Power Lunch - Dealmaking in AI and the latest from the Fed 11/3/25

Episode Date: November 3, 2025

OpenAI’s first-ever compute deal with Amazon Web Services saw the company choose Nvidia chips over in-house. Plus, as the AI industry looks for new ways to generate power, AWS taps Cipher for partnership. And Steve Liesman reports on news regarding the Fed’s latest meeting. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

The NASDAQ kicking off November higher as more energy and AI deals help lift the tech trade. Welcome to Power Lunch, everybody. I am Brian. Kelly will be back in early December. Amazon flowing today as it inks a massive $38 billion deal with Open AI. Invidius chips, they'll help power that partnership, both of those stocks rocking higher. Speaking of power, the CEOs of a new name that is popping today up 30% at one point, shares up 350% this year. There's your mystery chart.

The CEO joins us exclusively with the how, the why, and, oh, yeah, the name ahead. And don't lose sight of D.C. The partial government shutdown now on day 34. The funding impasse nearing a record for its longest shutdown ever. We'll ask if we are really any closer to a real deal that would reopen the government. Welcome, everybody. So much to do on this busy Monday. Let us start, though, with Amazon.

It is signing a $38 billion deal with OpenAI to help run cloud computing. It is the first deal between OpenAI and Amazon. And it's even more interesting because OpenAI, remember, is partly owned by Microsoft, which is a big Amazon competitor on the cloud side of the business. Bottom line, with all these deals, it's getting pretty cloudy on who is doing what. So let's open up the clouds a bit, try to shine a little light in. McKenzie Seagalos in San Francisco to lay out exactly what's going on here. Mack.

Hey, Ryan. So what we're seeing is pretty notable for OpenAI. They haven't had this flexibility to sign whomever they wanted in the Cloud Wars. They're now for the first time teaming up with Amazon. And you said it. Amazon is not only a big rival to Microsoft Azure Cloud Services. It also actually is a big investor in OpenAI rival. Anthropic, they've invested $8 billion into them. But this really shows just how much Amazon

needs notable customers like OpenAI in order to build up its AWS strategy. It really needs to bring all those AI customers in-house. So even though it's not invested in Open AI, they want them using their cloud services. Well, I talked to us about this Amazon, Open AI, but Microsoft thing, because what it feels like is one of those like real housewives of Silicon Valley where people are kind of dating everybody, or not at the same time, but kind of at the same time, because Microsoft owns part of Open AI. They're fierce competitors on parts of the cloud side with Amazon, and yet this deal is with Amazon, but it's partly owned by Microsoft. I'm so confused. Yeah, I mean, part of it, a nice way to think about it is these proxy wars being waged by the

tech giants through these AI startups. So you've got Microsoft that's invested $13 billion into open AI, a return, by the way, that's 10xed as of that deal that they renegotiated last week. And then you've got Amazon invested into Anthropic. And both of those companies, notably Microsoft and Amazon, like their AI strategy is really being the cloud backbone to these players. A lot of the compute that they're providing is how they're capitalizing on an AI strategy, but that is different to having an in-house LLM that is competitive, like that Anthropic, like that OpenAI.

So they're kind of betting on those players to fill in that piece of the strategy for them. But to your larger point here, now we're seeing this cross-pollination where you've got Microsoft, you know, its investment into OpenAI. OpenAI is now using cloud compute from Google, from Microsoft, from Amazon. And that's because there is this voracious appetite for compute. So Open AI needs to diversify who it's working with. And on the flip side, Amazon, because it isn't trying to like get its own chatbot out there, they need to get more AI customers on their cloud. So it's not who's dating whom. It's everybody's dating everybody. So that the entire commune, for lack of a better term, can win.

But you know who is exclusive? Open AI only appears to be working with Nvidia GPUs right now because they signed with Google Cloud in July. And they did not want to use their in-house TPUs. So they're using Nvidia there. This Amazon deal, at least for now, it is not using its in-house tranium chips. It's using all Nvidia. Now, it's my understanding with people who are close to this deal that they are willing to potentially work in the AMD chips that they signed. to deal for. So maybe the new capacity it's being brought online because AWS, Brian,

is building out new data centers specifically for OpenAI, just like they did for Anthropic, doing it for Open AI. And so they might be able to port some of those chips both from their Nvidia deal, from their AMD deal. Just say the word Tesla. Tesla. Good. We got all the Mag 7 in one story. Nailed it. That's it. We got all the, well, we didn't say Netflix. We'll get that in the next story in just one second with Laura Martin, McKenzie Seagalos, make it sense of it. Mac, thank you very much. All right, let's stay with that story. Your next guest says there is no clear winner just yet, because maybe everybody's winning. We're also kind of largely waiting

on the dust to settle on the latest multi-billion-dollar AI deal. That aside, where are the best places for your money right now? Joining us, Laura Martin Managing Director, Senior Intrinet Analyst at Needham and Company. Laura, I tried to, you're the media whiz, so I tried to get like the TV show reference in there? Was it directionally correct? I mean, or does it, it seems like everybody is kind of semi-dating everybody? I would make a more practical point here. Amazon is the only cloud company that does not tell us their capacity constrained because they doubled their power over the last three years. Google tells us their capacity constrained. Microsoft tells us, so part of this Open AI is they do not want to fall behind.

So if they can get to Navidia chips by using AWS, they are incentive to do that real time. And AWS has capacity. So I would say this is a much more practical than long-term view, and that they cannot afford Open AI. Now, the other thing is, Brian, Open AI is going to do $20 billion of revenue. And I think they have a trillion dollars of commitment. So let's hope they can meet this $38 billion of commitments they just committed to AWS. You know, the Amazon power story is one we're going to get to with our next guest in just a few

minutes, so I don't want to front-run that. But let me go back to the word you just said. The H word, hope, I have been told hope is not a strategy. How serious, how much do you believe this? For Open AI? For all of it. So, so the point I'd make on that, let's just stick with meta, Amazon, and Google. They are telling us that they are linking revenue to capacity, that when they tell you their capacity constrained. They're telling you they have excess demand. And you watch Google's margins and also Amazon, they're going through the roof, which means each, because of the demand, the over-demand, they're raising prices at the speed of light. So we're getting expanding margins. So I believe there is a positive return on capital for every one of these increases in

capital spending that they are projecting. The capital spending numbers continue to go up. And I feel like people can't, it's like every day someone's raising their capital spending estimates on these names. The stocks keep going up. But here's the only downside of that, Laura, I kind of tweeted out that the mag seven, forget about the mag seven. We're going to call this the eight is enough. That's a TV reference for people of a certain age, by the way, market, because there's eight stocks that are now about 37% of the entire S&P 500. We've never seen this kind of concentration, at least in modern times. Laura, does that concentration worry you?

So I'm going to go with no, but what I am going to do, the implication here is that hurdle rates are rising. In order for the capital market to allocate you capital, you better have revenue growth like CoreWeave did, which is 12 to 300 to 2 billion to 4 billion in four years after founding. That's the kind of numbers that the market is assuming

these hyperscalers are going to get. And that means if you aren't seeing that kind of growth, it's going to be hard to raise capital in these markets. Yeah. Is there anything you're seeing, Laura, from alphabet, from meta, right? Because your world in media and the hyperscalers sort of overlaps a lot. Is there anything you're seeing or hearing from any of these managements? That would imply that that pace of capital spending is in any way close to slowing down.

I mean, I would say yes, because we're here. hearing their capacity constraint, and Amazon explicitly ties their increases in capital with incremental revenue, because they're great businessmen. And I would say we do have media companies like Reddit that are growing 68 percent. It has been nothing to do with generative AI, but they can raise capital and their stocks going up because they're growing at a rate that is commensurate with these generative AI names. I loved your Reddit note from last week. I'm going to quote Laura Martin, back to Laura Martin, a little light airplane reading as I flew back from the West Coast, your coast, Laura,

which was you called Reddit a money-making machine. I don't, I know, I'm going to do the same thing. It's like the home alone. I don't think I've ever heard you being a very pragmatic person, use that kind of ultimate, very bullish language. 68% revenue growth. They converted 30% into free cash flow. And of the incremental revenue in the 90 days,

and in September, they converted 80% into free cash flow. I've never seen economic, A, they've had five quarters, six quarters of over 60% growth, which means whatever multiple you think you're paying, they're doubling their size every two quarters, and then they're flowing at all $80 dollars down to the free cash flow in a world where people are spending a ton of CAPX in my hyperscalers, Reddit spends $2 billion, which means everything comes to the shareholder in terms of their earnings.

Our earnings went from 16 cents last year to 80 cents, four quarters later. I mean, a money-making machine. It's amazing. Reddit's doing excellent work on the execution side. A money-making machine. I'm going to leave it there because I got nothing better than that. Laura Martin, you're always better than that. Managing Director at Needham, senior internet analyst, Laura, thank you.

All right, we got some breaking news from Washington Fed Governor Lisa Cook, who is speaking in D.C. Steve Leashman has those headlines. Steve. Thanks, Brian. The outlook for Fed policy continues to be somewhat confusing with Fed officials offering different forecasts from each other. And even some officials taking both sides of the argument. Lisa Cook, speaking in Washington, said she supported the decision to cut rates last week, but says downside risk to employment are greater than upside risk to inflation. However, she says the current restrictive policy is appropriate given inflation above the 2% target. Every meeting, she says, including a December meeting, is a live meeting. Meanwhile, San Francisco President Mary Daly said the Fed needs to bring down both inflation and support the job market. We need to continue to put downward pressure on inflation and keep our policy modestly restrictive, but not hold the reins so tight that we injure the labor market unnecessarily and give people lower inflation, but fewer jobs. Stephen Myron, the Fed Governor on loan from the administration, continues to advocate for

deep cuts while Chicago Fair President Austin Gould be saying the bar was higher for a December cut than one in October. All of this following comments on Friday where three Fed speakers spoke against the October cut and were skeptical about one in December and two, including Myron, supported both. Brian. Back to you. All right. So what's obviously no government data right now? Hopefully this shutdown will in very soon. Steve, outside of that kind of what's the next most important thing that not only you but Mary Daley, Lisa Cook and others are likely watching. I think there's the Wednesday ADP report that is kind of like the proxy new official jobs number. And I do think, Brian, that one thing that will bring Fed officials together is, or clear and obvious signs of a weakening in the job market. I think there's a big question right

now as to what is the economic momentum, what is the state of the job market. We got some high frequency data from ADP showing that last week showing the job market may have rebounded from that September swoon that we recorded with a negative number. The jobless claims numbers, we are able to put together. We get those on Friday morning and they're sort of a little bit partial. There's a couple states missing, but they have been continued to be relatively low. So it's the job market. The Fed would be addressing the inflation side of the mandate if it didn't have concerns about the job market. If that concerns go away or they become clear, the outlook for policy will become more clear.

And I think these like hundreds of billions of dollars of AI deals have to play into the Fed thinking as well, Steve. It's truly our worlds are colliding. And I love it. Steve Leesman, thank you very much. Brian, let me say one thing about that real quick, which is it's pretty important because there's also some discussion about the idea that what's going on an AI with capital investment could be pay bring over some other weakness in the economy. So there's a lot of talk about trying to look through that AI investment boom and decide what's going on underneath it. I think you read our show notes because later in the week, we're going to show some consumer

sensitive stock, Steve, that are actually collapsing. There is a lot of worry underlying that AI hood. Steve, really appreciate it. Thank you very much. All right, on deck, we're going back to AI, an exclusive interview with one of the hottest AI and energy stocks in America, the name, the news, and the trade. Next. All right, welcome back. AI, certainly the new Gold Rush crypto miners. They want in. Cipher mining, just to ink to $5.5 billion, $15-year deal with Amazon. They're going to lease out 300 megawatts of computing power all to fuel AI workloads. This is the latest in the seismic shift of computing as many crypto-related companies that have power are pivoting more to the

AI side, which demands even more power. Cipher shares, they're. They're going to be able to. It's are popping right now up 22%. They're up about 30% earlier today. Tyler Page, the CEO of Cipher Mining, and he joins us now. Tyler, welcome. Good to have you on a Power Lunch exclusive. We just heard Laura Martin of Needham say, well, she doesn't worry about Amazon because kind of like he man, the old cartoon, they have the power, I guess is the best way to say it. And so do you. Tell us about this deal that you're making today, which is getting a little bit overshadowed by that OpenAI news. Yeah, well, we're the foundation that I'm sure drives a lot of the headlines.

Look, the power they've secured is now about 300 megawatts more than they had previously. So Amazon has signed a 15-year lease with us for a 300-magawatt data center to do AI. It's going to pay us in the neighborhood of $5.5 billion of revenue over those 15 years. So we're pretty excited. We've been pivoting the company in this direction for really the last year, year and a half, and it's very exciting to finally see what we've been planning for come to fruition. Well, so tell our viewers and your investors who are watching and listening right now, Tyler, where you are in the process.

Do you have all that power or is this something where you've got to get the money and then you've got to build this out? Yeah, so look, we've been in a very privileged position because we began our lives about five years ago as Bitcoin miners. And that business is driven by finding large power interconnect. often in remote areas. And so we've developed a lot of expertise building and operating five data centers in Texas in what would be considered remote areas if you asked an incumbent data center developer.

When we have looked over the last couple years at just the growth in adoption of AI and the speed with which these large users of compute want to get to market, it's been our thesis that we should acquire and develop a very, large portfolio. So we actually, for the site in question with Amazon, we'll be delivering that next year. The site already has a fair amount of infrastructure in place. The substations

completed. It already has energy. We're currently doing Bitcoin mining there. But the whole campus will be reorganized as an AI data center ready for their full use in 2026. So the knock, if you will, what the bears might say on Cipher and other companies like yours that are sort of pivoting from the crypto side, more to the AI side, is that the model for crypto is different on AI. The scales are bigger and it's hard to replicate. How would you respond to sort of those, you know, not knocks, but sort of if the bears are going to make the case, that's probably what they would say. Yeah, it's a bad day to be a bear because what we've done over the last couple of years, is acquired a lot of talent from the traditional hyperscalor data center space that saw the industry evolving. And so our entire construction and operations team joined us from either hypers

or companies that do construction and development work for hyperscalers. We actually had to teach them how to build a lower KAPX Bitcoin mining data center. So this is a return to the familiar for them. We use the same exact providers. We manage the support. supply chain, the timeline's very similar. The difference when you build an AI data center is in a lot of the cooling infrastructure and some of the nuances of the networking, et cetera, downstream from the power. All things they've done many times. So this is a return to the familiar for us. And that's a critical point because cipher aside, I actually had dinner with somebody who builds out data centers on Saturday night. And we talked about it. And honestly, it's a really

boring business in a way because their worries are where do we get the concrete? Can we get the fiber optic cable? Can we get the labor to build the walls on the data center? Talk to us about that side of the business. Just constructing the building can be hard. Sure. I mean, these are, you know, multi-billion dollar buildouts. But I think what we've seen in the market is that there is a bottomless appetite to do project finance. if you have a long-term lease with a high-quality counterparty. You can't get much more high-quality than AWS and at our other recent deal with Google and Fluid Stack.

So we're really working with the best counterparties in the world. So access to finances there. And then when you talk about West Texas, again, we're now building our sixth data center in that area. So things like labor, supply chains, etc., in that area of the world, we have a lot more expertise than traditional data center builders. I guess I'll see you hanging out in Marfa soon. Tyler Page, CEO of Cypher Mining.

Tyler, really appreciate the view. Stock's red hot up 22%. Have a great day. Thank you. Thanks so much. All right. You're very welcome. Coming up, we're going to stay on that issue and ask this. If AI's lofty dreams might be headed for a bit of a reality check, the CEO of Microsoft said maybe you'll hear his comments on that. Next. Well, the Dow is down a little bit, but the NASDAQ is up about 410. of 1%. Welcome back, everybody. I tweeted out earlier today that, quote, there are no AI deals. There are only energy deals. It was a little bit tongue in cheek, but hopefully the message was clear that all these big AI deals that we talk about today and last week and the week before that are kind of just energy deals in disguise. And it sounds like the head of Microsoft

agrees, at least in part. In Brad Gerster's latest podcast, Microsoft CEO Saaned Indadella voiced his concern on whether the power grid itself can keep up with AI demand. The biggest issue we are now having is not a compute glut, but it's a power, and it's sort of the ability to get the builds done fast enough close to power. So if you can't do that, you may actually have a bunch of chips sitting in inventory that I can't plug in. In fact, that is my problem today, right? It's not a supply issue of chips. It's actually the fact that I don't have warm shells to plug into. Now, according to the latest note from Morgan Stanley's strategy, Stephen Bird,

we are likely about 49 gigawatts or so of power short to make all these AI dreams happen. Now, gigawatts can be confusing. Very, very, very roughly. One gigawatt is about 750,000 American homes worth of power. So we're kind of talking about 40 million homes equivalent worth of electricity. That would be enough to basically power three times, all the homes in Texas.

So how do we hit that supply and what companies may it ultimately benefit? Stephen Bird joining us now. He has a must read. You've made me sound like I actually know what I'm talking about, Stephen, so I appreciate you and your entire team. Did we lay it out okay? I mean, how much power are we really talking about being short right now? Yeah, it is a huge amount. It is hard to think about gigawatts, but you have the math exactly right.

What I'm encouraged about is in that issue, new opportunities are coming to the four. So, for example, some really creative ways to provide that power much more quickly than the traditional grid approach. So, for example, natural gas turbines, fuel cells would be a big part of the solution, siding at nuclear power plants. And then recently, Bitcoin conversions where you take all that power that these Bitcoin sites, use and repurpose those as hosting HBC data centers, the economic arbitrage. It's really exciting on everything I just mentioned. And the value to the AI community to providing power more quickly is quite massive when you put on the lens of a power person, the numbers we're talking about in terms of that arbitrage

is. It's really quite exciting. It's what we just talked about was Seif. Right? I don't know if you were able to hear that interview. And I'm just so angry with myself because I talk about energy for a living in part, Stephen, 10 years ago, I wasn't smart enough to say, I'm going to go buy a really old power plant, because everybody would have said, well, you can take it.

Nobody wants that power plant. It's old. We're going to phase it out. These people bought these power plants for Bitcoin mining, largely. Bitcoin mining takes a huge amount of power, but it's a pimple on the back of AI. And so now they're sitting on these power sources where it's like suddenly you had something nobody wanted. And now 10 years, everybody wants it.

And you can sort of kind of name your own price. That's exactly right. I mean, frankly, no one saw this coming. But we started to write about this last April. April of 2024. We described how these Bitcoin companies could repurpose their sites as data center host. And at the time, none of these deals had occurred. At that time, those stocks are trading at a metric that no one in Bitcoin land cared

about at the time. They really should care about it, which is the enterprise value divided by the watts of power they had. At the time, most of those stocks traded between 50 cents a watt and $2 a watt. And now the companies in that world of Bitcoin that have converted tend to trade at $7, $8, $9 a watt. What I'm really excited about is there are several Bitcoin companies that have not yet converted, but are signaling very clearly that they're in discussions to convert. And they trade between, call it $1 and $3 a watt.

And I just see so much opportunity. Can you tell us? Absolutely. Yeah, the one that's. that is the cheapest would be Hut. Hut 8? Yep. Absolutely. By the way, I have these Power City indexes where I track city returns, whatever, you know, whatever. Number one stock in Miami is HUD 8. Yep. For good reason, it has such great opportunity. The company's been very good at acquiring additional power. They have a lot of power of expertise within the company, a lot of data center

expertise, and they're clearly engaged in discussions to try to convert the very smart management team. They get it. Riots is another one. that trades between $3 and $4, ride signal they're very much not just engaged, but they're really laying the foundation now to convert their sites. And they have two of the best sites in the United States for conversions.

So there are a handful that are essentially going to catch up to the companies that have converted, and the arbitrage really is quite exciting. It really is. And a good question was asked in my ear, one of my team members, Joe, who said, well, what does this mean for Bitcoin then?

If everybody's pivoting from Bitcoin to AI, is this the death knell of crypto mining? Not crypto, obviously, but crypto mining? I think not in the sense that what will happen is not everybody is going to convert every site, first of all. But what we will see is essentially what they would call the hash rate or the degree of competition to mine what one Bitcoin can go down as these sites are repurposed. That would mean actually greater margins for Bitcoin mining. So it does provide a sort of unintended positive side effect for everybody in the world of Bitcoin mining around the world, just because essentially the amount of competition to mine that limited number of Bitcoin will go down.

It's really unbelievable. Is there any part of you? I'm going to ask you the exact same question I asked Laura. Martin at the top that says, this is great, but we're decades. We're not years away. We're decades away. Well, so I am worried that we just won't have enough power, as Attina-Del. I see these creative solutions. But for most of these solutions, there's some degree of execution risk, though I would say Bitcoin conversion has lower execution risks, but we see all kinds of local challenges to getting data centers built, opposition to power price impacts, concern about water, a variety of local concern. So I see lots of execution issues. When it comes to the question of whether we need all of this power, I think the answer is clearly yes,

in the sense that we build our power model from the chip level up. And the way we think that's good because it avoids double counting. if you were to just follow all the data center projects, you might come up with some giant number for data centers, but a lot of those will never get done. We think grounding your math in the actual chips that are going to be sold is a good way to make sure you're not overstating things. And even then we're coming up with a big power problem.

I'd also note every three months when we update our power number, our chip numbers always go up because the demand for compute continues to go up, and there's been a lot written and assessed about the complexity of AI, which requires so much more compute. more compute. So I'm fundamentally bullish. We're going to need this power and bullish on the arbitrage. It's a spectacularly interesting time to be alive and you don't realize you're in history until you end up making it later on. We're in history right now for power and it's fascinating.

Stephen Bird, thanks for making me smarter. Really appreciate it. Thank you. All right. Up next, the tech heavy NASDAQ, on the rise to kick off November. Yeah, the market Dow may be down, but NASDAQ's not. It's up four-tenths of 1%. But with the shutdown in DC, nearing a record tomorrow, could D.C. kind of hit any Santa Claus rally. We'll ask David Zervos and Ali McCartney. Next. We got a news alert on NVIDIA. Let's go to Aman Jabbas in D.C. with that, Amen, what do you got? Yeah, Brian, we've got some new reporting here from the Wall Street Journal just within the past half hour or so the Wall Street Journal reporting on some of the context behind President Trump's meeting with Xi Jinping in Asia last week. The journal reporting,

this according to them, that the president wanted to talk to his top aides just before the meeting with Xi Jinping about the idea of allowing exports of top invidia chips to China. And the president was met with some opposition to even discussing that with the Chinese side. The opposition, according to the journal, coming from Secretary of State Marco Rubio in particular, but also U.S. trade representative Jameson Greer and also Howard Lutnik at Commerce, all across, according to the journal, standing opposed to this push by NVIDIA to allow exports of the top chips or at least second ranking chips to the Chinese side on national security grounds. We have a statement here now from the White House.

They're giving us the same statement that they gave to the Wall Street Journal, which is in their story. They say President Trump listens to a variety of insights on policy matters, including from top business leaders. President Trump's historic meeting with President Xi proves, however, that the only factor guiding his decision making is the best interest of the American people, that according to a White House spokesman. So Brian, what you have here is some emerging details of the context around what we all noticed at that meeting, which is that the president didn't offer a concession on new sales of sophisticated chips to the Chinese side. Now the Wall Street Journal explaining that

part of the reason for that is this sort of united front of opposition from the Secretary of State and Secretary of Commerce and the U.S. Trade Representative. So kind of the idea, basically, that though Jensen Wong is being very, very nice to Trump. In fact, using some of Trump's own language last week, right now it doesn't appear to be helping? It doesn't appear to have solved the problem from Jensen Wong's perspective, right? I mean, Jensen Wong has been embarked on a very public lobbying campaign. He's gone to a lot of Trump events. He's, as you say, using Trump language. He's showed up at the White House quite a bit, is in contact with President Trump quite a bit from what we understand. Clearly, what he wants is to be able to sell as many chips as he can into

that Chinese market, which is enormous and super valuable for NVIDIA. The question from a U.S. national security perspective is, is that the right thing for the country or not? And the president appears to have decided, at least at this meeting, that he wasn't going to discuss it with Xi Jinping. And so that's a dead letter for now. But it doesn't feel to me, Brian, this is just an instinctive thing. It doesn't feel to me like this is over, though, at this point. I don't know if anything's over. Amon Javers. Really appreciate the breaking news on Nvidia, NVIDIA.

By the way, still up. Amon, thank you very much. Right? Let's talk markets and your money and macro. Here's kind of an RBI you may not hear elsewhere. More stocks actually rose in October than declined, and it's day 34 of the government shutdown. Does this mean that the shutdown or partial shutdown

is not impacting the equity markets? At least it is not yet. Let's talk about more in the Fed and the markets and your money with David Zervos, Jeffries. He's chief strategist and Allie McCartney managing director of U.S. Private Wealth Management, UBS. Ali, start with you because you promised when you sat down, you said, Brian, I've got some historical context and stats. Would you bring? Okay, here we go.

So you just had a guest on there was talking about the tenuousness of CAPEX and the investment that has to happen in this transformational catalyst. If we look back at everything from the steam engine to telecom to electrification of the grid, for an extended period of time, so over years, not over months, somewhere between two and five percent of GDP was spent on infrastructure to keep those things going. Right now, we're at about a year, maybe two, of slightly less than one percent being spent. So when you think about the context of where Capix is, is a bubble, you know, all of that sort of relevant information, I think if we believe, and we're talking about this in the break, that we are making history and transforming and catalyzing something very new, scary and very new, that there is more to go, both in terms of KEPX and upside. This is, I love this view because a week and a half ago, I was on Route 80, right here.

Took Route 80, which is right here by CNBC. I flew across the country on an aeroplane. I landed and I got back on Route 80 outside of San Francisco. And I thought somebody somewhere made the decision, we're going to build high. across this country. We're going to spend a ton of money building asphalt into farmland. Is that kind of, David, what we're doing right now? Up in the sky, I think it is what we're doing. I mean, we're creating, you know, the network of the future. We're creating a technological advance that seems to be

moving forward, I think even differently than what happened in the 1990s. I think Jay Powell last week said a lot of things I didn't agree with, but one thing he did say that I agreed with was that this is different than the 90s. These companies are making money, Brian. They have significant revenues. The concept is proven. And now it's a matter of scale and investment. And I think we are going to see, we have a very high probability of seeing a transformational AI trade, transformational tech trade that has with it a lot of creative destruction. And I think that's going to be the real issue for the policy. I agree. And I was at the NASDAQ in the late 90s, a lot of those companies were garbage. And we found that out.

That's not my opinion. We found that out. The stocks went to zero. They realized there was no, there was no man behind the curtain cranking the wheel and making the thing. You referenced the Federal Reserve earlier. You might have heard me sort of quickly asked Steve Leasman, how does AI and all this money factor into Fed's thinking? It's got to, we influence them somehow. You would hope it would, but I don't think it is. I think with the new Fed folks that will be coming in next year, it will probably play more of a factor, Brian. You've got more supply-side thinking rather than traditional Keynesian demand-side thinking. And that should have folks looking at it more like a greenspan would have looked at it in the 90s, where he pushed back on the staff and he said, you know what,

yes, growth is strong, but we don't need to hike because growth is strong because these are supply-side driven. He also said a grational exuberance in 1996. The market briefly wobbled, and it turns out it was like the wrongest thing ever, ever said. I mean, now, now, Now, he was four years later, he was right. But in Wall Street, that's wrong. If you're four months early, you're maybe four years early. You're totally wrong. But a lot of stocks keep going down, Allie.

Like a lot of the consumer stocks are lower. It just feels like everything is about the same 10 stocks. I like that you brought that up because I think it feeds off of what David was saying. This is a, this is a K-shaped economy when we talk about that. Right. So the fact that we've had governments since the financial crisis dumping money and leaning into risk and the rich have been getting richer and richer. And we have a technology AI that is likely going to disempower lower level employees, right? So you're hearing a lot about now.

That's scary. That's scary. It's very scary. Rich are getting richer. Middle class workers are more and more afraid of their jobs or losing their jobs. And younger people feel not hopeless, but that. they can't participate in the American dream? The unemployment rate in young people right now is very high.

And so I think that the Fed not only has to think about the technological advancement and the quantity of money that's going into infrastructure build, but they really have to start to think about the labor market and unemployment differently because there are so many structural things in the labor market that are changing, whether it's due to immigration, AI, aging population. Can we guys, they call this producing from the set. I call it being Peyton Manning. have Rick Santelli next. I was wondering, can we roll Rick Santelli into this conversation right now? Because if we can, I think it's absolutely perfect for Rick. Rick, are you there? Can you

hear me? Rick, are you out there? I'm here. He's there. Fantastic. I'm here. Because what Ali just said and what David was just talking about, I think it's critical to what you talk about every day. AI is amazing. But if AI can't, doesn't matter in bringing down mortgages from six and a half to 7% and people can't afford to buy homes. The bond market just does what it wants because worried about inflation related to the spending. I don't know if it's a net positive for the economy at all. What do you think? I think that we're just going to have to wait this out. I think that the mortgage rates are moving towards adjustable, which makes me a bit nervous because an adjustable rate mortgage might be about 70 basis points lower than an average 30-year

fix. But ultimately, it's a bet on the future. And would if rates stay the same or go higher, there won't be much of a benefit. So I think you're exactly right. A good chunk of the economy is predicated on the cottage industries and directly correlated to the health of the housing market. And if interest rates stay here, it's going to take time before a five and three quarters to six and a half range of a 30-year fixed mortgage makes sense to people, just like about 40 years ago when interest rates started to move down and started to move back.

up. So being rate locked with rates under three and a half percent, I don't think that dynamic's going to change. But over time, even people with a very competitive, significantly lower rate will find reasons over time to be moving and changing residences. And that dynamic could take years. Was that wrong? David Zervos, about that? AI bringing down mortgage rates. All people care about is, can I buy a house? Mortgage rates and energy, and energy, your favorite. I mean, I think gas prices and mortgage rates are the two things this administration's going to focus on going into the midterms. And, you know, that's a difficult task getting these longer rates down.

But I think Secretary Besson has made it a goal of his to get that done. I think he's looking at a variety of ways of doing that outside of the Federal Reserve that will likely make that happen. If AI can bring down mortgage rates and we've perfectly, and energy, we've perfectly combined all the big three stories. Allie and David, nobody does. It's better than you. Thank you all very much. Appreciate that. Let's get to Courtney Reagan for a CBC News Update. Hi, Ryan. More than 20 Democrat-led states are challenging a new Trump administration's student loan forgiveness policy.

It blocks non-profit and government workers who work with immigrants and transgender youth from being eligible for the public service loan forgiveness program. The states argue the administration is overstepping its authority, and they're asking a judge to declare the policy illegal. A federal probe into Tesla's electronic retractable door handles is expanding. Regulators gave the electors. electric vehicle company until December 10th to present extensive records around the design, failures and resolution of customer issues with its handles. It comes after Bloomberg recently reported several incidents in which the handles stopped working and trapped customers inside their vehicles. An Alphabet-owned Waymo announced today it will bring its robotaxi service to three new cities, San Diego, Las Vegas, and Detroit. Waymo is already in five cities. The company says

it plans to launch in San Diego and Las Vegas next year and in Detroit this winter. Brian? Back over to you. I was in Waymo's all last week, and I have to say it felt like the future. Oh, that freaks me out. I'm not ready. No, it's cool. Do it. You won't want to go back.

We got to say goodbye because up next we're going to talk about how poop is helping remove carbon for the atmosphere. Poop. Stick around. All right, welcome back. Waste management is a dirty business, not just on the ground, but in the atmosphere as well, that garbage releasing carbon as it degrades. But what if it didn't happen? What if we could collect, you know, human waste and not have it go into the atmosphere?

Diana Oleg explains how we can do that in her continuing series on climate startups. You know, Diane, I just had to say poop. I had to say it. I knew you did, Brian. I would expect that. But anyway, when organic waste decomposes in a landfill, it emits carbon dioxide and methane, two of the biggest climate offenders. Now, each year, landfills produce 120 million tons of CO2 equal to the emissions of 26 million

passenger cars. So why not get all of that waste out of the landfill and put it deep underground? Food and agricultural waste, excess manures, human waste, and paper sludge. These are the materials that a Houston-Texas-based startup called Vaulted Deep deals in. Building on expertise from oil and gas waste engineering, it converts all this waste into so-called slurry and injects it directly into the earth. It's sort of a win-win. Both a win for the climate. because we're helping lock away carbon for hundreds of thousands of years. We're avoiding methane emissions. We're also contributing to cleaner local environments and protecting community health.

Vaulted adapts oil and gas deep well technology to build new wells and repurposes old salt caverns like this one in Kansas. The slurry is injected up to 5,500 feet underground more than three times the height of the Empire State Building. Once underground, the slurry naturally mineralizes and stabilizes, ensuring per seattley. permanent containment. Vaulted Deep not only contracts with waste management firms, but sells credits into the voluntary carbon markets. It's all voluntary commitments from these large companies saying, you know, hey, we care about the climate. We have ambitious climate targets, whether that's net zero or carbon neutral. And so they are voluntarily buying our carbon removal. Large companies like Google, which is a client. We think that vaulted has an approach that because of its benefits to

local communities and because if its impact not only on CO2 but also likely on methane is poised to deliver that scale of impact over the long run. Vaulted Deep is backed by Prelude Ventures, Lower Carbon Capital, Earthshot Ventures, Woveen Earth Ventures, Rethinth Ventures, Rethink Impact and Fall Line Capital. Total funding, $40.3 million. $1. Valted Deep CEO says the material is stored in deep geological formations capped by thick, impermeable shale, the same type of rock layers that have safely contained hydrocarbons for millions of years, Brian? Does Google just buying carbon credits, or are they actually doing something?

Well, they're actually investing to purchase 50,000 tons of CO2 removal through the system with vaulted deep. But in this collaboration, they're also studying how to account for methane emissions, which is much harder because that's the bigger offender than CO2. They really want to get rid of methane. So they're working with Vaulted Deep to understand how to measure and get rid of the methane emissions as well. Diana Olock, are the continuing climate-related series, Diana, thank you very much. All right, coming up, a big Palantir preview. That is next. All right, welcome back. So let's wrap up Power Lunch with really three quick big stories. The first one is Kenview. That is the maker of Tyler.

and other products being bought by Kimberly Clark. It's a nearly $49 billion deal. The combined company would really bring together brands like Huggies and Kleenex with the lives of Band-Aid and Tylenol would generate an estimated annual net revenue of about $32 billion. That deal expected to close in the second half of the year. It ends a very rough ride or theoretically ends it for Kenview. Also tonight, Palantir.

Their earnings out tonight, but right now a new intraday record high. shares up not only today, but more than 170% this year. And our friend Jay Woods of Freedom Capital pointing out that Palantir could see another 20% gain post earnings, noting that Palantir has popped an average of 20% in eight of the last times it is reported earnings. The market tends to love Palantir earnings. Then we're going to wrap it up with energy, something we've talked a lot about today, particularly Diamondback Energy. That is the original Fang stock. The ticker is actually Fang. Diamondback is an oil and gas driller. It is beaten estimates in three of the past four quarters. The stock is slightly lower today. It's not about 15% of the year as oil prices

continue to come down, but we'll see what changes when they release their numbers and give guidance after the bell. Well, that's it for us. We'll see it tomorrow. Closing bell starts right now. Thank you.