Larry Fink Explains Why Hyperscaler Stocks Will Go Higher

In today's video, I want to show you some highlights from a recent interview that I found with Larry Frink, who is the CEO of Black Rockck, one of the largest asset management companies in the entire world. He is extremely well connected into what is going on with AI, the data center buildout, and where artificial intelligence is going. And in this interview, he provides a lot of insights that I think are extremely relevant to the market today, especially with the sell-off that we saw with the AI sector and AI relevant stocks last week. So, what I'm going to do is play you a few clips from this interview and then we will discuss the highlights from each clip afterwards. So, with that being said, let's dive right into it and get started with our first clip. >> You know, you mentioned 50 years, so I am curious. I mean, have you ever seen anything quite like this, Larry? We've got what were the most profitable companies the world has ever seen making a choice to incinerate, wrong word, to decide not to have any free cash flow anymore, to not return capital to shareholders because they have to be part of this headlong race to make sure they are not left behind in the AI future. When I listen to myself say that, are you there's some worrisome aspects to that, are there not? I mean, I've spoken to some of the uh leaders this week and last week in in that sector. Um their biggest worry right now, supply is not keeping up with demand as you and you see that showing up in the value of of the memory stocks. Okay, we have more demand for memory than we have supply. Obviously, they've been able to do big price increases. I don't know how sustainable that is. Maybe in three four years we'll have enough supply but you know um and so what we see as a big investor in data centers the demand for compute is not slowing down it's growing faster the problem we have as a country we're not investing fast enough so I have an opposite >> so you're we're just not even we need to do even more >> I mean a trillion dollar capex spent from six companies not enough. I I believe it's well not just there. We're not investing in our grids fast enough. We do not have enough adequate supply of power and that's what's creating some of the issues at certain state levels that they're worried about electricity prices for the consumer which is a legitimate issue. But other states that have more power and more capability of delivering power which and and delivering power that is not going to raise the price to the consumer are going to be the big winners for growth. Yeah. So in the last quarter we you know we contracted close to 1 gawatt of power in Pennsylvania. Okay. We're financing another big data center in in another state. And in my conversation with every hyperscaler, demand is exceeding supply. And yes, you're right. These large companies used to be balance sheet light. And now their business has changed. And that is we but that's the role of the capital markets. And I promise you, these are going to be great investments for individuals if we could if >> but you've got to know that the underlying business model ultimately that we get to is going to generate enough >> enough revenue and profit to be able to pay for all the data centers. They're not cheap. >> Yeah, they're 50 60 billion dollars for 1 gawatt. >> So I took some notes on what I believe are the highlights from this first clip here. And the first key point is that Larry has spoken to the leaders of the hyperscaler companies and supply is still not keeping up with demand. Demand for compute is also growing faster and is still accelerating. Larry also believes that we still need to do more and invest more capital especially investing into energy. And every hyperscaler CEO that he has talked to said that demand is still exceeding supply. And I think that this is interesting because the hyperscalers and the AI trade in the stock market on Friday sold off dramatically. And this was mostly because of the new Chinese model that came out. And we're going to get into that. But what Larry Frink is saying here is that every single one of these companies is still capacity constrained and the demand for compute is still higher than supply and it's actually accelerating. And this gives me confidence in my Amazon and Meta positions, especially heading into earning season because Larry Frink is the CEO of the largest asset management company in the world. He is extremely well-connected and he is very clearly saying that this tailwind is still there. It's still massive and these companies are still seeing unprecedented demand. So heading into earning season, I think that Amazon, Google, Meta, Microsoft are all going to have fantastic earnings reports and maybe dispel some of the fear that we are continuing to see in the market right now because it seems like the AI trade is getting a little bit sold off and maybe a little bit fearful, which I don't necessarily agree with because I think these companies are going to see tremendous amounts of demand for years to come. Now, getting back to my list of key points here, Larry Frink then said that these businesses are no longer capital light, but they are still great investments. And I agree with him. And I also find the argument that hyperscalers are investing in more physical assets, and that's a bad thing to be kind of silly. And think about it, AWS was seen as the crown jewel business for Amazon for over the past decade. AWS has always been an extremely capitalintensive business that requires a lot of upfront investment. And then you see significant amounts of cash flows over the longer term. Now again over the past 10 years, this was seen as a fantastic business model that still had very high margins and everyone wanted to buy Amazon just for the AWS exposure. But now that AWS is seeing even more demand than they can capitalize on, they literally cannot throw enough money to even meet the demand they're seeing. Now investors are skeptical of this business model. Don't you think that Amazon has a very long track record of proving they know how to operate this business profitably over the longer term? And now that the business is seeing more demand than they can even capitalize on, shouldn't that be viewed as a good thing? Even though it's a capital inensive business and always has been, that's kind of how I think about this and the capex spend that they're doing. But for some reason, the market is still not showing faith in the capex and is now questioning AWS's business model as a whole. So, I do think the bare case here doesn't fully make sense. And again, just because a business is capital intensive doesn't mean it's a bad business. It can still be a very good one, produce a lot of cash flow, and also produce high returns on invested capital. So, now let's move on to the second clip from Larry Frink. I worry about can all a society benefit from AI because right now compute is so expensive. My questions to everybody in the hyperscaler business is not whether it's not whether they have enough demand is how quickly can they bring down the cost of compute. I'm not worried about Black Rockck paying for um the the AI and I think you know we have been a huge investor in it and our systematic equity team has really benefited with the flows and the return but I worry about the small and medium businesses. How are they going to be able to compete in this new AI technology world? So the biggest question I ask behind the scenes is how quickly can we bring down the cost of compute and and okay they you know they tell me it's like a Moor's law and how quickly they could develop faster and faster analytics to to have faster compute but unless we develop better technology and better systems in our grid to provide more power more consistent power you know I believe the United States must be power agnostic. We should we should not care if it come from solar or hydrocarbons. We have to be there. But we don't have, you know, we need to be manufacturing the solar panels in the United States building the battery storage. At least we can do the battery storage here. >> We do. >> So those are my worries. My worry about is not a bubble. My worry about we're not we don't have the ability to build fast enough. And then I go watch China. China's building 100 gawatts of nuclear. They're building close to 100 gawatts of solar. Okay, they are getting set up for this AI revolution and the need for power. We're we're not doing this enough. And I actually get frightened when I see states saying we're going to do a moratorum. That's not the answer. The answer is how do we get how do we deliver more power quickly? Let's why don't we why don't we all start because again how do we deliver more power so we don't raise electricity prices but we can be the center of the AI revolution. So I took more notes from the second clip and the highlights that I think are important. So the first highlight is that Larry Frink believes the cost of AI needs to continue coming down and he actually wants the cost of AI to come down which I have been seeing people say is a bearish thing for the overall trade and it's actually largely what is causing the panic in the stock market today. Now, in my opinion though, I believe that as the cost of AI inevitably comes down, it will allow more and more people to use AI and the return on investment of AI will actually increase. I mean, think about it. If the ROI of AI increases for more companies, then more companies will adopt it. So, the cost of AI going down actually feeds into the bullcase and increases the overall usage of artificial intelligence. In my opinion, this will directly benefit the hyperscalers in the infrastructure layer of AI even more. It's maybe not bullish for the large language models, but for the ones powering artificial intelligence and using it internally to make their businesses better, it is absolutely bullish. That is also why I chose to invest in Amazon and Meta because Amazon is an infrastructure layer that benefits from the increased adoption of artificial intelligence and Meta is using it inhouse to make its recommendation systems better. So as AI becomes cheaper, they should be getting a higher return on investment by using it. Now the second main point that Larry Frink brought up is that to get AI cost down to the point where we can increase usage and get that ROI better for your everyday person and your smaller businesses, we need to build a lot more energy. And he says that one of his fears is that China is building so much more energy than the United States and North America. So, they could actually have a competitive advantage because their energy is going to be cheaper, which will allow them to see a better ROI and develop artificial intelligence even more. And I actually have some screenshots here that I want to show you. And this first one shows us that China is building power much faster than the United States is. Just in the last 4 years, China added more power capacity than the entire existing capacity in the United States. Think about that. China basically built the entire US power capacity just in the past four years and it now has about triple the United States total power capacity. This next screenshot shows what the projections are like if this trend continues. And we can also see that the AI demand for power is projected to continue growing and that by 2030 it's projected that that AI's total power demand will meet the current total energy capacity of the United States. This means that if the United States does not start bringing on a ton of power and pretty quickly, then energy will quite literally stop artificial intelligence progress because it will use up all of it. And again, this is not an issue in China and could give China a huge competitive advantage. Now, in my opinion, this could actually start to become a political issue if the United States does not start bringing on power and it becomes more and more clear that China has this huge competitive advantage over the United States in artificial intelligence. I mean, think about it. If it starts to become clear that China has this huge edge and lead over the United States, then the president is going to start to push for energy development and start cutting red tape all over the country to start building as much power as possible. And I think that based on the trends that we're seeing, this is probably going to happen. And it's also partly why I am invested in Brookfield Asset Management, as I have shared previously on my channel because they are one of the asset managers that is specifically investing in the energy buildout and infrastructure buildout in the United States, in North America, and also over in Europe. So, as energy scarcity continues to become a larger political issue, I think that there's going to be a further push to invest in energy infrastructure in North America specifically because what the trend is showing us is that we need a lot more power. We need it to continue progressing artificial intelligence, continue pushing the frontier of AI and also simply to keep our energy cost down and make artificial intelligence affordable. If we don't, then China is very much going to get the lead and a lot of the AI spend and benefits and the return on investment is going to go over to China. So I think that it's a necessity for us to build a lot more energy infrastructure in the US and in North America and I think that is going to be the reality of the situation. I agree with Larry Frink here. So now let's move on and watch the third and final clip from Larry Frink which I think is an important one as well. I'm very bullish on the markets over the next 12 months. I think the tech technological revolution is going to power uh better margins for more companies. I mean, think about Black Rockck. We've raised our margins increased by 260 base points over the last 12 months. A lot of it is using more and more technology. >> Can you keep those margins going? You said you don't see 45 to 46% margins as a ceiling. You pointed back to 2021. Your business though is very different than it was in 21. So, I'm sort of curious. What's the mechanism that gets you back above 47%. >> Uh continuing driving uh more growth in our private markets area which we we see incredible momentum. Uh continue to drive opportunities in retirement which I believe these are all going to be creating better margin and just making sure that we're utilizing technology as fast as we can to do more with less. I mean, our headcount, David, is unchanged and our our, you know, we're up a trillion dollars in assets. Okay, that's that's what you're you're seeing. We're able to use technology to process more trades to process more uh activities. So, we're able to do that. We're able to leverage our human capital using technology, working alongside of that. we're using um the amount of uh code we're writing alongside our coders with AI is accelerated dramatically. So we're able to do more and more and more and I and as the ability of AI and compute as I talked about earlier becomes faster and cheaper that will drive even higher margins for not just for Black Rockck for other firms. All right. So, Larry Frink finishes this interview by saying that he is very bullish on the markets over the next year because BlackRock is internally seeing the benefits of leveraging artificial intelligence. And as artificial intelligence continues to become cheaper and better, more and more companies will see the same benefits. He's not worried about Black Rockck being able to afford artificial intelligence because they have pretty much unlimited access to capital. But right now, AI is still pretty expensive, especially for small and medium-sized businesses. So, as AI continues to get cheaper and cheaper, then he believes that more businesses will be able to benefit and leverage artificial intelligence to make their businesses better, expand their margins, and ultimately increase their cash flows. And as more businesses are able to adopt AI and see the positive ROI on it, he thinks that that's going to increase the overall earnings of the market as a whole and a lot more businesses. So, I actually went over to Black Rockck's earnings and in their most recent quarter, we can see that their operating margin hit 45.9%. Which is the highest margin the company has produced over the past few years. Earnings over at Black Rockck are also growing very quickly and hit an all-time high of $13.91 per share in the most recent quarter. So, the margins and profits over at BlackRock are continuing to expand. And Larry said that this is because they're not having to hire more people, but they can do a lot more work with their existing workforce. He said that the assets under management of the business have grown by over $1 trillion without their headcount expanding. So AI is allowing their employees to get a lot more done. And that is where a lot of this margin expansion is coming from. Now, another example that I have here is from Marcato Libre's most recent investor letter that they put out in June of 2026. So I have a couple of screenshots and in the first one it says Marcato Libre runs the largest technology operation in Latin America. 40,000 plus applications a multicloud infrastructure spanning millions of compute instances hundreds of thousands of data services 20,000 plus developers across the region billions of lines of code changed every year and thousands of deployments of code every day. At the scale, any improvement in productivity has a huge impact. With generative AI, we're not seeing peacemeal improvements. We're seeing transformational gains in productivity. Then they continue on to say, "The clearest demonstration of this is autonomous work that agents are doing. As a result, the team is delivering far more than a year ago. Code submissions are up 90% year-to- date versus a year ago. Merged code, which has accepted releases, are up 80%. and deployments are up nearly 60%. In the second quarter alone, these figures more than doubled. The number of unique code bases that each developer is working with has doubled as they work across a broader range of tasks. Rollbacks, which are code that is returned to its developer due to errors, fell by an entire percentage point year-over-year as the quality of rollouts has improved. More than 30% of managerial roles are now contributing to code again freed up by agents handling the review burden. So Marcato Libé is also saying that artificial intelligence is materially impacting the business and allowing them to do a lot more with a lot less. The overall productivity of its employees and the business has increased tremendously. This is also why Larry Frink believes in the AI buildout again because his business and many others are seeing tremendous benefits and productivity gains. But currently a lot of these productivity gains are only for the huge companies that can afford to really invest in artificial intelligence. So as AI continues to become cheaper then these productivity gains should be more widespread and benefit more companies. As more companies see the benefits the adoption grows and the hyperscalers will see more demand. And this is the most important point in the interview right now relative to the markets in my opinion because last week the market sold off the AI stocks because a Chinese model came out that was cheaper than the leading frontier models in the United States. Basically the market freaked out because AI got cheaper. But I actually think that this is what we want to happen. We actually want AI to become cheaper because as it does the adoption of AI will increase. And again, as I said earlier, I believe that the infrastructure layers providing AI are going to see significant tailwinds over the longer term as AI continues to become cheaper. That's mainly Amazon, Google, and Microsoft, the cloud providers for artificial intelligence. So, I actually agree with Larry Frink and I am a bull on this overall market because I think that the tailwinds from AI are massive and I think that the buildout is still in its early stages based on what I have heard from the hyperscaler CEOs, the asset managers and the people actually leading this revolution. We're still in the early stages and the buildout could last well into the 2030s. We need everything. We need more data centers. We need more compute. We need more energy. And I think that this demand and this investment cycle is going to continue well into the 2030s as I just said. So ultimately I don't think that the dips that we're seeing in the market when people start freaking out about the ROI on AI capex spend is necessarily justified. And I think all you have to do is take a look at Meta's financials. Take a look at Microsoft's financials. These companies are investing a significant amount of money. Yes, but their revenue growths are all accelerating. Their operating cash flows are all accelerating and at all-time highs. Amazon's gross margin is at an all-time high. Its operating cash flow margin is at an all-time high. So, my overall takeaway is that Larry Frink is still very bullish on this market, he's very bullish on artificial intelligence. And based on what he has heard from the hyperscaler CEOs, AI and compute demand is only accelerating. And since this demand is so high, I think it's going to continue propelling these companies earnings and cash flows forward for years to come. So every single dip we see, I'm going to be a buyer and I think that this upcoming earning season is going to be quite bullish for the hyperscalers. But with all that being said, that's going to wrap up today's video. And if you enjoyed this video, then as always, please remember to leave a like on it. And if you want to see more content like this, then please consider subscribing to my channel. Thank you so much for tuning in. I truly do appreciate it. And I hope to see you again in my next