137. Beyond the Interface: Dreamforce, the SaaSpocalypse, and Who Owns the AI Production Stack

Hello, welcome to the cube pod episode 137. I'm John for with Dave Volante. I'm in New York City. Dave's back in the home ranch in the Boston area. >> I need to do it like this because I'm redeyed. >> I love the goatee development there. Full beard fisherman look. Um you just came back from Dreamforce. Uh we had a great event with IBM the CFO chief human resolver. You were at Dreamforce. We had a lot of action going on here. Just a bunch of people this week. Gemini and the team are just cranking away. But there's so much going on. First, I'll start I want to start with you because you know Dreamforce is always a an event and Jensen was there. Enthropic CEO was there. Jensen made a comment. This is the best sales conference. Kind of a ding on Beni off. Um but um and then Benny said we everyone's following you. It's like walking around doing that town hall vibe. And then uh you know Jen says follow me and he said everyone's following you. Um he's definitely following Jensen obviously and then Daario was there obviously investment made by them and him for anthropic anthropic a threat as some say in the SAS apocalypse to rewrite Salesforce. Um the rise of Slack and the importance of Slack. Last year we highlighted at at Dreamforce how Slack would be great unstructured data for things like graph. I just came back from the Neo4j graph talk New York City and ontologies graphs picking up data from Slack, Gmail, new tools like Instinct is out. You know, you got Meta's got a new tool. So the agents are best of times, worst of times. What's your take on Dreamforce? >> Well, so first of all, Beni off is the keynote master. Um, he is very performative. I happen to love his style. uh great energy and he calls all his people up whether they're customers or employees of his and every speaker at Dreamforce has this incredible energy uh because they're they're trying to keep up with Mark. So, and and you're 100% right. Uh the the banter between Mark and Jensen was fabulous. Jensen's got to be the alpha. Whatever show he goes to, he's the alpha. A couple weeks ago, we were at CrowdStrike and he and George Curts were kind of going at it talking about George is talking about how he's the young gun and Jensen was talking about he had bigger guns, you know, so George was, you know, but but ultimately Jensen's the alpha. I love the way that to your point that Mark put his arm around Jensen and said, "Everybody's following you, bro." And so you get a huge sort of tip of the cap to him. But but I I thought Dreamforce was is actually quite amazing. And here's why. A couple years ago, George Gilbert did some deep research and started to uncover how Salesforce wasn't just a collection of Tableau and Slack and all these other acquisitions that weren't integrated. They were actually doing the engineering work. And even though it was early days, they had a vision. Mark was allocating the the R&D and doing investments to bring that all together with the system of intelligence we called it. Well, we saw that on full display uh this week at uh at Dreamforce and and so the other huge thing is headless, right? Headless 360 and you saw Claude now generates the user interface >> using headless. So you don't even have to use the interface from Salesforce, the native interface, but you can same thing with Slack can now generate an interface. What does that mean? the the new data partner that we have, Qualitate. Wait till you see this thing. John, you've been on boarded. It's like crack. I'm running surveys right now. I'm running one survey on Service Now control tower. I'm running another one on Coreweave to set up for our event there. Anyway, Qual just ran a survey around agent force you just this week. they dropped it. And what it showed is that customers actually expect spending to go up as a result of headless because you're using even though you're using another client surface, you're actually using more data and generating more agent activity from Salesforce. So it could actually escalate. The big question is pricing. How are they going to price this thing? Is it going to be seatbased pricing or consumptionbased pricing or userbased pricing? Is it going to be outcomebased pricing? Is it going to be an ELA? and and that's going to ultimately determine, you know, how this thing plays out. And I will say this last thing is software companies are usually really good, John, at preserving their revenue and making sure it continues to grow. >> You know, I think this whole headless thing points to the fact that AI agents are no longer a feature cycle, meaning agents are feature of the enterprise. They're forcing a complete reconstruction of the enterprise stack. We're hearing this even here today. I was talking to folks from Boston consulting group and investors around the AI infrastructures but this enterprise stack from the applications the end interfaces down through security infrastructure and capital okay spending investments and categories. So you know your commentary and your survey kind of reinforces that shift not just for Dreamforce things like Splunk they had their conference the AI infrastructure summit that was in San Jose and all these security issues. I mean, this is the the industry is basically saying, look, it we're moving too fast, but all full steam ahead. What do you say? Damn the torpedoes, Dave. I mean, that's your favorite line. That's what's happening. We got to be safety. You know, go slower. We need to pace the frontier. I don't think that's actually true. I think that's a false narrative. You know, it may be fast, but everyone's full steam ahead and agents are getting more autonomous at exactly the most concerning around this noise around safety, right? So, it just doesn't make any sense to me because what you just said is we're have a composable stack. >> Yeah. >> So, this is really not really talked about. So what is how should people think about this? Because half the population thinks AI is bad. >> Half think it's going to be great and the half of the bad is politicized, but also they're not in the in the no. >> AI is going to be great and it's got the potential to be really bad. So both things can be true. And I think that you know when you listen to Sachs, oh this is all BS, you know, and you listen to Bernie Sanders, oh we got to stop AI, you know, the truth is probably somewhere in the in the middle. Nobody's ever said we don't want to govern AI. I've never heard anybody say that. Um, and of course there's all these conspiracy theories, but one thing we know for sure is that open source is not going to stop. And so I think that, like I say, the truth is probably somewhere in the middle. I don't know what the outcome is. Is it a 0% probability that AI could do some serious damage? I would say absolutely not. You know, >> well, I mean, do you think do you think we're really going to slow down AR? is simply going to be another >> stack component. Safety and governance kind of go together. We covered a lot of governance conversations. Safety just becomes another layer of the stack in my opinion. What what I mean it just can't slow down. >> You can't you can't make it safe until you build it. And so you got to build it first. And so you know what are you protecting if you don't if you haven't built it? You don't know what you're protecting. So people say, "Well, you got to design it in." Well, you got to design it first. And so I don't see AI slowing down, I I I you know, unless the there's some external government pressure to do it. But unless, you know, Trump and Xi get together and say, "Hey, let's slow this thing down." And even then, would you trust that China's going to slow it down? And even then, if it's out there in the wild with open source and open weights, you know, it's not necessarily going to slow down. I think the question I would have John is collectively can the open source community kind of match the frontier models or do they need them I would say it's likely they need them to lead but will the delta between when a frontier model comes out and the open source matches it or roughly matches it will that compress says it absolutely will compress it might go from six months to four months or from four months to two months you know it's not going to go from you know four months to zero you know, but really at the end of the day, you know, open source is pretty close. >> Let's talk about the um Dreamforce real dynamic that everyone's talking about. Salesforce is essentially betting the company on agents. Would you agree with that? >> Oh yeah, they I'm wondering when they're going to change their name. >> Okay, so we agree. Let's just Okay, now do you believe that you got AI Force COA and you got the deeper integration with AWS Google suggest they see this next application architecture? What's your analysis say on that architecture? Cuz it's an interesting move to say, okay, betting on agents is one thing. That's clearly what they're doing. But what does that mean? Cuz they're betting the company. That's what's happening. What does it mean? >> So here's what it means. So headless becomes that system of intelligent. That's where the data is harmonized. The process knowledge gets gets injected and and and that becomes we this is really one of the first true enterprise harnesses that we've seen. you know, you see harnesses within the LLMs, but they're not deterministic. So, this is really the first example of that AI Force, which is what they announced, that allows any client uh surface to basically create the interface. And so, I think the reality is this stuff is is not mature. I mean, Agent Force is still in its early days, but there's there's clear business value. We heard it at Dreamforce. We heard it from the qualitate data. people are getting, you know, improvements of 30 to 40% more uh of benefits. They're they're they're just knocking out half of the inbounds that are crap inbound leads. Um you know, they're they're agents are doing more and more calls, you know, so you're very clearly starting to see value there. Um the other thing I say is that it's unclear with spending right now. There's no evidence that spending is shifting from the existing sort of Salesforce budget in into agent forces. We're not seeing that cannibalization. But let's face it, it's mostly early days and it's mostly proofs of concept. Over time, we'll see how pricing plays out and how that affects Salesforce's business. They did announce that they're moving out of the 50s, 50 billion into the 60s. You know, they upped their projections. you know, the stock went down quite a bit. Um, so but but the bottom line, John, is to your point, they now have this full agentic stack with it's and much of it is their own, but a lot of it to your I love that word composable. When you think about their relationship with like u dream uh uh uh data bricks or snowflake, those are composable elements that work well within the snowflake uh the Salesforce ecosystem. I mean, composable agents is definitely what's what they're talking about. What's interesting to me to your point is that, you know, they could be the agent interface, but they don't have to own um every agent interface, right? Because agents have standards, but they have the data. So to me, you know, with SAS with Salesforce becoming the poster child for SAS apocalypse because they are probably the number one example people use if you look at it and they're a proxy for like a workday and everybody else, but if you if you look at what they're doing, their advantage might be the remaining system of data, the workflow, the identity, then and the business context underneath it all. And that's why I think Slack becomes a potential agentic front door. unstructured data works well with crafts works well with kind of the infrastructure that AI likes. So it' be very interesting to see how they develop that. I mean that's going to be the the the marker for me is okay how does Slack evolve? >> I I I think if you are here's how I think about it. If you're using Salesforce for not only u a system of record but also for your workflows and it it's deeply embedded into your processes >> and again the qualitate data bears this out. It's highly likely that you're going to do a lot more inside of Salesforce. If however you're just using Salesforce as the system of record and it's not deeply embedded in your workflows and your processes through headless, you could actually kind of recreate a lot of that function that Salesforce has outside of Salesforce. And that's where the SAS apocalypse actually comes into play. So the the the interesting and unknown right now is, you know, what's that going to be? What's that midto long-term effect? I I think it's undeniable that they're that that that some of that is going to you know occur with other you know whether it's uh harnesses claude code cursor uh uh open AI you data bricks etc. If developers have an affinity toward those and Salesforce is not deeply embedded in your organization I think it's highly likely that some of that Salesforce data through headless is going to be accessed and you're going to see some siphoning off. The question then becomes is is the incremental business that Salesforce is going to get in the momentum because they're early on in this cycle. Um is that going to offset any of that drift? >> Yeah, the whole mode is based on the data workflow identity and the context. Okay, that I would say that would be a fair statement. But we what we don't know yet this is what I'm watching closely because talk about the AI infrastructure and the boom and the agents that'll come out of that as more computing power emerges. data. Here's a thought exercise. Right now, the value is the data that they have and and that's true for most enterprises. What's the 1% roughly of the data in the enterprise is unlocked because OpenAI and the frontier models haven't yet ingested that proprietary data. But what happens when agents start to learn about how enterprises actually work over time? Then you have a new training inference data set. Okay. And so I was just talking to someone about this um Spirit Airlines bankruptcy. There was a bid for the data you know 20 years of data roughly 10 20 years worth of data of every transaction and the the buyers of that data were for training purposes. Okay. So that kind of brings up the commoditization of the system of record. So at what point does that become disposable? So that's a a interesting thing because you say, hey, the frontier model's trained, use GPUs and Nvidia and then inference became the killer app. Okay, that's great. We're seeing that boom now. But what happens in the SAS apocalypse scenario that you're talking about where okay, what if their advantage is all those workflows, all that systems of record? What if the agents figure out how companies work, >> right? Well, >> exercise, but you can say given the fact that these new tools like Instinct are doing transactions, changing your flight seat for you, um using iMessage interface, WhatsApp is saying, "Hey, you know, I don't want a middle seat cuz I just got a boarding pass. I got a middle seat. Change it." And it just goes and does it and cranks on the support line, does the does the the work for you. That's happening right now. I just talked to someone who had their middle seat changed by instinct, the new AI agent app. So that's interesting scenario. So I am I'm going to watch this very closely because if this AI boom continues as we saw in the AI infrastructure summit, chips, memory, networking, optics, database are they're all moving at full speed. The next step is can they in train and infer generically what an enterprise does >> well and and you know Beni off understands this well. You saw you might have seen the Finn acquisition. So the Finn acquisition is a standalone agent to counter you know Brett Taylor's company Sierra and Finn is you know it's best to breed from an ease of use standpoint. So so Ben off sees this coming and so that there in lies the Finn acquisition but the the whole point of our breaking analysis and you just nailed it this week John is it's called it's titled Salesforce after Dreamforce how CRM the company can grow beyond its own interface. And this is the crux of the issue. We we're seeing this evolve. And I got a shout out to to George Gilbert. I mean, he nailed along with with with Jeffrey Moore for the economist article that they're doing. He nailed the the emerging software stack and it's evolved, but we're now seeing we saw it with Snowflake, we saw it with data bricks, we're seeing it here with Salesforce. You can certainly see the hyperscalers, the Google and Microsoft and even Amazon to a great extent all coalesing around this new software stack. It's changing right before our eyes and Salesforce is very clearly in the mix. >> Yeah. And it's going to move from agents to to me infrastructure. I mean digital infrastructure, you're talking about digital infrastructure at this point. When you talk about headless systems, composability, you're talking about agentic infrastructure, not applications because those will be managed or built on the fly maybe. So, we'll see. Um, the infrastructure boom continues. Dave, again, the the data is coming out. Caruso 30 almost $ 31 billion valuation. You know, the infrastructure continues to go. Uh, >> we knew them when they were like less than a billion, John. I mean, it's >> like on nobody even heard of Cruso. I'm not even sure they were funded yet. It was like a gleam in his eye. >> Post post on AI got a almost a $600 million value uh raise at a $5 billion valuation. Series C by the way. Uh >> no bubble. No bubble though. >> There's no bubble. I had three here today and I said to them, is there a bubble? They said absolutely not. No bubble. Brian will bust your chops. I guarantee you're gonna get a text. >> Brian, no bubble. >> Enjoy it while it lasts. People, >> for the record, guys, there's no bubble. Dave, there's no bubble. I just interviewed managing director of digital bridge and I'm investing in digital infrastructure. He said no bubble. There's too much demand. Now, he lived through the internet bubble and he lived through the telecom bubble. He says there's too much demand and the financing of these unsecure not what was the word um where you don't have AAA ratings um these these these these valuations are not standard so people are worried about >> but investment he's talking about investment grade >> investment grade yeah non-investment grade assets so uh and he was also an early investor on both the core weave rounds the two billion and the $7 billion financing and they'd looked at it from a credit stand And he it's interesting. He said they finance the chips and the physical buildout. So they got the real estate and the physical plant as well as the chips. They they they holistically finance both in the analysis and again non-investment grade. So obviously they can apply their domain expertise. But this is kind of what's going on. The financial community is leaning into these um these non-investable grade deals and make them bankable. That's what's going on with the NeoCloud. That's why I got to ask you. So I'm looking up the definition. What's the definition of a market bubble? A market bubble is an economic cycle where asset prices surge far above their true intrinsic value driven by speculation and emotion rather than real financial fundamentals. Let's let's actually unpack that. Well, I just told you that the real financial people are valuing a non-investment grade asset, okay, as valuable and it's delivering >> kind of like they did with housing. >> So then it says the anatomy of a bubble displacement. >> Fraud fraud is different than actual. >> Okay. But the anatomy of a bubble displacement investors get excited about a new technology. Boom. Prices start rising and gain momentum as more buyers join. Euphoria, asset prices skyrocket. Everyone believes prices will rise forever. There is no bubble. Uh now profit taking profit taking hasn't really started yet, right? >> It's not everyone and everyone's not running to this. It's very unique financial market. >> But here here's what I say though some interesting dynamic John >> and by the way the demand curve. The question to ask is how big will the tide be? It's high before we go there couple observations. Moore's law was deflationary. Jensen's law is actually inflationary. He said this to Beni off. He goes, "People are bummed out because they they locked in forward pricing a year ago and now they could get 30, 50, 90% higher prices." So it's actually inflationary. The second thing I would say is there is monetization very clearly. SpaceX AI, the hyperscalers, anybody who Nvidia, anybody who has compute can monetize. The question is, are mainstream enterprises monetizing? And I think the answer is some are, but most are not. And that is the thing to watch. So, like I I've always said, bubbles don't always burst. Sports teams, they've been in a 30-year bubble and hasn't burst yet. >> That's called a market. >> I think I would call that a market. >> Market. Yeah. >> Okay. So, Gardner projects worldwide AI spending growth of 49% in 2026. um the capital spending far beyond the found is going far beyond the foundation model. You're starting to see different kinds of investments and the investment thesis is here in in Wall Street at least that I'm observing is migrating from who has the smartest model to who owns the production stack. What I mean by that is to your point about ROI and this is where I don't think it's a bubble is that whoever can get stuff into production. So if I'm a Neo cloud and I got a hundred billion dollars in orders for offtake, I better I got to build. So the risk goes to can you put that into production and service that deal or deals. Now there's two types of deals. There's the physical deal building the data centers. There's risk involved there. Then there's the chip financing. And what Jensen was pointing out to your point is that the chips aren't actually losing value as much as people think. So they're the they're creditworthy. >> They're gaining value. They're increasing in value. that doesn't make it a real estate thing because the investment thesis is migrating from who has the smartest model to who has the production stack and the chips and the real estate kind of work together right so when you look at like say for example this came up in my last interview fiber optic is the most commodity product on the planet well guess what people are acquiring land and power for put a data center if it doesn't have fiber optic it's dead so the fiber is becoming valuable again so things like that copper the price of copper high. Why? Because they're all being used in data centers. Well, guess who has a lot of copper? All the telos who are ripping it out of the ground. Their trash is now treasure, right? So, you started to see the asset balance sheet discussion. Um, I don't think that looks and smells like a bubble to me. That's demand curve. So, the infrastructure will be interesting to follow the money. Where's the spending going? And what's the investment thesis? It comes back down to what you said like multiple pods ago about ROI. If you don't have stuff in production, aka data centers, you're screwed. The data center is the new enterprise, Dave. It's the new on premises. So, okay. If if you believe that to be true, then there's the unlock of the enterprise market. >> I think that's true. Modeled market. >> Well, very clearly onrem is getting an unlock right now. You can see it with with Dell, with HPE. Of course, a lot of Nvidia stuff is going on prem. I'd say that, you know, even the even the neo clouds, even though it's not on prem, I mean, you're seeing alternatives to hypers scale. Uh I, you know, and this this the narrative about the smartest model is interesting because, you know, the every LLM says, well, it's not about the smartest model, and you just made that statement, but but everybody's trying to get the smartest model. And like Sarah Frier is saying, the one who has the the the best compute, the most compute access, which Sam Alman's been trying to lock up for years, and now, you know, Dario is doing the same, is going to help them get a smartest model. I actually think smartest model still matters. It seems to be what we're what we're talking about because that's what's driving the frontier. Now, the question is, is that tail off? You know, >> well, the investment thesis was who's the smart, but I think everyone wants the smartest model and then they're going to have our power laws kicking in. Again, back to our multi-year projection on this one. That's true. >> That's true. So, I'm not they're not investing in the smartest model. They're investing in who can do the production stack. >> Fair enough. But but the smartest the smartest model is what's driving the demand because that's what's driving the compute. >> Well, because the compute needs to go into smaller footprints like the edge. We talked about the hypercon converge edge. Who's going to put there? It's a chip game, not model game. Now, you have to have a smaller model. What's the smallest smartest model? That's not the investment thesis. That's how do you get it in production? How do you turn a telecom or a headend into into a >> fair how do you deploy it? I I agree with that. >> Um so you know models are becoming specialized and embedded. I mean look at anthropic. Um you got Salesforce COA launch. Anthropic got the biology lab. These are like the beyond the chatbot era, right? Like so it's like >> way way beyond the chatbot. >> These are workflows. These are workflows and domain expertise and operating environments. That's where the AI infrastructure production comes in. So you to me that's the dots that connect there. >> Well, okay. What else you want to talk about? >> Well, I mean agent stack thing because like to me, you know, um the thesis that came out of Dreamforce, which we've been tracking here as well, and you've been tracking with George, is you got models, agents, context, and data. the Neo4j event, you know, they're all engineers, so you can read the tea less a little bit there. The context graphs are huge, context layers, that's a data thing. And then orchestration, security, and governance, okay, and then what infrastructures that run to put it in production again. So if the enterprises aren't going to be standing up, you know, gigawatts of factories, they'll use the neocloud. So I think the neoclouds like Michael Dell's been taking advantage of will be the enterprise new data servers, right? So the data center. So, you know, it's there's a strong argument that these data centers are the new enterprise >> because they can do multi-tenant features but give single tenant experience for workloads. >> That's to me the number one thing. >> Okay. But but so tie it back to Salesforce. So, so coming out of Salesforce, the premise is very clear. I mean, their next growth opportunity is going to come from customers that probably spend less time on the Salesforce interface and having agents do more of the work and and the the we've seen a progression. Benny off talked about this. We went from command line to graphical user interface to web to to mobile and now we're going to sort of agentic creating on the fly. You're seeing Claude create the interface. You're seeing Slack create the interface. virtually any harness is going to be able to create the interface and so so that that's going to drive token demand. Where's that token demand going to come from? Back to your point core infrastructure in the data center and a lot of that data center you ties into Amit's uh sovereign AI stack. A lot of that is going to be on prem. There's no question about it. >> Well, I want to talk about some breaking news. I just saw CNBC have an article just hit Nscale announcing IPO of the on the NYSC. I was at their investor meeting. I saw their numbers, saw their traction. Great management team. Um it's out there. They're going public. >> Wow. Great. >> On NYC. We'll be in the building. We'll be tracking that puppy. So, they're an example of kind of the playbook Coreweave built, which is fully integrated um business. Again, they came out of the whole cryptocurrency mining um firm that they started. Now, they went full scale. They bought any scale company we've been tracking with Robert Nishimar's company. They're now in the stack fully integrated. Um so, huge cloud provider opportunity for this AI infrastructure again serving enterprises. So, you have that vertical specialization. Then you got more of the horizontal neo clouds. Um so, Nebius and Cororeweave have a new competitor. Okay. And this IPO is expected to be pretty smashing. Um, and we'll see how this goes. But, you know, we'll be here on the ground. You know, I don't know. I'll have to ask Brian Bal at the NYSC how it all works. But I think when they flip, they're finished through the SEC process. They could go within weeks. So, you know, we're September 18th, Friday could be two weeks IPO. You'll have to fly in Dave and check it out. Jee and I will be certainly on the ground doing live streaming from the floor here. um capture all the data and all that. >> Do we know the date of the IPO? >> I I don't know. I don't know. You know, they're certainly it's the IPOs. They're like the c I'm learning about the cage of the IPOs. That's why I have to ask Brian because um they could just if they're well they got their they're if they're buttoned up, they'll go faster. You know how it works. Um but they're they're stacked up. They just add a new board member uh Fiji Simo, former OpenAI, Instacart Meta. She's in my friend group with Cheryl Sandberg and PaloAlto. So you got the Cheryl Samberg crew. Um Nick Kle, he took over comms from Elliot Shre who was at Google with Cheryl which we did work there with them with the cube and Silicon Angle. So I know them personally. They are strong leaders. So their board is stacked with with uh with talent. So I think this is going to be pretty solid and the fact that they bought any scale shows that the management knows what they're doing because with disagregated serving you got to know the resource. So they they check all my boxes, right? So I love this company. We'll see how they do. I mean, who knows what's going to go on with the IPO market. It could happen. Anything could happen. Trump could do something in China. You never know what's what the politics scene is. Um but if the market stays the way it is, it should be a smashing success. >> So I just asked Qualitate. I mean, you can you can ask an LLM who is Nscale. Anybody can do that. But what's cool about Qualitate is they have actually have customer interviews. Now, I'm shocked that they have an interview with NScale, right? But it's in their system. And then this one customer says, "We've also looked at Ncale as a burst option for when bad jobs spike beyond what we've provisioned." So that's the only customer example we have. But like I said, right now I'm running a survey on Coreeave, a 15N survey, and we'll see what comes out of that. I want to get the results ahead of our uh of our show there. I like that mixture of expert, you know, smaller ends, highly valuable targeted results, and given our data on the cube in Silicon Angle, we can contextualize the hell out of that. So, you know, this is interesting news, Dave. So, again, >> you're going to see the first contextualization of that uh this week in breaking analysis that George and I did. It was awesome. We took anybody can say what happened at the this this at Dreamforce that was there and anybody can do customer surveys. The key is bringing those two things together and putting them in context and then of course you know mapping him at George's system of intelligence framework which he's just awesome at. So you're going to see that this weekend. >> Well, we got a lot going on in California at the Corv event. I'll be at the proof point event. They're looking at another IPO prospect. Octa Octane. Uh we got the uh UiPath Fusion event. Uh Meta Connect's going on virtual. Um we have to leave. I have to leave now because we're having a meet up. There's a big DJ coming in here. Our room for our big >> just a plug. We got we have a we have an Oracle security event um launching uh next week. I was just out in PaloAlto recording that with all the the big guns SVPs. We had customers in there uh led led by Juan Loza which was awesome. what what they're doing in security is just remarkable. >> Yeah. Well, we got a lot of actually got a lot of investors coming. Got like 60 70 people signed up. We're going to do a little talk with Post on I AI. They got the new chip. They got great valuation. We got um Dmatrix coming in. Um bunch of hot companies. So, it should be fun. Again, >> what's that thing that you're doing? Aren't you speaking at some Dell event? >> Yeah, next week I'm keynoting at the Dell Symposium in San San Jose. Uh I think you're going to be in town for their New York event. Y um Michael Dell is going to come into the NYSC. Obviously, they're one of the the darlings of Wall Street right now. They're up for the year 250% around there, give or take, depending on how the market's down a little bit today. So, like, okay, they're kicking ass and taking names. So, that's good. >> The Dell data the Dell data platform is uh we're we're there some news coming out there that Dell is like the hottest company going these days. It's unbelievable. >> Yeah. Yeah. And so, you know, hey, good for them. They did the right thing. They they work with us. They we advise them. We give them the right decisions to make. They made good calls. They're smart. They have the big ears open. They're always listening to the market. One of Michael Dell's tenants of first principles. >> You know who just wrote a book. Speaking of big ears, Matt Baker. >> Really? >> I'm reading his book. It's unbelievable. And this it's unbelievable what he's sharing. He's like, uh, the book is called Unconventional Wisdom, an unlikely career in tech by Matt Baker. I mean, he lays it all out there. A lot of personal stuff, which is quite amazing. >> What's the tell inside baseball there? >> A lot of inside baseball. Um, it's very cool. Obviously references Jeff Clark a lot, Michael, Travis Vill. He talks about his career. I mean, he's just I didn't I mean, I've known Matt for quite some time, but just sort of in a business case, a business setting. You got to check it out. Um >> I'm definitely going to download that. >> It's pretty amazing. >> Is it hard copy or ebook? >> No, I'm reading it on my Kindle. >> Great. >> It's it's it's wild. And I'm learning a lot, too. I mean, he you know, he was, you know, he's he was a tech a tech guy, tech like it guy, which I didn't know. And he became became the head of strategic planning at Dell, right? He was at Intel. >> Um pretty cool. He went to he he installed like remote access for for Gordon Moore's house which you know where that is. It's right near you know Don's place. >> Right near Don's place. Yeah. >> Don's place. Yeah. >> Yeah. They have a hard time to get connectivity there because it's in in a hole basically in the mountains. >> Well Dave I run to this event here. We'll see you next time. Cube pod. Next one will be 138. Cranking along. We'll see you next time. Thanks everybody.