You know the late '80s, early '90s downturn where it made a lot of sense to deploy capital. The same thing after the financial crisis. The next few years will be similar, but most investors are still licking their wounds from the past. Their real estate's underperformed. Their committees want to look in a different direction. And that's when the time of maximum opportunity tends to exist, but there tends to be minimal interest. And so, it's in a moment like this where back to this sort of stay calm, a little bit of a contrarian view that we have at Blackstone, you try to take advantage of a moment like this.
So, I love you mentioning office. I had our mutual friend Peter Linneman on the Walker webcast back in January. By the way, did you take Linneman's class at Wharton?
I didn't, but I know Peter. He's terrific.
I love that you built the largest commercial real estate owner on the face of the planet and you didn't take the one real estate course at Penn. That's fantastic. So I had Linneman on and I said to Peter, 'Okay, so if you had to pick one asset class to invest in right now and obviously it's all location specific and basis specific and everything else, but just give me a broad brush. Where would you go?' And Peter said to me, 'Stay rich: multifamily.' And so then my question to him was, 'Okay, get rich.' And he goes, 'Office.' And then I said, 'Get poor.' And he said, 'Data centers.' So with that as the preamble to data centers, Blackstone has put a lot of capital and a lot of focus on the data center space. You've put 50 billion of equity into buying data centers and investing directly in data centers. I think you have another 50 billion teed up in 2025 to continue investing. I've heard you talk about investing in the picks and the shovels, back to the sort of the gold rush analogy of you might not want to buy the gold mine, but you might want to be Levi Strauss and supply the dungarees to the people who are working in the mines. But when you and I last saw each other at the JP Morgan asset management conference, Michael Semble got up and talked about AI and he talked about data centers and by JP Morgan's calculation, the hyperscalers have put enough capex into data centers to basically have 12,000 ChatGPTs run on them. And whether we get 12,000 ChatGPTs or two ChatGPTs, what he's trying to basically say is we may be headed towards oversupply in the data center space. I guess the direct question, John, I'd have for you is, given the backdrop and the amount of capital that is going into data centers, what gives Blackstone the conviction that we're not going to see a bubble here in data centers as it relates to the amount of capital and infrastructure that's going into the data center space?
It's absolutely the right question. I'd start with what we do in data centers, which is we build only when we have long-term leases with some of the biggest companies in the world. So, if you contrast that with building condominiums in Miami or Dubai when the price is high, and then lots of people come in and you get this variability and so forth and overbuilding, or what happened back after the Civil War with railroads. In this case, we don't go out and build a billion-dollar data center speculatively. What we do is we get a long-term lease from a major company and then we invest the capital and then those leases are call it 15 plus years. So to me the backbook of what we own here is very safe, very stable. The question then is, will we find ourselves soon where a bunch of the hyperscalers will say they're full up, the DeepSeek dynamic, that the cost of compute has come down, that we don't need as much training? And what I would say is we continue to believe that the data centers are the physical manifestation of the migration of our lives online, data storage, what's going to come from autonomous cars, the migration to the cloud, obviously AI, which we could talk about, which I think is going to have a very profound effect on all sorts of businesses. And even if there's less creation of new large language models and training them because the cost of compute is coming down, there's going to be more and more usage. And so based on what we're seeing with our tenants, what you've seen with the biggest hyperscalers, their demand continues not only in the U.S. but in Europe and Asia and we will continue to deploy capital when they come to us and say, 'Hey, I want to build a data center.' We may have an option on the land, but the vast majority of the capital only goes in the ground when we have a lease. And my gut today is the world's going to need more compute power and therefore the world is going to need more data centers. If it turns out it slows a lot, then our development pace will slow. And then adjacent to this, what's really interesting is the power that goes with it. Power demand in the U.S. has basically been flat for 40 years. And now I think we're going to see because of this digitalization, because of AI, ultimately because of robotics, I think you're going to see a very big increase in demand for power, which impacts transmission and generation and utility services. It's another way, picks and shovels, to play this mega trend. So if you said as a firm, what are we trying to do in investing capital, which is think about some of the big things that are happening out there like we did with single family homes or the movement to logistics online, now with data centers and what we're doing with energy and power, find those great neighborhoods and then put a bunch of chips against them and do it in private equity and real estate and credit and infrastructure. That for us, that thematic approach has really paid dividends in terms of delivering for customers.
So, two other things, quick things on AI. First one is, do you use one of the AI services, and if so, are you willing to put an endorsement of who Jon Gray uses for his AI chatbot?
Well, I mean, I think everybody, we at the firm have our own white-labeled AI. Here on my phone, I think I've got OpenAI is what I'm using on my phone. But I do that for personal use. And I would say to me that's like the early stage. I was in Europe last week with one of our CEOs who's now using AI. They give incentives online to customers and they used to use what they thought was advanced technology. They're now using one of the large language models and it's allowing them to reduce their customer acquisition costs by two-thirds. Massive increase in margins. One of my partners here took the Blackstone earnings call transcript and a couple of equity analysts' reports and put it in the Google large language model and it created two virtual bots having a conversation for 20 minutes about the pluses and minuses of Blackstone stock. Now, it wasn't a 10 out of 10, but maybe it was a seven and a half. And so when you think about customer engagement, you think about content creation, you think about coding, we're going to see a huge movement. And I think this is going to have a very large impact. So that's really what underpins my continued belief that there'll be demand for data centers.
Really funny that you talk about the use of AI. We had a Goldman Sachs CNBC report that circulated inside of Walker & Dunlop yesterday to give a sense of where the market was. And one of my colleagues, Aaron Appel, took it and ran it through a chatbot to just have it spell out all the different acronyms that were in it that allowed someone to actually read it as if it was real English and you didn't have to be an expert on CNBC like you are to be able to understand it. And it makes me think about, I went to the 2023 annual report of Blackstone, John, and in the first page there were 14 different acronyms of various groups inside of Blackstone. And I was just curious, it was seven years ago that you took over as president and COO and moved from just a focus on real estate to the broader focus on the firm. And as it relates to the data, you now sit on top of, I believe it's 230 operating companies and 13,000 individual real estate assets. That's obviously everything from a big office tower, and not many of those, to multifamily businesses to logistics centers to single family homes. Given your view over all of that, how are you using data to manage such a vast organization? Because as I ask you that question about AI, I'm certain you have certain data feeds that come to you that tell you something about how PE is doing, how your insurance and credit business is doing, how your commercial real estate business is doing, and how your hedge fund business are doing, which are your big sort of four big food groups. But if I were to sit there and look at the data you look at, what are you looking to to help drive Blackstone forward?
So what we're trying to do is capture these data feeds on what's happening in some of the key areas. Obviously revenue volumes, inflation, what's happening with wages, what's happening with input costs, and the bigger businesses that have broad reach. So, some of our infrastructure businesses, we own the largest ports business in North America or we own the Rome airport. We own all sorts of roadway services businesses, all sorts of things that track GDP. You love to see those inputs. The real estate, as we've talked about, when you own tons of apartments and warehouses and hotels, you get a lot of insights. We have some very large operating businesses in private equity that are servicing the healthcare industry or they may be automotive parts related. And what you're trying to do is pull this all in and also do it on a global basis and then on a real-time basis share it and try to do it on a more automated basis. But to be fair, you need human interaction with this because sometimes someone sold a division or there's a one-time factor. So somebody has to look at this and we've really dedicated a team of people here now whose job it is to assemble this information because what we want is a competitive advantage. If you think about investing as connecting dots, seeing more, we should have more dots than anybody. We should real-time be able to see is the economy strengthening or weakening, are we seeing inflation going up, going down, where are we seeing pockets of strength or weakness and therefore can we deploy capital. So we're trying to automate a bunch of this. There is a human element of it. Interestingly, we send out to the senior partners the board summaries of all the boards of companies to the people on the private side of the business. So you can look at and see what's happening across major businesses. And we have the AI sometimes do these summaries which can be helpful if you're trying to consume all this information. And then as human beings, yesterday I had my operating committee, all the people running the different business lines at Blackstone, talking about what they're seeing real-time. So I'd say some of it's a little bit old-fashioned communication, our regular Monday meetings, our management operating committee, but a bunch of it is moving more automated. And I think as a firm like ours, we should make scale one of our great competitive advantages. And information at scale should be our best advantage. But I'd say it's still a work in progress.
When I looked at your four major groups, if you will, of PE, which is obviously the founding of the firm and it's what Steve set out to build at the beginning, then you've got commercial real estate, then you've got your insurance and credit business, and then you've got your hedge fund business. I was interested, John, that in those three big food groups of PE, commercial real estate, and credit, that the three of them now have about the same AUM. They're all 350 billion, give or take, as it relates to AUM. I was surprised at that, but I also have heard you talk about that we are in a golden age of private credit. Talk for a moment about Blackstone's growth in the private credit space. And the other question I have for you there is one related to the composition of your investor base there and individual investors versus institutional investors. But for a moment talk about the private credit space and the great growth you as well as some of your competitor firms have seen in that space over the past couple of years.
So on the first part of the question, I don't think there was a plan for all of them to get to roughly the same size, but it's due to Steve Schwarzman to push us to diversify. So you could have a business when you're in a higher rate environment, higher spreads, you have a credit business that does really well. As you know, rates fall and multiples expand, you have equity-oriented businesses. Having that balance, I think, is really important. I would say there's been a cyclical element to the growth in private credit and then there's a secular element. The cyclical would be rates went up a bunch, spreads widened, the returns you got for being a senior lender in corporate land, in asset-backed finance, in real estate were very high. Equity-like rates of return for debt-like risk. Some of that is going away as spreads come back down and as base rates come down. The part that's not going away is us bringing up our investors. And you talk about the composition of investors. It's certainly institutions, our pension funds, it's insurance companies, and now individual investors, and bringing them right up to borrowers and them trading away a little bit of liquidity. It's not a public bond, but they've made essentially a private loan, but capturing in the case of investment-grade debt, 175 basis points of extra spread. That is incredibly valuable. And so what we're seeing with our clients is a lot of interest to earn that extra return. I'm taking the same A-minus risk that I was taking before, but I'm earning a spread that may be more than double than what I was earning. And that is what's causing this sea change movement into private credit. And if you think about insurance companies, what's interesting is if you're an annuity provider who writes five-, seven-year annuities or you're a life provider who's got 30-, 40-year liabilities, if you can trade away a little bit of near-term liquidity for those higher returns, that makes a ton of sense. And it's also helpful to the financial system. If you think about First Republic failing, they had a $70 billion super-prime mortgage book. They basically had no defaults, but they had 20-year mortgages and 20-second deposits. And so what we're really doing is sort of match-funding the assets and liabilities. And we're also doing it in this farm-to-table mode that gives higher returns. And I'd say we're doing it a little differently than some of our competitors. They've chosen to become insurance companies and take on the balance sheet risk. We're doing it no different than we've done private equity or real estate or anything. We're a third-party investment manager where what we're trying to do is deliver our customers higher returns. So, we're sticking with that model in credit and we're doing it on an open-architecture basis with a whole range of clients. And today, obviously, there's a lot of momentum in that space. If you include real estate debt plus the corporate credit area, it's now up to over $450 billion. So, that combined is actually the biggest thing we do at Blackstone on an AUM basis. And when I look at what I'm hearing from clients today, I would say that business has a high likelihood of growing a lot.
So when I hear you say farm-to-table, I sit there and go, I thought that's what banks do.
Well, it's interesting. Banks do some of that. But they also, as you know, distribute in asset-backed finance. They distribute in direct lending through the broadly syndicated loan market. They distribute in real estate and CMBX. They'll take fees and distribute. So some of that they're in the distribution business. And then the other thing, banks I think today are trying to drive their returns on equity higher. So it may make sense for banks. You've seen us and others announce a bunch of partnerships where they say, 'I've got all these consumer loans on my book. Maybe what I should do is retain 20%, bring in somebody like Blackstone to hold 80% of the debt. I'll get origination fees and I'll get servicing fees.' And so you're seeing more and more partnerships between us and the banks. But this evolution of private credit is actually driving down the cost for consumers and businesses and I think making the financial system stronger. And of course banks are going to continue to be a huge part of the whole package.
And when I think about the consumer, you clearly don't take deposits, but I believe that of your 1.1 trillion of AUM, about 230 billion of it is from private investors. And so insurance companies would be the 230 billion.
Oh, you're talking about 230 from insurance, 260 billion from individual investors.
