an important message. This should not necessarily be a race to the bottom for human thinking.
Raphaelle d'Ornano
>> Well, I'll start with the second point indeed on the impact of jobs. So the first statement I would make is that when you look at like investor days, when you look at like all of the communication that is done by public companies, tech or non-tech companies, it's the same thing. You see like two opposite directions. You see, number one, the CEOs who, I don't know, are ignoring that what is happening. They're like, "My headcount is going to be the same. We don't anticipate any changes in headcounts." That's a very strange declaration to make because I understand that you don't want to have 50% less of your headcount, but to say that you don't think there's going to be an impact is a form of hypocrisy and either hypocrisy or like, okay, are we really living in the same world? That's number one. The others that are like axing jobs like crazy and getting sometimes compensated by the stock market, sometimes not, a lot of them are doing that because they had bloated headcounts from the past.
Gemma Allen
>> Like Block?
Raphaelle d'Ornano
>> Maybe, exactly. I mean, to take an example, but there are many others. And it's like AI is the perfect excuse. So as you say, there is like a clear middle where for the companies that have the ability to augment their throughput, who are able to say, "My employee is going to be doing more with AI. We're going to have more product releases. We're going to have more material or content material. We're going to have more X, Y, Z." Because the company is in a competitive position by which it can actually do more. Because in some industries, there's just nothing more to do. You don't have any more business, you're bounded by competitive limits, you have pricing constraints. And so even if you wanted to have more throughput, there is no end market for that throughput. So the companies, once again, that are going to win this whole AI wave are the companies that are going to, again, augment that, price that correctly, and be able to like have maybe a bit less employees. And I don't know if that's a quarter, one third, one fifth, I think no one knows, so let's not give statistics, but with less employees and reasonable proportions, getting more throughput, for sure. Now to go back to your first point on, well, what does this mean for what software is becoming, I think we're right now in an interesting position in the sense that first, the SaaSpocalypse is not gone. And that's hypocrisy number two. It's not because we had a war, it's not because we had other key geopolitical and macro environment moments that diverted from this problem that AI is hurting SaaS is gone. And some investors were quite like, I don't know, they saw that because there's a war in Iran, because there's macro decisions that are being made or not made, well, okay, we can go and invest back into SaaS. So I think the problem is still here. The problem is starting to take a different form where we are understanding that we're not going to vibe code all the applications. We're not going to vibe code your payroll system, your CRM, your Shopify system, whatever. That is naive and that is not even what the labs are saying. No one is saying we're going to vibe code away Workday or ServiceNow. That's not the point. What we are understanding is that software is going to be about providing value through agents that again are in this playground that I mentioned. So it's about controlling the playground of agents that are providing value, and the business model is going to have to change. I have been saying for a long time that we are moving away from a price per seat to consumption in one form of another. But again, it's going to be what is the value that you bring to your customer through your solution? How do you price that based on what the agents are actually accomplishing for you in that workflow? And so it's a question of redefining the superpowers that these platforms have, again, from the trust, from the context. You know I like the context mode a lot. I think of course trust, governance is critical of course, but I think what's really valuable is the context in which decisions are grounded, like the years of pattern matching this works, this doesn't work, this is the right way to do this decision, this matters a lot. And for agents to actually work with the intelligence that is available, the context is what will make software work.
Gemma Allen
>> So let's talk a little bit about context, especially from the perspective of the SaaSpocalypse. You've written a lot about how it's the wrong type of panic. But there is also a very real event happening right now whereby the story, the response, the reply from these SaaS companies isn't exactly succinct. It's either it can come across somewhat defensive, somewhat naive. If you think about the messaging on the other end, if somebody was to say... I heard a CEO yesterday and he gave the example, took Netflix 20 years to make Blockbuster bankrupt. Is that an answer right now? You know what I mean? We cannot compare speed and scale 20 years ago to today, for so many reasons. And as a CEO of a SaaS company, you of all people should know this. So I think it's important that the messaging becomes a little bit clearer. And I think to Jim Kramer's point to Jensen, it is a crazy time. There's a lot of noise, a lot of signal, a lot of patterns that just don't seem to be consistent as well in terms of how Wall Street's responding to what's happening in the tech market. So we see more and more kind of almost like the same sounding replies from these firms on their earnings calls, et cetera, where they're asked to talk about their response to the AI era, talk about how ready they are for the agentic world we're kind of very, very rapidly facing into. And it seems to me as well, Raphaelle, that a lot of their messaging sounds a little bit the same. I've had moments as a journalist where I've thought, do LLMs write these messages? Am I crazy to ask that? But what are your thoughts? What do you think they need to be clearer on?
Raphaelle d'Ornano
>> Well, so my thoughts are the fact that there are two things that should not be underestimated and that are lethal if that is the case. Number one, it's speed. When you look at the speed of reasoning, intelligence, everything that these LLMs are producing, in over the past three months, okay, from November to, the past five months, from November to now, like the pace of progress has been like astonishing. When you see what's happening in coding, which is really the application case for everything, this is very impressive. So first, I think that taking comparison points from, oh, this happened in 2005 and in 2010, this is what happened. Okay, well, things are going to happen 10 times faster, if not 20 times or 100 times faster. So that's the speed is not something that the human brain is accustomed to. Again, you need to be thinking about this and obsessing about this all the time. And even when you're doing that, it's hard to grapple with. So I think number one. Second, I think that we are now at a point where CEOs and investors are recognizing that this is really profound in terms of a redefinition of what is a given business model. Like what does it mean to be a software platform? I don't think at all that software is dead. I think there's a lot of value in software, like a lot of value. I think right now we're at a point where the market is not making any difference and it's kind of like throwing software as soon as it's written software because everyone is panicking, which creates a lot of opportunities in doing that. But there is this need to understand the profoundness of the impact. So speed, profoundness. So I would say that a lot of CEOs are like, what do you want to say when you're getting hammered every day? Because I mean, what it was, yesterday, Amazon didn't... So they didn't even release a specific agent like Claude releases their Claude Cowork, et cetera. They said that they were building agents to replace a part of the jobs that they had eliminated in the part of their last job cuts. And we lost billions of dollars of enterprise value in software just yesterday. I mean, it was like a crazy day yesterday. People were like, "Oh, really? Something's happening." Yes, software is down massively. That was just yesterday because of something that is a very random announcement like Amazon is doing that and so are thousands of companies that are all building agents. But you see, it's enough to wipe away billions of enterprise value because people are in panic mode. So to not panic, you have to acknowledge like, okay, this thing is going to hit me. It's going to be a violent hit." Now is the time to start thinking about moats, what makes a durable moat in this new paradigm.
Gemma Allen
>> So if you were advising the CEO of one of those companies who had a terrible day yesterday, it was a bloodbath, and you were to say to them, "Here's five things you need or three to five things you need to be super, super clear on. This messaging needs to land in every audience from investor to customer to employee," what do you think those things are?
Raphaelle d'Ornano
>> Well, I would say what is your proximity to the human person? What is your capability to capture the intent and to translate that into an agentic workflow? Are you able to sit next to the human and understand what he or she wants to get accomplished to allow agentic AI to get to that business outcome, yes or no? If you're far from that human intent, you're in bad shape. I can tell you that right now. Number two, what is your context that is unique, that is defensible, and is it architectured correctly with the right memory systems that allow agents to go and query that context and to actually make the decisions now? So one and two. One and two are sufficient right now. Don't go and talk about governance, security, and domain expertise because everyone is saying that and now Nvidia is saying that too.
Gemma Allen
>> Exactly.
Raphaelle d'Ornano
>> So that's not the right argument.
Gemma Allen
>> It's almost pointless talking features. They mean so little in this current moment.
Raphaelle d'Ornano
>> I mean, they mean a lot, but it's like if you don't have that, then, well, you should not even be speaking. I mean, because then you have a much bigger problem. It's kind of like it's baked in that enterprise grade software has everything that protects the user, that protects the data, that protects the global enterprise systems that are using it. If you're not able to comply with that, then you cannot be playing in enterprise software. So of course it's a critical component if you don't have it, but it's like it's not the cherry on the cake, it's the flour in the cake. Sorry for my expression.
Gemma Allen
>> For sure. I've been thinking about this and I've been thinking, I grew up in a world of like mission statements and value statements in tech. You wanted to really like sell your product to the market, to the audience. I think now what a lot of these companies need is an agentic through line.
Raphaelle d'Ornano
>> Yes.
Gemma Allen
>> They just need a short descriptor of exactly how it is that they will play, defend, survive, and add value 15 years from now.
Raphaelle d'Ornano
>> And play offense.
Gemma Allen
>> Exactly.
Raphaelle d'Ornano
>> I would even say play offense.
Gemma Allen
>> Exactly.
Raphaelle d'Ornano
>> And that is not, if I may, oh, I have an AI feature or I have an agentic feature. The market doesn't want to hear that. The market wants to understand that you have acknowledged that this is a paradigm.
Gemma Allen
>> .
Raphaelle d'Ornano
>> Who are you in this paradigm and not just, "Oh, I did that. I did that." That's not going to bring any value, like any money on the table. So I'm actually very confident that there's a path to getting there and that we're getting there. There's so much damage that has been done. Some companies are going to emerge. The public markets right now are still really in expectation mode and it's difficult, so it's going to take a bit of time. But I'm confident that in the next weeks, in the next months, some companies are going to emerge very strong with all of this in mind.
Gemma Allen
>> I hope so. Well, we're up on time for today, but tell me what's ahead for you for this next week?
Raphaelle d'Ornano
>> Well, a lot of work, a lot of research. I'm releasing a very exciting manifesto in a couple of days, so I will be happy to talk to that. So that has taken a lot of time and it's like the most fascinating work I've ever done.
Gemma Allen
>> Well, if ever there was a moment that an industry needed one, Raphaelle, it's right now. So I'm so excited to read it and to see you at the cusp of this. And that's a wrap for this week's episode of AGNT. Tune in next week where Raphaelle and I will break down what's happening in the industry and what's really happening behind these moats.
