Raphaëlle d'Ornano, founder and Chief Executive Officer of Decoding Discontinuity, joins Gemma Allen of theCUBE to discuss the future of agentic artificial intelligence in the enterprise sector, broadcasting from the iconic New York Stock Exchange. The AGNT podcast examines how intelligent systems reshape businesses and markets, initiating a new chapter of AI innovation.
In this inaugural episode, the discussion begins with an exploration of d'Ornano's expertise in agentic AI, a growing field that emphasizes collaboration between human intelligence and AI systems. The hosts delve into the evolution of intelligent systems and highlight Salesforce's strategic vision and readiness to introduce Agentforce, paving the way for technological transformation. Insights from theCUBE Research and Allen's engaging hosting style facilitate the conversation.
Key takeaways from the episode include d'Ornano's assertion that context, rather than data, forms the core moat in agentic AI development. The discussion also covers the groundbreaking initial public offering of Chinese company MiniMax, which has taken significant strides as a leader in the large language model space. According to d'Ornano, understanding how enterprises can leverage their existing assets in the new agentic paradigm is crucial for gaining competitive advantage and driving sustainable growth.
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. 4 with Gemma Allen & Raphaëlle d'Ornano
Raphaëlle d'Ornano, founder and Chief Executive Officer of Decoding Discontinuity, joins Gemma Allen of theCUBE to discuss the future of agentic artificial intelligence in the enterprise sector, broadcasting from the iconic New York Stock Exchange. The AGNT podcast examines how intelligent systems reshape businesses and markets, initiating a new chapter of AI innovation.
In this inaugural episode, the discussion begins with an exploration of d'Ornano's expertise in agentic AI, a growing field that emphasizes collaboration between human intelligence and AI systems. The hosts delve into the evolution of intelligent systems and highlight Salesforce's strategic vision and readiness to introduce Agentforce, paving the way for technological transformation. Insights from theCUBE Research and Allen's engaging hosting style facilitate the conversation.
Key takeaways from the episode include d'Ornano's assertion that context, rather than data, forms the core moat in agentic AI development. The discussion also covers the groundbreaking initial public offering of Chinese company MiniMax, which has taken significant strides as a leader in the large language model space. According to d'Ornano, understanding how enterprises can leverage their existing assets in the new agentic paradigm is crucial for gaining competitive advantage and driving sustainable growth.
AGNT Podcast Ep. 4 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 to theCUBE and NYSE Wired.
Raphaelle d'Ornano
>> Thank you.
Gemma Allen
>> What an unbelievable week. I mean, this morning, things were so loud on the options floor behind us that we had some guests who were struggling to concentrate on the interview.
Raphaelle d'Ornano
>> Wow.
Gemma Allen
>> And there's a couple of names that we've heard a lot this morning from that floor, and we're going to talk about them, the obvious ones, Nvidia, Salesforce. But before we go there, I want to talk to you about one of my favorite reads of the week, your Substack, and specifically, as a creative person myself, Figma. So you have done a bit of a 360 on Figma, right? You were a little bit on the fence and now you're a little bit all in. Tell me what you're thinking.
Raphaelle d'Ornano
>> So Figma is a really interesting company. Look, it was one of the flagship IPOs, one of the flagship SaaS IPOs of this year, I mean of 2025, way before this whole SaaS-pocalypse. And when it went public, you had a company that had exceptional figures. The numbers that they were reporting in terms of like ARR growth, net retention, margins, those were outstanding. I had underlined the research that I had done, that the numbers were disclosed in a very selective way nonetheless, and that the numbers that were effectively like wow were not on the whole population. Anyways, the company at that time, at IPO, was priced to perfection. It was priced much more than companies that were trading at the upper end like Datadog, ServiceNow, which was trading very well at that time, Snowflake, and I was like, "Wait." This company, which has impressive numbers, if you put aside again the fact that it's selective, which matters a lot, it was priced for this orchestration future as if it had already earned its orchestration future. So it was like, it's not that I don't like the company or that I don't believe in the product, which seems to be very good, though I'm not a user every day, but does it warrant that valuation? I was like, no. So I had a very bearish view, and the bearish view played out effectively during the year because Figma was down approximately 70.8% up until its very low, which is ... I mean, I know a lot of companies got crashed, but I think minus 70.8% is one of the big cracks. And then last week, in the middle of all of this noise around SaaS is dead, what is happening to SaaS, they launched a very smart partnership with Anthropic by which they're able to really integrate themselves in the orchestration graph ... and the word matters, the orchestration that we have been talking a lot about here ... in terms of Anthropic's world, and they built a reciprocity between Claude and Canva, their product, by which developers that are doing work in Claude are able to see the effect in Figma, vice versa. And so you have this new frontier between not the human UI but the agentic UI, and in a very smart move. And so when I saw that, I was like, "Well, look. This is one of the proof points that I had been waiting for to see whether Figma could earn its orchestration future." And so there we go, and so now my position is quite long on the stock.
Gemma Allen
>> So here's the thing, right, about Figma. It's a great product. I think it's an excellent brain. It has some very loyal customers and users. But if you think about what makes it, its USP, and especially from the perspective of defensibility. That MCP protocol, that's repeatable, though. There's nothing to stop Canva doing that. Adobe, which is already a very highly competitive product in the space if not a leading product, already has some of the integration at its core. So when you think about the USP for a company like that, how do you assess its market position in terms of its orchestration layer, versus how it has something unique to its competitors?
Raphaelle d'Ornano
>> Right. So I think that the whole space is really being shaped right now, and I think being one of the first movers actually matters. And so Anthropic had their famous Developer Day this week, which was expected to be like, "Oh wow, what is Anthropic going to release and who is going to die next," because this is what is happening. Anthropic releases a plugin, stocks go down, and so the question is that ...
Gemma Allen
>> Look at IBM.
Raphaelle d'Ornano
>> Right, who is next? So they had their Developer Day and they showed this. "Okay, we are partnering with the ecosystem," which is not new for me, "and this orchestration graph, you need to be positioned in it." So I think first-mover advantage matters a lot. Number two, not all horizontal tools, not all design tools, are plugged into that yet. So I mean, it is technically feasible, but many things are technically feasible, so having a first-mover advantage matters. And I think, again, this protocol only makes you grow to the extent that you already have a very strong user base. You already have a lot of developer love, what Anthropic has. And so I think it's translating the superpowers that they have in the current paradigm into the agentic paradigm, if it makes sense. I don't think it's magic. It's not a magical wand, but it will help them be even better, and hopefully they will also have the first-mover advantage. I mean, we're very early on to this, and this also matters as long as design actually matters. So maybe in two years this will be completely obsolete, and the moat which exists as of today will not exist. It's a very fast-moving space.
Gemma Allen
>> Well, let's talk with the SaaS-pocalypse for a second, right? Because I read something I thought was quite interesting as well this morning around how we're assessing these SaaS tools from the perspective of defensibility and orchestration and entry points, in terms of how much of your day is spent on only one tool creates some sort of understanding around its future defensibility. But some folks argue that actually the true orchestrators are companies like GitHub, or the real agentic entry point is around those who provide STKs or those dev tools, which is really at the very root, like the source of the supply chain, per se. What are your thoughts on that as a counterargument to the whole SaaS-pocalypse debate?
Raphaelle d'Ornano
>> Well, so I've written a lot that coding is the wedge. Coding is one of the first entry points into the enterprise, and so that's where Anthropic has been very smart, we need to say. I mean, I remember writing in August when GPT-5 was out, and Anthropic ... there was this debate already, like who's winning, Anthropic, OpenAI, et cetera. And I had said, "Well, coding is the wedge, and that is how Anthropic is choosing deliberately to have this Trojan horse into the enterprise, because that's where, if you're able to conquer this territory, then it's much easier to move up stack into different extensions." So I think they've been doing that well, and I think coding being the wedge and being adopted by developers is critical in penetrating the enterprise in a bottom-up way, like, for sure. I think that, as we think around defensibility of SaaS ... which is the number-one question that everyone is having right now, because again, the volatility on the market is just crazy right now, and Anthropic is a big cause of this craziness ... we're still applying the frameworks that are, "Oh, vertical is defensible, and so Procore is defensible maybe, and horizontal tools, well, we can all throw the horizontal tools away." So I don't know, like Box being a name that comes to my head, and there's others of course, but that is a very old-school way, if I may say, of looking at defensibility. So it's done because it's vertical or horizontal, that something is defensible or not. I propose looking at it this way. First, there's a distinction between systems of records and tools. I think that's two categories that we could make, and systems of records can move up into systems of actions. That's path to defensibility number one that we had two weeks ago, if I recall well. And then there's tools, and this is where the Figma example comes in. Well, if you are a tool and if you do not have that pathway from the system of record up, well, how do you plug yourself into this new agentic AI stack, the orchestration graph, and what does that look like? But I don't think to say that because you're vertical, you could be a very clunky, old and non-defensible piece of software. It's done because you're vertical and because you have data, which is the other mental obstacle that we have. So if a company today is vertical and has data, oh my god, it's very defensible. No, maybe it's not defensible at all. And maybe a horizontal company, which business is not to have that data, can find a way to be very defensible in a new reality that is just being built day after day. So I think there's a lot of ... I mean, I understand people are trying to put frameworks. That's a part of my work too, but let's really go into the architectural understanding of the stack. Let's go into understanding what the AI labs are actually doing. And so using that as, okay, this is their roadmap. This is how we as an incumbent, imagine if I was an incumbent software company, this is how we defend and this is how we compose in this new architectural paradigm, and right now we're not having the right lenses.
Gemma Allen
>> So if we think about companies in the system of record to great tools to systems of action space, let's talk about some earnings, and let's talk about the ultimate example of that, Salesforce CRM. They had an interesting week. Again, earnings are pretty good. Everything's looking very, very solid. They're having a good day. What are your thoughts on that company across the board? Do you think that they are an agentic orchestrator? Do you think that this agentic platform that they're all in on is the multibillion-dollar solution to staying in the longevity game? Because some might argue as a product, it's chunky, and the change management is tough. It's not necessarily seamless and fluid from the perspective of disruptor comparisons, right?
Raphaelle d'Ornano
>> So, look. I think CRM is going through a lot of changes. And yesterday their earnings call was very interesting, because they actually had some of their key customers that were in the room with Mark or on Vizio and that were actually explaining how they were using Salesforce in today's context and with their product agent. So first, look at the numbers. This was their strongest quarter ever, and the level of growth at a company that has this level of ARR, because let's not forget that Anthropic's growth is impressive. I'm the first one to say that. But when you're starting from their revenue base, which is huge, I mean, delivering that growth over 10%, it's pretty impressive. Guidance was very good. And if you look specifically at their numbers on Agentforce combined with Informatica, their recent acquisition, which forms this assemble from the context at the bottom, Informatica, and Agentforce as the system of action, well, the numbers, they were growing at like 169%. I mean, 169%, maybe it's not the 1000% that Anthropic is growing at, but it's still very huge for SaaS. And so they're demonstrating that Agentforce has that explosive growth and that it's being used in a way that is fundamentally different than the traditional CRM. And this is what their customers, not me, their customers were saying yesterday during the earnings call, and the market actually reacted quite positively. I mean, CRM is up 4% just before we started today. So I think they're in the middle of a transition that Mark has been extremely vocal about when he started, I mean, Agentforce at the very beginning when no one believed that this could even be a reality, which is something we again need to recognize that we had recognized when we were on the ground at Dreamforce this year. I mean, they have been having this idea. Now the numbers are starting to translate that into real numbers, real growth, defensible margins, defensible cash flow. Let's see how that plays out and let's see how this conversion ... they introduced also this notion of agentic work units, which is the first KPI on understanding how agentic work actually is used in the enterprise at their customers, so let's see how this plays out. Look, I think it's in the course of being built. We need to see. Right now, let's see,
Gemma Allen
>> When we think about a company like Salesforce ... and by the way, I'll say I've used Salesforce and I've worked in a number of rollouts. It's a fantastic product. It's very well integrated. The plug-and-play is phenomenal across a resource like LinkedIn. I think Tableau is a great product too, even as somebody who's Power BI for a very long time. Slack is a great product, but again, in some way they're siloed touch points for enterprise. You might be a Slack Teams user and a Salesforce user. Do you think that the real competitive advantage as in mass integration, do you think that's why companies like Google or even Microsoft ... and I know you have mixed feelings there ... have such a unique position, in that in the agentic world, you touch those products in so many parts and aspects of your day? It's not just a data management system or a productivity tool. It's a bit of everything.
Raphaelle d'Ornano
>> Well, look. I think one of the first laws of value in the agentic world is being close to the user intent. I call this proximity to user intent, and I think that one of their key modes is Slack and Slackbot. And I think that when you're a user and you're effectively expressing an intent, on, "I want to have this work completed, to produce this outcome through Slack," I think it's one of Salesforce's most powerful assets. And I think, again, during the course of making that a reality, I don't have the answers, but it's work in progress. I view Slack as a very strong item of defensibility, and Informatica allows it to ground that reality and to be that enterprise brain. So it's interesting,
Gemma Allen
>> We have to talk about the big, big earnings of the week. A company that has literally defied capitalism, Nvidia. Those numbers, they're monstrous. They're just insane. The stock, Wall Street isn't exactly reacting as I'm sure Jensen had hoped they would, but it is insane to think the growth this company is having, right? We had a guest on this morning, and sometimes when we're doing short-form content, we ask a guest, "Hey, what are you listening to," and I expect them to say bad Bunny or Britney Spears, and then he said Jensen. I was like, there is just such a cult following around this company. What are your thoughts on what we heard versus the market reaction?
Raphaelle d'Ornano
>> Look, I've been quite vocal on the fact that, with this agentic AI buildout, we are going towards what I term as the inference economy. It's one of my favorite terms, the inference economy. Everything is going to be compute, everything is going to be inference, and we're going to have outputs that are going to be multiples of compute. Imagine that we're doing this show and that we're agents. We would be using compute, but of course, we're not agents and we're human persons, but imagine. You can complete a lot of tasks with agents, and agents are fundamentally units of compute, and so multiples of the GPUs that who, Nvidia, is selling. So as we progress into the agentic economy, as all of these workloads become agentic in the enterprise, going back to Agentforce that we just mentioned, you're going to have a tremendous use of compute not for training anymore, but for inference. At the same time, the scaling laws are continuing, and so up until now, you're having massive use of GPUs for training. So Nvidia has the training that's going well in inference. So I think, number one, Wall Street is starting to get used to 70% growth on that base, which I agree with you is not something that is normal. But I think that was kind of, "Oh, well, if they don't have that number, then we will be disappointed and punish the stock severely." And that's not what happened, because again, the numbers were there, both from a revenue standpoint and from a gross margin standpoint. I think that we're starting to have doubts. It's really around, I mean, back to the circular economy thing. And I've always said if OpenAI doesn't get their financing, we have a big problem. And I think just this week we saw, oh, well, actually Nvidia, they're not that convinced that they're going to really invest in OpenAI.
Gemma Allen
>> Also, on the circular financing, we saw some novel deals to speak to with AMD and Meta, these mass codependencies in a whole new sphere. We haven't seen those types of scenarios before, really, not at this level.
Raphaelle d'Ornano
>> No, and I think Wall Street is starting to be ... I mean, it's not starting ... is very nervous around that, and to some extent that's reflected in the earnings. It's like, well, you take all of this around, and who is buying all those GPUs? What is the probability of that financing? Well, AWS now is conditioning that. Amazon is conditioning that to either they reach AGI, which is a very strong technological milestone, or they go to IPO, but to go to IPO, they actually need financing, so what do we do with that? Nvidia is going to give them like $30 billion, but it's like Nvidia needs to give that money to OpenAI because OpenAI is buying their GPUs. So look, where I'm very confident, the inference economy is more than real and this is happening, and Jensen has been very vocal about that. I agree.
Gemma Allen
>> Maybe just to counter the Nvidia being, again, the lion ... and I'm sure they will continue to be the lion, there's no denying that ... but to counter the argument as Nvidia is the only show in town in the inference economy, one thing we do hear sometimes from some of the chasing pack that are chasing the incumbents is that the training world and the inference world, from a GPU chip, from a structure perspective, is quite different.
Raphaelle d'Ornano
>> Oh, yes.
Gemma Allen
>> Blackwell is exceptionally hard to run. Even some of the requirements you have around rolling out Nvidia chips across data centers, it's huge, so they have a supply chain clog issue. So there is definitely, it seems to me, space for some of these chasing pack members to make some ground on them, especially if inference is creating a whole new capacity layer.
Raphaelle d'Ornano
>> Oh, for sure. I mean, I'm very long on AMD personally. I mean, I love AMD, and so they did announce a very key partnership with Meta this week. So I mean, I love that company and I think they have a very strong play.
Gemma Allen
>> The CEO is female.
Raphaelle d'Ornano
>> She's great. So I mean, I'm very bullish on AMD, and I mean, again, they're also doing circular financing. I'm not saying that it's bad. I'm saying the market is nervous and we need to be careful, especially when it's around a company that is OpenAI, which it's not. I mean, circular financing in general, if we can avoid it, that's better. But I mean, that is happening, so let's not make a judgment right here. But there's other companies, of course, there's AMD, there's all of the GPUs that could be used for inference, and there's a lot of developments that are happening around there right now. The chips that are being built are mostly for training purposes, but there is an expectation that it's going to migrate to inference. I've always said that neoclouds are not specifically meant to serve inference workloads. We have CoreWeave earnings in a couple of minutes. This whole market is changing, where I think Jensen is completely right that inference is going to be so much bigger than we think it is, because you have to translate the whole economy into everything is compute. I still hear some hedge funds that are like, "Are you serious? This is really happening? We're becoming compute?" Yes. And so there's an arbitrage between human labor and compute, and then you go back to the physical bottleneck. Well, then do we have enough electricity? Do we have enough this and that? What do we have that could be a physical bottleneck, like atoms meets bits, that could make this inference economy not be true? So this is the nervousness. I don't think it's a reaction to the numbers were bad. The numbers were really good, so I don't think there was any disappointment here. It's more, okay, well ...
Gemma Allen
>> It's a fever dream.
Raphaelle d'Ornano
>> Yeah, the magic effect has gone. I mean, to some extent, the deal that AMD announced, I mean, we didn't have plus-20, 30%. You have companies that are not that great that are going up 30% after their earnings. So again, Snowflake announced amazing earnings last night. The stock was up. The stock is trading, I mean, almost in neutral territory. So look, things are difficult to understand at some point, but the direction Nvidia is going is the right direction, provided that the whole thing doesn't blow up with OpenAI, which we need to be careful about.
Gemma Allen
>> Okay. To finish, let's talk about a company I think we seem to talk about a lot, but let's finish on this, Anthropic. A really interesting week. I mean, their COBOL revelation in terms of what Claude Code can do has certainly had some big impacts on IBM, and it impacts the market broadly, right? They've also had an interesting week geopolitically and politically. I mean, we saw the news around the Hegseth meeting. They have until tomorrow, I guess really, to decide if they're going to bow the hat or not.
Raphaelle d'Ornano
>> Right.
Gemma Allen
>> Who knows? What are your thoughts?
Raphaelle d'Ornano
>> Look, I think from an investment perspective, Anthropic needs to be understood as, "Track Anthropic and you'll get a sense of what is happening." I mean, again, we had Claude Code, then we had Claude Code work, then we had the vertical plugins, then we had Claude Code Security, and it's going one by one. So what comes next? The Developer Day they had on Tuesday, everyone's like, "Oh, this is cool. No one is getting killed and actually they're having a friendlier approach." That is not the question. Again, the question is what is Anthropic building that is the new platform that both OpenAI and Anthropic are building, and how do you fit into that?
Gemma Allen
>> Here's the thing, though, right? You say understand Anthropic and you understand what's happening in the market. But if anything, this week has shown us that really when it comes to the reaction of the markets to these tech companies, often the folks that are perhaps making the decisions maybe don't fully understand the technology, because Claude Code Security is a perfect example of that in terms of what it did to CrowdStrike this week. It's not competing at a code vulnerability level. CrowdStrike actually operates in a very different part, a different space completely, but the market's reacting to heightened narratives, not necessarily to exact product realities, right?
Raphaelle d'Ornano
>> Well, look at how the market, if I may, reacted to the Citrini Research report.
Gemma Allen
>> Insane.
Raphaelle d'Ornano
>> I mean, isn't that insane? And you have Citadel that is responding to that article. I find that ... and the French press, you know I'm French, actually covered that. So it's like, wait. So I'm more surprised by how people are reacting, because right now everyone is trying to get a sense of what is happening. We have been talking about this for a month and writing on that. There's a new reality. It is called orchestration. You only need to understand what is being played out. You can have actually very clear analytical frameworks to understand and not go like, "Oh my God," and panic on every single announcement, because otherwise, that's not how you can run an economy. Like, "Oh, this company is down." I mean, poor companies, right? American Express, Capital One, Visa. I mean, they have nothing to do with Anthropic. They get like oh, minus 5%, minus 7%.
To be a bit provocative on cybersecurity, I do think that security is moving, number one, to the cloud and is moving to the LLMs, and security is becoming a feature more than a product. And so I do think that even though on the CrowdStrike specifically, there was overreaction and the product is different, but I do think that CrowdStrike, Zscaler and their earnings this week are trading in very high territory, and that SaaS cybersecurity is not immune to what is happening, in the same way that JFrog completely went down last Friday on the day that Claude Code Security was announced. So I think there's always a bit of truth and a lot of panic, but you must always ... I like looking for the bit of truth and use that as like, oh, okay, what's next, because the market is not completely rational. There's a lot of irrationality, but there is also a bit of reality. So let's track what is true, and then it's our work to pull this thread and try to make sense out of all of this.
Gemma Allen
>> If anything ... maybe we'll end on this ... it was a great week for Substack, because it's showing the power of that platform. So we are looking forward to seeing what you have to say on Substack this weekend, and look forward to having you back here in two weeks. And again, I know we say this all the time, but God knows what the future will hold between now and then.
Raphaelle d'Ornano
>> Thank you. Thank you.
Gemma Allen
>> Thanks so much for listening to AGNT. Tune in next time.
AGNT Podcast Ep. 4 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 to theCUBE and NYSE Wired.
Raphaelle d'Ornano
>> Thank you.
Gemma Allen
>> What an unbelievable week. I mean, this morning, things were so loud on the options floor behind us that we had some guests who were struggling to concentrate on the interview.
Raphaelle d'Ornano
>> Wow.
Gemma Allen
>> And there's a couple of names that we've heard a lot this morning from that floor, and we're going to talk about them, the obvious ones, Nvidia, Salesforce. But before we go there, I want to talk to you about one of my favorite reads of the week, your Substack, and specifically, as a creative person myself, Figma. So you have done a bit of a 360 on Figma, right? You were a little bit on the fence and now you're a little bit all in. Tell me what you're thinking.
Raphaelle d'Ornano
>> So Figma is a really interesting company. Look, it was one of the flagship IPOs, one of the flagship SaaS IPOs of this year, I mean of 2025, way before this whole SaaS-pocalypse. And when it went public, you had a company that had exceptional figures. The numbers that they were reporting in terms of like ARR growth, net retention, margins, those were outstanding. I had underlined the research that I had done, that the numbers were disclosed in a very selective way nonetheless, and that the numbers that were effectively like wow were not on the whole population. Anyways, the company at that time, at IPO, was priced to perfection. It was priced much more than companies that were trading at the upper end like Datadog, ServiceNow, which was trading very well at that time, Snowflake, and I was like, "Wait." This company, which has impressive numbers, if you put aside again the fact that it's selective, which matters a lot, it was priced for this orchestration future as if it had already earned its orchestration future. So it was like, it's not that I don't like the company or that I don't believe in the product, which seems to be very good, though I'm not a user every day, but does it warrant that valuation? I was like, no. So I had a very bearish view, and the bearish view played out effectively during the year because Figma was down approximately 70.8% up until its very low, which is ... I mean, I know a lot of companies got crashed, but I think minus 70.8% is one of the big cracks. And then last week, in the middle of all of this noise around SaaS is dead, what is happening to SaaS, they launched a very smart partnership with Anthropic by which they're able to really integrate themselves in the orchestration graph ... and the word matters, the orchestration that we have been talking a lot about here ... in terms of Anthropic's world, and they built a reciprocity between Claude and Canva, their product, by which developers that are doing work in Claude are able to see the effect in Figma, vice versa. And so you have this new frontier between not the human UI but the agentic UI, and in a very smart move. And so when I saw that, I was like, "Well, look. This is one of the proof points that I had been waiting for to see whether Figma could earn its orchestration future." And so there we go, and so now my position is quite long on the stock.
Gemma Allen
>> So here's the thing, right, about Figma. It's a great product. I think it's an excellent brain. It has some very loyal customers and users. But if you think about what makes it, its USP, and especially from the perspective of defensibility. That MCP protocol, that's repeatable, though. There's nothing to stop Canva doing that. Adobe, which is already a very highly competitive product in the space if not a leading product, already has some of the integration at its core. So when you think about the USP for a company like that, how do you assess its market position in terms of its orchestration layer, versus how it has something unique to its competitors?
Raphaelle d'Ornano
>> Right. So I think that the whole space is really being shaped right now, and I think being one of the first movers actually matters. And so Anthropic had their famous Developer Day this week, which was expected to be like, "Oh wow, what is Anthropic going to release and who is going to die next," because this is what is happening. Anthropic releases a plugin, stocks go down, and so the question is that ...
Gemma Allen
>> Look at IBM.
Raphaelle d'Ornano
>> Right, who is next? So they had their Developer Day and they showed this. "Okay, we are partnering with the ecosystem," which is not new for me, "and this orchestration graph, you need to be positioned in it." So I think first-mover advantage matters a lot. Number two, not all horizontal tools, not all design tools, are plugged into that yet. So I mean, it is technically feasible, but many things are technically feasible, so having a first-mover advantage matters. And I think, again, this protocol only makes you grow to the extent that you already have a very strong user base. You already have a lot of developer love, what Anthropic has. And so I think it's translating the superpowers that they have in the current paradigm into the agentic paradigm, if it makes sense. I don't think it's magic. It's not a magical wand, but it will help them be even better, and hopefully they will also have the first-mover advantage. I mean, we're very early on to this, and this also matters as long as design actually matters. So maybe in two years this will be completely obsolete, and the moat which exists as of today will not exist. It's a very fast-moving space.
Gemma Allen
>> Well, let's talk with the SaaS-pocalypse for a second, right? Because I read something I thought was quite interesting as well this morning around how we're assessing these SaaS tools from the perspective of defensibility and orchestration and entry points, in terms of how much of your day is spent on only one tool creates some sort of understanding around its future defensibility. But some folks argue that actually the true orchestrators are companies like GitHub, or the real agentic entry point is around those who provide STKs or those dev tools, which is really at the very root, like the source of the supply chain, per se. What are your thoughts on that as a counterargument to the whole SaaS-pocalypse debate?
Raphaelle d'Ornano
>> Well, so I've written a lot that coding is the wedge. Coding is one of the first entry points into the enterprise, and so that's where Anthropic has been very smart, we need to say. I mean, I remember writing in August when GPT-5 was out, and Anthropic ... there was this debate already, like who's winning, Anthropic, OpenAI, et cetera. And I had said, "Well, coding is the wedge, and that is how Anthropic is choosing deliberately to have this Trojan horse into the enterprise, because that's where, if you're able to conquer this territory, then it's much easier to move up stack into different extensions." So I think they've been doing that well, and I think coding being the wedge and being adopted by developers is critical in penetrating the enterprise in a bottom-up way, like, for sure. I think that, as we think around defensibility of SaaS ... which is the number-one question that everyone is having right now, because again, the volatility on the market is just crazy right now, and Anthropic is a big cause of this craziness ... we're still applying the frameworks that are, "Oh, vertical is defensible, and so Procore is defensible maybe, and horizontal tools, well, we can all throw the horizontal tools away." So I don't know, like Box being a name that comes to my head, and there's others of course, but that is a very old-school way, if I may say, of looking at defensibility. So it's done because it's vertical or horizontal, that something is defensible or not. I propose looking at it this way. First, there's a distinction between systems of records and tools. I think that's two categories that we could make, and systems of records can move up into systems of actions. That's path to defensibility number one that we had two weeks ago, if I recall well. And then there's tools, and this is where the Figma example comes in. Well, if you are a tool and if you do not have that pathway from the system of record up, well, how do you plug yourself into this new agentic AI stack, the orchestration graph, and what does that look like? But I don't think to say that because you're vertical, you could be a very clunky, old and non-defensible piece of software. It's done because you're vertical and because you have data, which is the other mental obstacle that we have. So if a company today is vertical and has data, oh my god, it's very defensible. No, maybe it's not defensible at all. And maybe a horizontal company, which business is not to have that data, can find a way to be very defensible in a new reality that is just being built day after day. So I think there's a lot of ... I mean, I understand people are trying to put frameworks. That's a part of my work too, but let's really go into the architectural understanding of the stack. Let's go into understanding what the AI labs are actually doing. And so using that as, okay, this is their roadmap. This is how we as an incumbent, imagine if I was an incumbent software company, this is how we defend and this is how we compose in this new architectural paradigm, and right now we're not having the right lenses.
Gemma Allen
>> So if we think about companies in the system of record to great tools to systems of action space, let's talk about some earnings, and let's talk about the ultimate example of that, Salesforce CRM. They had an interesting week. Again, earnings are pretty good. Everything's looking very, very solid. They're having a good day. What are your thoughts on that company across the board? Do you think that they are an agentic orchestrator? Do you think that this agentic platform that they're all in on is the multibillion-dollar solution to staying in the longevity game? Because some might argue as a product, it's chunky, and the change management is tough. It's not necessarily seamless and fluid from the perspective of disruptor comparisons, right?
Raphaelle d'Ornano
>> So, look. I think CRM is going through a lot of changes. And yesterday their earnings call was very interesting, because they actually had some of their key customers that were in the room with Mark or on Vizio and that were actually explaining how they were using Salesforce in today's context and with their product agent. So first, look at the numbers. This was their strongest quarter ever, and the level of growth at a company that has this level of ARR, because let's not forget that Anthropic's growth is impressive. I'm the first one to say that. But when you're starting from their revenue base, which is huge, I mean, delivering that growth over 10%, it's pretty impressive. Guidance was very good. And if you look specifically at their numbers on Agentforce combined with Informatica, their recent acquisition, which forms this assemble from the context at the bottom, Informatica, and Agentforce as the system of action, well, the numbers, they were growing at like 169%. I mean, 169%, maybe it's not the 1000% that Anthropic is growing at, but it's still very huge for SaaS. And so they're demonstrating that Agentforce has that explosive growth and that it's being used in a way that is fundamentally different than the traditional CRM. And this is what their customers, not me, their customers were saying yesterday during the earnings call, and the market actually reacted quite positively. I mean, CRM is up 4% just before we started today. So I think they're in the middle of a transition that Mark has been extremely vocal about when he started, I mean, Agentforce at the very beginning when no one believed that this could even be a reality, which is something we again need to recognize that we had recognized when we were on the ground at Dreamforce this year. I mean, they have been having this idea. Now the numbers are starting to translate that into real numbers, real growth, defensible margins, defensible cash flow. Let's see how that plays out and let's see how this conversion ... they introduced also this notion of agentic work units, which is the first KPI on understanding how agentic work actually is used in the enterprise at their customers, so let's see how this plays out. Look, I think it's in the course of being built. We need to see. Right now, let's see,
Gemma Allen
>> When we think about a company like Salesforce ... and by the way, I'll say I've used Salesforce and I've worked in a number of rollouts. It's a fantastic product. It's very well integrated. The plug-and-play is phenomenal across a resource like LinkedIn. I think Tableau is a great product too, even as somebody who's Power BI for a very long time. Slack is a great product, but again, in some way they're siloed touch points for enterprise. You might be a Slack Teams user and a Salesforce user. Do you think that the real competitive advantage as in mass integration, do you think that's why companies like Google or even Microsoft ... and I know you have mixed feelings there ... have such a unique position, in that in the agentic world, you touch those products in so many parts and aspects of your day? It's not just a data management system or a productivity tool. It's a bit of everything.
Raphaelle d'Ornano
>> Well, look. I think one of the first laws of value in the agentic world is being close to the user intent. I call this proximity to user intent, and I think that one of their key modes is Slack and Slackbot. And I think that when you're a user and you're effectively expressing an intent, on, "I want to have this work completed, to produce this outcome through Slack," I think it's one of Salesforce's most powerful assets. And I think, again, during the course of making that a reality, I don't have the answers, but it's work in progress. I view Slack as a very strong item of defensibility, and Informatica allows it to ground that reality and to be that enterprise brain. So it's interesting,
Gemma Allen
>> We have to talk about the big, big earnings of the week. A company that has literally defied capitalism, Nvidia. Those numbers, they're monstrous. They're just insane. The stock, Wall Street isn't exactly reacting as I'm sure Jensen had hoped they would, but it is insane to think the growth this company is having, right? We had a guest on this morning, and sometimes when we're doing short-form content, we ask a guest, "Hey, what are you listening to," and I expect them to say bad Bunny or Britney Spears, and then he said Jensen. I was like, there is just such a cult following around this company. What are your thoughts on what we heard versus the market reaction?
Raphaelle d'Ornano
>> Look, I've been quite vocal on the fact that, with this agentic AI buildout, we are going towards what I term as the inference economy. It's one of my favorite terms, the inference economy. Everything is going to be compute, everything is going to be inference, and we're going to have outputs that are going to be multiples of compute. Imagine that we're doing this show and that we're agents. We would be using compute, but of course, we're not agents and we're human persons, but imagine. You can complete a lot of tasks with agents, and agents are fundamentally units of compute, and so multiples of the GPUs that who, Nvidia, is selling. So as we progress into the agentic economy, as all of these workloads become agentic in the enterprise, going back to Agentforce that we just mentioned, you're going to have a tremendous use of compute not for training anymore, but for inference. At the same time, the scaling laws are continuing, and so up until now, you're having massive use of GPUs for training. So Nvidia has the training that's going well in inference. So I think, number one, Wall Street is starting to get used to 70% growth on that base, which I agree with you is not something that is normal. But I think that was kind of, "Oh, well, if they don't have that number, then we will be disappointed and punish the stock severely." And that's not what happened, because again, the numbers were there, both from a revenue standpoint and from a gross margin standpoint. I think that we're starting to have doubts. It's really around, I mean, back to the circular economy thing. And I've always said if OpenAI doesn't get their financing, we have a big problem. And I think just this week we saw, oh, well, actually Nvidia, they're not that convinced that they're going to really invest in OpenAI.
Gemma Allen
>> Also, on the circular financing, we saw some novel deals to speak to with AMD and Meta, these mass codependencies in a whole new sphere. We haven't seen those types of scenarios before, really, not at this level.
Raphaelle d'Ornano
>> No, and I think Wall Street is starting to be ... I mean, it's not starting ... is very nervous around that, and to some extent that's reflected in the earnings. It's like, well, you take all of this around, and who is buying all those GPUs? What is the probability of that financing? Well, AWS now is conditioning that. Amazon is conditioning that to either they reach AGI, which is a very strong technological milestone, or they go to IPO, but to go to IPO, they actually need financing, so what do we do with that? Nvidia is going to give them like $30 billion, but it's like Nvidia needs to give that money to OpenAI because OpenAI is buying their GPUs. So look, where I'm very confident, the inference economy is more than real and this is happening, and Jensen has been very vocal about that. I agree.
Gemma Allen
>> Maybe just to counter the Nvidia being, again, the lion ... and I'm sure they will continue to be the lion, there's no denying that ... but to counter the argument as Nvidia is the only show in town in the inference economy, one thing we do hear sometimes from some of the chasing pack that are chasing the incumbents is that the training world and the inference world, from a GPU chip, from a structure perspective, is quite different.
Raphaelle d'Ornano
>> Oh, yes.
Gemma Allen
>> Blackwell is exceptionally hard to run. Even some of the requirements you have around rolling out Nvidia chips across data centers, it's huge, so they have a supply chain clog issue. So there is definitely, it seems to me, space for some of these chasing pack members to make some ground on them, especially if inference is creating a whole new capacity layer.
Raphaelle d'Ornano
>> Oh, for sure. I mean, I'm very long on AMD personally. I mean, I love AMD, and so they did announce a very key partnership with Meta this week. So I mean, I love that company and I think they have a very strong play.
Gemma Allen
>> The CEO is female.
Raphaelle d'Ornano
>> She's great. So I mean, I'm very bullish on AMD, and I mean, again, they're also doing circular financing. I'm not saying that it's bad. I'm saying the market is nervous and we need to be careful, especially when it's around a company that is OpenAI, which it's not. I mean, circular financing in general, if we can avoid it, that's better. But I mean, that is happening, so let's not make a judgment right here. But there's other companies, of course, there's AMD, there's all of the GPUs that could be used for inference, and there's a lot of developments that are happening around there right now. The chips that are being built are mostly for training purposes, but there is an expectation that it's going to migrate to inference. I've always said that neoclouds are not specifically meant to serve inference workloads. We have CoreWeave earnings in a couple of minutes. This whole market is changing, where I think Jensen is completely right that inference is going to be so much bigger than we think it is, because you have to translate the whole economy into everything is compute. I still hear some hedge funds that are like, "Are you serious? This is really happening? We're becoming compute?" Yes. And so there's an arbitrage between human labor and compute, and then you go back to the physical bottleneck. Well, then do we have enough electricity? Do we have enough this and that? What do we have that could be a physical bottleneck, like atoms meets bits, that could make this inference economy not be true? So this is the nervousness. I don't think it's a reaction to the numbers were bad. The numbers were really good, so I don't think there was any disappointment here. It's more, okay, well ...
Gemma Allen
>> It's a fever dream.
Raphaelle d'Ornano
>> Yeah, the magic effect has gone. I mean, to some extent, the deal that AMD announced, I mean, we didn't have plus-20, 30%. You have companies that are not that great that are going up 30% after their earnings. So again, Snowflake announced amazing earnings last night. The stock was up. The stock is trading, I mean, almost in neutral territory. So look, things are difficult to understand at some point, but the direction Nvidia is going is the right direction, provided that the whole thing doesn't blow up with OpenAI, which we need to be careful about.
Gemma Allen
>> Okay. To finish, let's talk about a company I think we seem to talk about a lot, but let's finish on this, Anthropic. A really interesting week. I mean, their COBOL revelation in terms of what Claude Code can do has certainly had some big impacts on IBM, and it impacts the market broadly, right? They've also had an interesting week geopolitically and politically. I mean, we saw the news around the Hegseth meeting. They have until tomorrow, I guess really, to decide if they're going to bow the hat or not.
Raphaelle d'Ornano
>> Right.
Gemma Allen
>> Who knows? What are your thoughts?
Raphaelle d'Ornano
>> Look, I think from an investment perspective, Anthropic needs to be understood as, "Track Anthropic and you'll get a sense of what is happening." I mean, again, we had Claude Code, then we had Claude Code work, then we had the vertical plugins, then we had Claude Code Security, and it's going one by one. So what comes next? The Developer Day they had on Tuesday, everyone's like, "Oh, this is cool. No one is getting killed and actually they're having a friendlier approach." That is not the question. Again, the question is what is Anthropic building that is the new platform that both OpenAI and Anthropic are building, and how do you fit into that?
Gemma Allen
>> Here's the thing, though, right? You say understand Anthropic and you understand what's happening in the market. But if anything, this week has shown us that really when it comes to the reaction of the markets to these tech companies, often the folks that are perhaps making the decisions maybe don't fully understand the technology, because Claude Code Security is a perfect example of that in terms of what it did to CrowdStrike this week. It's not competing at a code vulnerability level. CrowdStrike actually operates in a very different part, a different space completely, but the market's reacting to heightened narratives, not necessarily to exact product realities, right?
Raphaelle d'Ornano
>> Well, look at how the market, if I may, reacted to the Citrini Research report.
Gemma Allen
>> Insane.
Raphaelle d'Ornano
>> I mean, isn't that insane? And you have Citadel that is responding to that article. I find that ... and the French press, you know I'm French, actually covered that. So it's like, wait. So I'm more surprised by how people are reacting, because right now everyone is trying to get a sense of what is happening. We have been talking about this for a month and writing on that. There's a new reality. It is called orchestration. You only need to understand what is being played out. You can have actually very clear analytical frameworks to understand and not go like, "Oh my God," and panic on every single announcement, because otherwise, that's not how you can run an economy. Like, "Oh, this company is down." I mean, poor companies, right? American Express, Capital One, Visa. I mean, they have nothing to do with Anthropic. They get like oh, minus 5%, minus 7%.
To be a bit provocative on cybersecurity, I do think that security is moving, number one, to the cloud and is moving to the LLMs, and security is becoming a feature more than a product. And so I do think that even though on the CrowdStrike specifically, there was overreaction and the product is different, but I do think that CrowdStrike, Zscaler and their earnings this week are trading in very high territory, and that SaaS cybersecurity is not immune to what is happening, in the same way that JFrog completely went down last Friday on the day that Claude Code Security was announced. So I think there's always a bit of truth and a lot of panic, but you must always ... I like looking for the bit of truth and use that as like, oh, okay, what's next, because the market is not completely rational. There's a lot of irrationality, but there is also a bit of reality. So let's track what is true, and then it's our work to pull this thread and try to make sense out of all of this.
Gemma Allen
>> If anything ... maybe we'll end on this ... it was a great week for Substack, because it's showing the power of that platform. So we are looking forward to seeing what you have to say on Substack this weekend, and look forward to having you back here in two weeks. And again, I know we say this all the time, but God knows what the future will hold between now and then.
Raphaelle d'Ornano
>> Thank you. Thank you.
Gemma Allen
>> Thanks so much for listening to AGNT. Tune in next time.
