Power Lunch - AI Financing Deal, Trane Technologies CEO, Planned California Pipeline 8/10/26

Episode Date: August 11, 2026

Markets are moving lower on Tuesday amid lingering doubts that the U.S. and Iran can reach a broader resolution to the Middle East conflict. Brian Sullivan and Kelly Evans are joined on set by Yardeni Research President, Ed Yardeni, to get his take on the $500 billion AI financing deal announced on Monday and where he believes stocks will go from here. Meanwhile, Trane Technologies CEO, Dave Regnery, also joins the anchors in studio to discuss his company’s latest earnings report and the growing demand for thermal management and cooling systems in AI data centers. Later on, Brian gives a preview of his next Power Insider piece that will cover the new proposed gasoline pipeline that will stretch across the U.S. from Texas to California. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

The $500 billion opportunity in AI, welcome to Power Lunch alongside Kelly. I am Brian, Invidia and a bunch of big Wall Street firms want to turn computer power maybe into a commodity like oil or corn. It's a massive opportunity, but is there a lot of risk there for you as well or maybe just another reason, Kelly, to own shares of the big banks? And the company cooling, cutting edge technology like data centers and pharmaceuticals. They just raised full year guidance on revenue and earnings, train technologies, Baron CEO, Dave Regnery, is here exclusively. All right, so we have got a big, big hour ahead, but let us start with a massive $500 billion $1,000 Nvidia financing deal with six of Wall Street's biggest power brokers. In an exclusive interview yesterday afternoon,

Blacklocks, Larry Fink, likening the current moment of financing data centers to the introduction of mortgage-backed securities in the 1970s. Invidia, CEO Jensen Wong, also saying that Nvidia's AI platform is an investable asset, and Goldman's Sackerel,

David Solomon weighing in on the road to get there. Will it be a straight line? No. Will there be points, to John's point where spreads widen out and it feels like things are going too fast? Yes. Will the returns from all of these things be ample? Of course not. There'll be winners and losers. But that's what the capital markets do.

And the capital markets are pretty effective and pretty efficient at getting those things right. All right. So we know all of this might be a little bit confusing. What exactly does this mean? Is it turning computing power into something like I mentioned, a regular old commodity like an oil or corn or wheat? Or is it something bigger, maybe something more risky or more profitable? Let's talk about it from all angles with us to kick things off. Ed Yardinney here on set. Leslie Picker here on set, Christina Parts on Evelace as well joining us.

Leslie, I want to start with you because you've been covering Blue Owl and private credit and all these derivative products which have had a rocky 2026. I think it would be a fair word. What is Wall Street's early reaction to what is being perceived as a new investment class here? I think that's the right way to describe this. A lot of people have been kind of distracted by the headline figure of $500 billion.

The real news here is that they're trying to create a product, a new investable asset class, an asset-backed finance class that could be akin to things we see with air quality. craft leasing or accounts receivables. Those are all kind of established ABF. A farmer. A farmer grows wheat, sells against the future crop to get the cash flow. This is Nvidia transferring some of its financing. Correct me if I'm wrong. And maybe some of that risk to Wall Street and thus maybe to our audience because ultimately I'm willing

to bet these will become things that we can all just buy. So here's a way to think about it is, you know, when you're, You go to purchase a car. Think about the car like the chips, for example. You go there, you can pay cash for the car if you want to, or you can finance the car. And, you know, there's an established depreciation schedule for the car. You kind of know what the collateral will be. All that's been established.

For chips and for compute, that hasn't been established. There are other parts of the AI infrastructure industry where you, you know, you already have ABF established in data centers and other parts of the ecosystem. But in terms of chips and compute, that has not been established because the depreciation is a little wonky. People haven't necessarily wrapped their head around what future cash flows may look like. So what this does is it basically establishes the market. It establishes the product. It has these six financiers who will come in and say, if you want to finance chips, you don't have to go sell tens of billions of dollars in debt in order to pay for the AI infrastructure buildout in order to pay for these chips. we'll arrange a

financing plan for you. I think the big question is going to be at what price. And this is something that Jonathan Gray of Blackstone alluded to that in the beginning you're going to maybe deal with some you know cash flow negative counterparts, the pricing could be high, but over time they expect that to come down

and get more of a, you know, the broader banks involved as well. Ed, does any of this bother you? Well, it's kind of early to really express a strong opinion on it. So far I have the same reaction is the market, which is kind of ho-hum. It's, first of all, it's a memorandum of understanding, and memorandums of understandings

don't have a real good track record recently. So it's just, that's all it is. There was no specifics about how it's going to happen. I see an analogy with REITs, kind of for data centers. We have those already, and those are kind of financial platforms that invest in the data centers and, you know, those have gone up. They've gone down. If you were, because you're widely followed, so let's say a retail investor said, and again,

I don't know if this is exactly who this we talked to, but said, there's an opportunity for me to pick up, I don't know what the numbers, those 8, 10, 12%, maybe less, this couple year, this couple year, two, three year thing, asset back financing deal and the, you know, it's being backed by basically compute some way, maybe rental income streams. Would you have any problem with that, or could that have a reasonable place in people's portfolio? It could have a reasonable place, but I think it's going to be acid. I don't think you buy the acid class across the board.

You're going to have to be pretty selective. And I think there's a little bit of hype so far in this whole thing. Calling it an acid class, it's a factory. It's a data factory. So I don't know that we're really creating anything radically new. Well, I think Christina Portsenov, maybe the new part of the news. Again, correct me if I'm wrong, is that there's been some, criticism of sort of this idea of circular financing, right?

Nvidia is going to pay its customers to buy its own product. This would be separating that. It would be pushing or transferring the financing to Wall Street. And like I said, ultimately, I'm guessing to Wall Street's clients. It's the de-risk, I think, de-risk some of the worries around Invidia, I think. Yeah. Yesterday, Jensen Wong, the CEO of InVidia, said it a few times that it would be third parties, these six financiers, putting up the capital. So the risk

wouldn't be on Nvidia's balance sheet. However, after the interview, which was a great one here on CBC, he did post on X that they would potentially backstop roughly 25%, or could, I should say, 25% of the residual value of a GPU. So that means just the end value of the GPU. They wouldn't be dealing with the losses beforehand if, let's say, a smaller AI firm couldn't pay back the loan or anything like that, but they would backstop the end value of the GPU, which shows that Nvidia has skin and then games, which also adds leverage to its books. So I was just looking the last, the 10Q showed a $3.5 billion in these leasing arrangements that Nvidia has. So there is some point of leverage that Nvidia has taken, but it shows that they're deploying their money.

They believe that their GPUs won't depreciate at such a quick cycle. Jensen Wong even said, too, that it would extend even to a decade. He gave some pricing for some A-100. and some H-100, so just know that these are older chips. And he's saying that even several years later, the chips are still holding a strong value in the market. And that's because demand is so strong. But to Kelly's conversation in the last hour, will that demand stay as strong

when the market is seeing competition ramp up and everybody else is creating their own chips? So does that really create a discrepancy in the value of these loans, even so that you would have to have much higher yield than a lot of these other asset-back securities because we don't know where the depreciation cycle will land just a few years out. And I think one of the things that's confusing to people is we all sense that with technology, things, you know, they depreciate. They, they're less use. But as Shreini Pajuri pointed out, there are A-100s from Nvidia that were made in the year

2000 that they're still being rented out. So as long as there's a deal to pay for a rental stream, It's not like you're exposed to the fact that token costs are 1% of what they used to be. Well, an analogy has been made with the current situation and the tech bubble and tech rec in 1999, 2000. Maybe now we're seeing an analogy more to 2008 where the excesses are going to be in the financing. If they're going to make financing this easy, we might wind up having too much capacity would be the issue. There may be too much competition. I also, you know, in talking to sources, I said, you know, how are you going to assess what is a good credit here? You know, what is worth lending to?

And, you know, the comments I got back was basically like, we'll lend. The question is more kind of what price, you know, people will be able to get access to this financing. The question is kind of what yield will be paying. That's kind of the old sad joke, right? If you owe the bank $100,000, they have a problem. Or you have a problem. If you owe them $100 million, they have a problem. It's always going to be a price in which you're going to be able to borrow money.

I mean, the Ardeni researchers thinking of raising $50 billion through this consortium, and I'll put all these chips in my basement. You too? Because power lunch. We're talking about a power lunch data center. Why not? But, you know, you're definitely going to lock in the costs of this thing, and the question is, are the revenues going to be there to meet it?

If not, then it's more like a 2008 problem than anything we saw back in 2000. Will it help Christina Partsenevelis? grow the proverbial pie. I feel like we're on the McLaughlin group. I put it to you. But you guys are talking, the main point is that Nvidia is going to act as the conduit is going to validate all of these relationships with these

private capital companies, these AI labs, these sovereign funds that are going to be spending. And so there's a lot of untapped money right there that Nvidia is going to validate, say that they have a great relationship with said company, almost act as a credit wrapper. And then that's why the Goldman Saches of the

world will decide, okay, well, if Nvidia's vouching for it. They're also providing a potentially up to 25% backstop on the life of the GPUs. Why not loan to these firms? So it could be an excellent source of revenue over the next little while just at a time when people are starting to question the return on investments, especially from hyperscalers with all the spending. When are we going to start to see it? Here's a whole pocket of new customers that could come in thanks to the validation of NVIDIA. And you can read it how you want, but they're the center of all this. And he has put Jensen Wong has put the company in the center of all this.

Ed, quickly before we go, perhaps this gets back to the larger point. Sundar Pichai has just announced that Gemini, which, you know, is my go-to. I love my Gemini. I love chat, GPT. Yeah, all the time. Do you know how many, can you guess how many, I don't know if there's monthly active, how many users do you think Gemini has every month now? You probably already know this because you're not going to.

No, I actually don't, but I'm a user, you're a user, and everybody uses Gmail would be a user. Exactly. They just crossed one billion. Right. What? One billion. I would think it would be bigger, though. That's small.

There's two and a half billion Gmail users. Gmail's in bet. Yeah, but that's, I'll pull that. That's confusing. Fastest growing product ever. And that's $20 a month at least. I'm afraid because I pay for the big, you know, no, hold on, hold on.

There's a lot, we don't know about that. I mean, first off, it's no way it's one billion unique. That's one and nine human beings. Half the world doesn't have access to electricity or water. Unless it's all the bots using Gemini. No, no, no. There's 2.5 billion Gmail users. So Gemini is embedded in Gmail.

So anybody who uses Gmail is using Gemini, so you can't by default. But he says Gemini app. That's per month. So you divide $1 billion by $30. I'd give you $250, whatever the number is, right? That would seem logical. Because we don't know if that's unique users. One of unique users would be a lot.

It's a lot no matter how exactly you slice it. And I just think that's when we're wondering, because you're totally right, Ed, that there is a risk here that you start to now just overbuild. Yeah. But right now, the constraint is there. Well, so what I was asking, sources, I said, so why is chips, why are chips and compute suddenly an ABF category when they weren't before? I mean, chips is a concept.

Asset backed financing. Asset backed financing category. Like, it's not like chips are new as a concept. Not like compute is new as a concept. And what they said is that the tie between, you know, compute and revenue is demonstrated by Open AIs ARR. Annualized. So quickly, Ed, do we just buy the banks then?

Forget about NVIDIA. Just buy shares the banks because ultimately they will be the ones that probably make the money on this. Absolutely. I mean, they'll be selling the products. So it gives them more to sell. But I think you can also go with financials, generally speaking. I think you could go certainly with industrials and health care.

I mean, all the beneficiaries of all this AI usage. Which is why you. Why you saw it to your point, Brian, that people are seeing this as the financial institutions as a beneficiary and not necessarily in video. Of course, we've regained it since then because of the interview and because of the depreciation cycle not being as bad according to Nvidia. But overall, I think, you know, the money would be going to the financiers at this point. We'll see if the MOUs trickle down. Christina Partsenevolous Leslie Picker. Ed Yardinney, we'll see you in just a moment.

Thank you all. Really appreciate it. We're just getting started here. power lunch. Still ahead from funding the AI boom to following where the money lands like we were just talking about. Our exclusive interview with train technology CEO about the business of keeping AI cool. But after the break, will CPI turn up the heat? What tomorrow's inflation print could mean for Kevin Warsh as he heads to Jackson Hole for the first time as Fed Chair. And Yardinney, we'll be back to weigh it. The major averages are moving a little lower today and the next test for the rally

comes up tomorrow morning with the July CPI report. Your Denny Research President, Ed Yardin, is still here on set with us. Last week was a great week. S&P went back to the highs first time since June. A little bit now. I mean, were things getting a little too good for you? No, I don't mind the market making new highs at all.

And especially since it's based on earnings. I mean, this is first quarter reporting season was fabulous and the second quarter was even better. It included some capital gains, which really added some rocket fuel to earnings in the second quarter. But I think the market is perfectly in the right place. The valuations are appropriate.

And the bond vigilantes have not yet kind of dislodged the rally. I mean, we see where we were at 470 or so we got to. I mean, you get that feeling that they're going to make a go for 5% again. Look, I really don't have a problem with the 10-year bond yield anywhere between 4 and 5%. So I'm thinking on average it's going to be 5%. I think we're back to normal. I mean, this is where rates were before the great financial crisis. But here's the problem, as you know.

I mean, in the 90s, we were running a surplus. Yes. We have way more debt now than we did that. And so when rates are where they are now, you're adding a trillion in interest cost to the deficit every year. You can't close it. There's nowhere else to cut spending. The Democratic Socialism and Medicare, it could get even worse. So that's the part I can't figure out.

Well, as you know, these have been issues for a while. I've been doing this for over 45 years. And throughout that period, people have been worrying about the deficit. When I coined the phrase bond vigilantes back in 1983, people were worrying about $250 billion deficits. Now we're locking in $1.5 to $2 trillion. And the bond deal is still affordable. I mean... Are you joining the MM tiers where it just doesn't matter?

No, no, no, no. No, I'm not joining the money marketeers or the modern modern theory at all. I would certainly much prefer that we do some things significant about the deficit. But to your point, if we're not putting, there's not enough pressure coming from the markets on the situation. It's not going to happen. And it's not going to matter. Look, my position on this is I will worry about the debt when the bond vigilante is worried about it. We did have a scare in 2023 when the bond you won from 4 to 5 percent in three months. But when it got the 5 percent, that's a free market.

And people freely decided, hey, this is a pretty good rate. Before we bring in, Rick, I want to ask you to your point about earnings, because earnings growth, when you look at the numbers, and I don't have all the numbers in my head, but you probably do. The earnings growth that we're seeing is not just good. Isn't it historic in some cases? I mean, the numbers that we're seeing 30% in some cases year over year is, I think, the all-time record or at least multi-decade record earnings growth. We are living through earnings history right now. I've been calling it FEMO fabulous earnings momentum to distinguish it from FOMO. FOMO is what we had in 1999 when nobody wanted to miss out on this amazing Internet story, even though the earnings weren't there. This time around was driving at the market is earnings. I'd much rather have an earnings-led melt-up than the valuation-led melt-up. So I think that's what we're seeing, and I think you're making a very good point.

Should we bring in Rick Santelli? Look, here's the thing. We all can kind of look at the landscape and say the 30 years, what was it, 528 this morning, that was a new, you know, couple decade high. But yet I know that Ed is talking some sense here, meaning that unless there's a strong velocity to the move and the bond vigilante is really, it's obvious this is not the kind of level yet where we're going to either derail stocks or give politicians any nudge to do anything about it. Well, the problem is there's not going to be a memo coming out to say, investors, beware, all the debt and deficits we're ignoring they're going to start to be important. But there is kind of a pseudo way to look at that potential memo, and I call it the corporate spreads. I think when we look at all the issuance that's going on in corporate America,

and most of it is investment grade for the buildout of AI. But to worry about debt and deficits and the note vigilantes riding aggressively when they're overlooking all the issues in corporate America being released, to me, that's the tell. To me, the market could at any moment in time completely start paying attention to interest rates and could potentially pull the rug out from underneath equity evaluations. But I think before any of that happens, you're going to see the spreads widen out. The amount you have to add on to treasury yields to compensate for the extra risk outside of the credit of the U.S. government. And until that happens, I think it's very difficult to get a GPS on exactly when you're supposed to be afraid of debt and deficits having a larger impact on interest rates. Well, maybe we should in a couple weeks, maybe we'll bring in little hats because it's going to be the three-year birthday of the four and a half percent tenure. I know it's fallen under that a little bit.

There's been some volatility. But overall, Ed, we're at year three of bond yields right about here. Right here. And the market hasn't given a you know what. Well, that's because the economy is dealing with it very well. Look, my view is that 4% to 5% is normal. That's where we were before the great financial crisis. That's where we wore before the great inflation of the 1970s.

And actually it's confirmation that the economy is in very good shape. And by the way, it's also confirmation that the market, that the economy really doesn't need lower rates. It's getting driven by productivity. It's getting driven by the AI investment boom. And consumers won't stop. Do you, and quickly before you both go, but CPI is in the morning, is a more hawkish reading, a more hawkish vet? Is any of that a concern to you? Well, I don't think one number is going to matter.

I mean, it could be, you know, we suddenly have a big move up or down in the markets. But I think the fact of the matter is the new Fed chair, Kevin Warsh, surprised everybody in June when he said he is committed to price stability. He reiterated that in July. And he also said that we got to get inflation down because we haven't done it for over five years. So I think the commitment is still very much there. Do you have any expectation, Rick, about the morning? Everyone says to focus on the year-over-year reading. You know, I think that's correct.

of course, these lower numbers, if you look at last month, that monthly number of minus four tenths of a percent, really incited people to start extrapolating what that means. But yes, we've seen year-over-year core CPI coming down pretty much in a moderate fashion. If tomorrow it comes in two and a half percent or lower, I would look at that as being an optimistic greeting, and I would look for investors to be more aggressive on the buy side than the sell side on the interest rate front. All right, two and a half percent or so. You never quite know until you get the numbers. Gentlemen, really appreciate it.

Thanks, Rick Santelli and Ed Yardinney. Thank you. All right. In the meantime, we've got a market flash on Apple. Another executive leaving Apple. Bloomberg reporting, long-time head of pay and wallet services, is retiring. Jennifer Bailey has been with Apple for more than two decades.

Shares of Apple took a dip lower following the news. They're down about 1.4%. All right, after the break, we'll talk more about data centers, cooling them, water use, Europe, up and more with the CEO of train technologies. He's here. He'll stay cool. We're on set right after this. All right, welcome back. We've been talking a lot on this show on this network about the growth of AI and data centers. And why not? It's powering nearly the entire economy. Well, your next guest sort of right at the center of both. He runs train technologies,

provides cooling systems to AI-driven data centers and more. Stock's been a winner up more than 24 percent so far this year. They just raised their full-year guidance on revenue and earnings. And here now for a Power Lunch exclusive is Dave Regnery. He is chairman and CEO of Train Technologies, a semi-recent spin-out of Ingersoll Rand. We know something about spin-outs. There you go. Dave, we're all good. And I want to be clear, you're not just data centers.

Train has been around for 100 plus years. Everybody knows you guys from the other business. How much is data centers AI, whatever, contributing to your overall and your core business? Yeah, well, let me just give you a little bit back. First of all, thanks for having me on your show. I'm glad to be here. Look, six years ago, we spun off our industrial business, and that became Ingersoll Rand, and we rebranded ourselves trained technologies. And here we are six years later. We were 12 and a half

billion dollar business. Today we'll be close to a $24 billion business. So we've had great success. That'd be revenue. Market cap was 30 billion. We're now over, well over $100 billion now. That's been out happening before. Before data center. That's why I said. With data centers, we don't disclose the size of our data centers. for competitive reasons. But I would tell you, it's a very, very strong vertical for us. And we've been strong in data centers for decades, and we'll continue to be strong well into the future.

But if you looked at our second quarter earnings results, we had 37% order growth. And if you look within our largest business, which is our commercial HVAC business in the Americas, we had 50% equipment growth. So we tracked like 14 different verticals in that business, and the great majority of them had growth of over-together, 20%. So data centers is very strong, but I would also tell you that we're seeing growth that's

very, very broad-based right now. There are over 800-ish, 840-ish mega projects in the United States, of all kinds. About 20% have actually begun construction, you know, a shovel in the ground, laying concrete, whatever it might be, which means I'm not a math-wiz, but I think that means the other 80 haven't been started yet. Do you see any signs of cancellations, delays, backlog issues? No, we have not seen cancellations. In fact, our order growth is very strong in these megaprojects, but I would also tell you the pipeline keeps growing. And every quarter, I get asked by the cell side analysts, how are we doing with the mega projects? And we're really good at triangulating on how to close these. But that pipeline continues to grow. Now, a good chunk of that

today is data centers, but we're seeing nice activity there. When you see Jensen Wong lining up a half a trillion dollars of potential finance. I mean, are you salivating? I'm smiling. You are smiling, yes. I mean, if we show the stock chart, that's over the past year, but if you zoom out over the past five years, it's been a rocket.

So that's obviously where people are most excited to hear about those capabilities. And what can you tell us? Because even as more and more money is going into this, there's more and more local opposition to data centers. Yeah, we work very closely with hyperscalers and Navidia with the chip manufacturing. to develop what's called reference designs. And these are data centers of the future.

And think of the future as two to three years out. And it's really about how you make a data center from a thermal management system more efficient than it is today. And it's a lot of fun. I can't tell you all the things that we're working on, but you can't expect that the data center of tomorrow is even going to be more efficient than the data centers out there today.

Say that in a way that the voter would understand or the person in a community who's worried that their bills are going up, their resources are going to be taxed? The rivers are going to run. There's a lot of misperceptions on data centers. The first one, as you just said, Brian, is water usage. The majority of the data centers' jobs that we're doing right now are closed-loop systems.

They're not wasting water. They're not even using water other than the prime the system. You'll hear people talk a lot about the temperature that's required to cool the chips. And whether it's 45 degrees C or 31-C, it doesn't matter to us. We're going to develop a system that's optimal to perform at that level. And when you think about, when you think of 45C, right, that's like 113F, okay, or 114F, I did my math right. But you could use a lot what they call free cooling. So in that environment, what you're doing is you're letting the ambient air cool the water.

So you basically run it through what's called dry coolers or smart chillers in our case. And you actually let the ambient versus running a compressor, which saves a lot of energy. Using your technology means that data centers will be less energy intensive in the future? Absolutely. Absolutely. Absolutely. That's our goal. But the chips are going to use more electricity. Well, they are, but there's tradeoffs, right? When we save energy, it's more compute power, right?

So it's not like a building. Like, we do a lot of work in buildings to make buildings work more efficiently. And most buildings waste about 30% of energy that they pay for today, which is just amazing to me. But in a data center, when we save energy, it's just more compute power for the data center. So they could have more tokens that are being generated. Well, it's because we have a lot of energy here in the United States. Maybe we have too much, and it's too cheap. I mean, I'm being a little bit sarcastic.

You go to Europe. It's not the case. Right? We've probably all of our viewers have been to Europe. I know somebody that has record heat right now. Okay, water levels are low. Air conditioning doesn't exist in much of Europe, but data centers need cooling. Europe is a big opportunity. Do you see Europe having the money and the ability to buy your products to cool any data center they want to build? We do very well in Europe. I would say that what we're seeing right now in Europe is the pipeline. So this is the planning phase. has really increased. And think of it as, you know, the equation would be,

do I have power on the grid? Do I have power at the site? Can I do self-generation? Or do I have storage capability? And when you take those three, you start to balance when you could get permitting for these jobs. So we're very bullish on Europe, especially in the second quarter.

Our order rates were up over 20% in Europe. And the pipeline, which is really, gets me excited, is really, really starting to become robust. So how much are you willing to invest in the data center? I'd invest a lot in train technologies. Well, investors, to Kelly's point, have. How much is train willing to invest in this market? Yeah, we've expanded our, what we call our applied systems,

which is a product that would be used in the data center. We've expanded it over 4X in the last three years. And I would tell you that we continue to invest at a very high level. Now, in the HVAC industry, we're pretty capital white. So think of our cap-exp spend being two to, and this year it'll probably be 3% of revenue. but it's relatively cap light, but we continue to expand quite quickly. It's, Dave, it's a great window into what's happening

with all of these data centers in the buildout. Really appreciate you being here. Well, thanks for having me on your show. Enjoy the conversation. Dave Regnery, CEO of Train Technologies. Still ahead, is there a glitch in the software space? While the sector looks flat on the surface, some names are up while others are down double digits

in just the past week. We'll talk the split and the potential winners from here when we come back. The AI trade is starting to create a, clearer divide in the software stocks. Take a look at some movers from the past week where Palantir, Twilio, and Cloudflare are soaring while App Loven, HubSpot, and Monday.com are all

heading the opposite way. So what is separating the winners from the losers? I would think cybersecurity, maybe, Sima, if I just look across those names, Sima Modi, is here to break it all down. You're exactly right, Kelly and Brian. It's cybersecurity and its infrastructure. Just take a look at Palantir. Now, of about 40% from reporting earnings on Monday.

It's showing that it is in the rights time, right place, with that Nvidia partnership just as the market sort of pivots towards open weight models, it's providing, right, that application layer that sits on top of a company's stack. And that was reflected in the company's earnings report, a hundred fifteen percent increase in its commercial revenue. When I spoke to Alex Karp, you know, he really made this point that he sees that growth rate being sustainable, which I think is key for investors and why it's being rewarded.

But even sticking with infrastructure, take a look at a company like Twilio, okay? This is a cloud communication company that allows companies to talk to their customer. Think about the Uber app. API that allows you to find and locate your driver, talk to them. That is being supported by Twilio. What they also said on the call is how they're positioning their AI agents to log memory. So that way, when you call into United Airlines, call disconnects, you have to call back and then tell them their problem again.

This time when you call back, they know exactly why you're calling. It reduces the time you're spending on the phone with a customer service agent. And record profitability, record free cash flow. So there are winners. 78% this year. It's incredible. Is there something in common with the names that are all kind of trading more heavily, shall I say? The names we just shared are all winning in terms of earnings metrics, but the names that haven't, HubSpot, App Loven, Monday.com.

Monday.com trimmed its outlook despite an increase in its second, in its first quarter annual recurring revenue. HubSpot, similar story. It's sort of AI strategy is still not paying off fully. It narrowed its guidance for the second half of the year. So those stories aren't being, investors clearly do not have patience at this point. Whereas back in January, that was a different story. Is the market trying to figure out like which companies might be killed by AI and which companies might benefit from AI? And it also, I think, I think goes to the risk of buying an ETF.

If you buy like an IGV and I'm not picking on the IGV, we show it all the time. But you're just buying a bunch of companies and half the companies may be getting crushed by AI and the other half may benefit from AI. and you just kind of don't know, it feels like this is the time now to really understand every company and what they do. I think. But their use cases, and you know,

so to that point, within the IGV software sector. But to your point, there are sub sectors within the IGV, Brian, there is cybersecurity, there are the infrastructure cloud companies and then software as a service. Think about Salesforce,

which, by the way, is set to report earnings last week of August, including Nvidia. But that will be a crucial moment for the company because it's very much in a turnaround phase. We're trying to,

conveyed the sense that it is in a turnaround phase with if agent force business, and yes, they're seeing an improvement in their annual recurring revenue, but CEO Mark Benioff just hasn't been able to dispel this bare narrative. Take a look at that stock, still down about 25% this year, despite many winners emerging in software. I think this is where the real battleground will be next. It's a little bit easier to separate maybe some of the cyber, I'm using the term broadly, from some of the names whose software is more replaceable by AI. With the likes of Salesforce, there's this tug of war, it seems, between going, okay, well, is it going to deploy AI and become an AI play as opposed to being a disrupted name? And, you know, time will tell, I guess.

Absolutely. And even when you look at the metrics for Agent Force, which is their sort of AI offering, the growth numbers are there. User growth is increasing, but it's still a very small percentage of their total sales. So I think for investors, they're thinking, well, if this business is really becoming a formidable player, why aren't we seeing this grow at a more rapid pace and play a bigger role in terms of driving the overall, business at a time where the frontier labs are doing very well. But listen, leave it up to Benioff, who has great experience and a great reputation in Silicon Valley, if they can make it work. I think this report will be very consequential to understand what the road ahead looks like for Salesforce. And then it can have so many more, you know, Matthew McConaughey tried to sell us on

that, didn't he? Him and Woody Harrelson. They can't help. Maybe they will. August, I think it's 24th when they report. They'll tell us. All right. Seema, thanks. They should be on the call. That would be fun. Seema, Mote. All right, let's get over to Pippa Stevens with a CNBC news update. Hey, Brian, White House officials have reportedly asked the Justice Department to consider a new prosecution of David Hearn, the former Olympian, accused of vandalizing the reflecting pool in D.C. That's according to the Washington Post,

which says discussions are preliminary and that prosecutors at the DOJ are exploring whether it's even possible. It comes after U.S. Attorney Janine Piro drew criticism from the president after she asked a judge to dismiss the case. Airline tracking platform Flight Aware is suing prediction market Kalshi, alleging it's using Flight Aware data without permission to allow people to bet on flight cancellations. Kalshi has yet to comment. CNBC and Kalsh have a commercial relationship that includes customer acquisition and a minority investment.

And dating app Bumble will now allow men to make the first move in messaging on the app, a reversal from longtime policy allowing only women to make first contact. CEO Whitney Wolf Hurd tells Bloomberg, it's a way to court younger users and represents how online dating has changed over the past several years. Brian, back to you. Pippa Stevens, thank you very much. All right, coming up, we're going to stick right with that massive AI theme. Your market navigator will talk about AI power bottlenecks and some of the companies that you might want to look at and maybe invest in.

Exactly. Welcome back. It is time now for our market navigator segment. Our next guest is taking a different look at the AI trade, not a traditional utility play, but something he calls scarce infrastructure. Let's bring in Matt Powers, managing partner at Powers Advisory Group.

Matt, good to see you. Explain why you like Constellation Energy. Kelly, hey, thanks for having me. Yeah, you know, we all know the AI infrastructure story at this point. It's nothing new. You know, still the immediate issue, is going to be electricity. We know that. So you can buy all the GPUs you want, but they're useless without a massive continuous power supply. So where do you get that from? It's coming from

nuclear. You know, that's the base case why I like constellation, energy here. You know, 22,000 megawatts of nuclear generation capacity, largest in the country. I think they're twice, the next in line. You know, the plants already exist. They're already connected to the grid. You know, and that footprint is almost impossible to replicate. So a few reasons if I'm an investor, I'd take a position now. You know, it's attractive entry point. It's pulled back. It's off 25%, 23%, I think, is now year-to-date, forward P.E's at 23 times forward earnings. You know, and that's with fundamentals improving. They just had earnings last week. They beat estimates. They raised their full-year guidance. You know, they signed up some new customer

agreements. And, you know, those agreements that's meta, that's Microsoft, that's Walmart. So long-term, think 20 years. The shares are down 16% over the past year. And I might say, hey, there's plenty of other ways to get more immediate upside in the AI trade. I mean, you can look at the other place. There's Vistra energy. I mean, closest comp you can find and they're fine. But again, you know, we're looking at nuclear. In Vistra's broader, their nuclear, their gas, their retail electricity, and Next Era, another huge power producer. Big chunk of that's coming from Florida power and light. So, you know, you're diluting some of that open market pricing with that. So, you know, if you want to bet on electricity in general, you know,

There's plenty of ways to do that. With CEG, you're getting just a very much more specific, again, scarcity value of the nuclear generation. And above all, I think right now it's trading at a really great entry point. All right. Matt, great to see you. Thanks so much. Appreciate it today. Thank you.

Matt Powers. Brian? All right. Still ahead. Could a $5 billion new pipeline help fix some of California's gasoline problems? We'll talk about news you may not have heard. Love that animation. It is time for Power Insider.

And today, we're focusing on a new pipeline or pipelines that California badly needs. California, as you know, some of the highest gasoline prices in the country. One reason? It lost two major gas refineries the past year. They shut down. So now the state has to turn to pipelines. A $5 billion project from Phillips 66, Kinder Morgan, and H.F. Sinclair, called the Western Gateway Pipeline. The plan, and it is a plan at this point, is to build a $900,000.

mile pipeline from the Texas Panhandle up to the Midwest, down to Phoenix. From there, it will connect with an existing pipeline into Southern California just outside of L.A. Remember, Arizona currently imports all of its gasoline, some from California, the new pipeline, expected to start in 29. Long way to go. But California, probably going to get a gasoline pipeline. Speaking of energy and oil and gas. Some investors, they've been printing money, not of the big names, but on some lesser-known stocks.

Look at these one-month returns. Helmerican Payne up 27.5%. Refiner, PbF Energy, up 29%. This is in a month. APA Corp, formerly known as Apache, 22%, Tidewater and Oceanering, up 23 and 21% over 30 days. And if you zoom out, many of those names,

have doubled in value in the past year. Yeah, we talk a lot about AI, Kelly and all the money made in Nvidia. I get it. But in the past year, some of these lesser known companies in energy, maybe pipeline companies have made more money for investors. And for California, it could solve a problem that you've been highlighting. Brian, thanks. We have more power lunch after the short break. Before we go, take a look at the alternative asset managers today. KKR up 7%. Apollo and Blackstone similar areas up there as well. They're leading the S&P 500. It's KKR's best day since mid-April. Two notable things about this, Brian. Number one, a lot of these names were under pressure as private credit fears worsened. That storm has blown over to some extent. Number two, exactly what you were

saying off the top of the hour. Is it any coincidence on a day we're talking about half a trillion dollars in financing commitments for NVIDIA's buildout that the alternative finance providers are having such a nice session? No, because they're going to sell that. They're going to, they're going to package those things up. Invidia, Wall Street, KKR, they're going to package them up, sum to wealth managers who are going to sell them to us. You're going to make more money. Compute is the new oil. It's going to be a new giant asset class. It's going to trade just like lean hog futures. Meanwhile, extreme heat is hitting everything from cheese to nuclear power over in Europe.

In Italy, scorching temperatures are straining the country's cheese banks where more than half a million wheels of Parmesan Regiano are held as collateral for farm. See, everything comes back to. That's the same thing as the compute. Cheese and compute. Power use of the vaults jump 30% as extreme heat has cut into milk production. And listen, this is a real story. I want to be clear, a swarm of jellyfish, jellyfish have forced a nuclear reactor offline at one of France's largest nuclear plants. Jellyfish apparently swarmed up the river, clogging the intake.

Because remember, new power plants need a lot of water to stay cool. They had to shut it down temporarily because jellyfish were gumming up, gunking up. And Romania has to deal with how the low danube is causing problems for its nuclear providers as well. So altogether, this is what they said, big El Nino season, creating all of these ripple effects. And then, of course, you have the winter to worry about if that's too cold. Yeah. It's been going out with that gas prices there. Cheese and compute are the next big futures markets.

Thanks for watching Power Lunch, everybody. I'll see on Fast Money 5 p.m. closing bell. right now.