Delay And Pray - The New Business Plan In Multiple Asset Classes
However, underneath that topical canvas the economy has distinct and dangerous problems, and one of those problems is the huge mark-up in assets. Many assets in the US economy have prices assigned to them that are not real, meaning if those assets were to transact today (with a buyer and a seller) those inflated prices would never hold on the buy side of the transaction. Basically, there are fake valuations all over the place in the US economy.
There are four main areas where these inflated (unrealistic) prices are being used to keep the 'run on the bank' from taking place. There four industries are commercial real estate, private equity, private credit, and the AI industry.
First, commercial real estate is known to be in a huge downdraft. CRE was struggling before Covid but in the aftermath there is an obvious gap in the marketplace. There is simply no need for many of the office buildings in cities and their suburban areas. Throw into that problem the overbuild in upscale apartments, a big buildout in warehouse and industrial space, and too many retail spaces.
All aspects of commercial real estate are in some state of struggle. Banks have been slowly but surely revaluing (meaning they are taking losses) every quarter. Another way to view the problem is that over $1.5T in commercial real estate is facing debt refinancing over the next 15 months. And lenders have already been operating an extend and pretend strategy, where they just pretend that the buildings have the same valuation they did back in 2019 - even though many of the properties are worth less by a factor of 30% if not 60% or more. At some point the lenders will have to put a true value on these buildings and when they do there will be massive losses by lenders and the owners of the properties will be wiped out.
Second, private equity went on a tear from 2013 through 2025, no one knows how many total companies they purchased, and some of those acquisitions were highly profitable. But on the downside, PE has some 32,000 companies that they are not able to sell. Those same companies are normally managing a large amount of debt, plus they have slow sales, high costs and business models that are not very stable.
For instance, many PE firms invested in niche software for specific industries and many of these companies have very limited business and limited upside. At the end of the day they are largely worthless. Another big operation of PE has been the build and build rollup, where PE will buy up a lot of businesses in a specific industry which can range from dental to car washes, HVAC repair to veterinarian firms. These rollups were very expensive to assemble and they're not able to grow with a slowing consumer.
The valuation question in this instance could be very inflated as they use a methodology known as Marked To Model, instead of a live pricing environment called Marked To Market. Basically the PE firms are putting a valuation on their companies that they dictate, they use either the cost they paid for the firms or some other internal valuation, but either way this is putting a value on the companies that is nowhere near reality. The value of many of these companies is zero or less than zero, however know one knows because they are privately held. It's a tremendous amount of money that is overvalued, easily eclipsing $1T and most likely way beyond.
The dollars at risk for these 32,000 PE-owned firms is not known, but many estimates show PE employing over 13M Americans. That's a lot of households in the US depending on the continuing function of private equity, so it's not just the PE-ownership and lenders that are at risk for these inflated valuations.
Third, the close cousin of PE which is private credit. In the modern era private credit has largely replaced bank lending in multiple spaces, including residential housing finance, construction lending, AI buildout, and funding private equity operations. Private credit is projected to be over $2T by now and in this instance it's the firms and assets that they loaned into that are at risk.
Many analysts are starting to get nervous about the lending into rapid AI expansion, as there is no obvious profitable revenue streams yet established by any AI operator. The dangers of residential and commercial real estate declining is a real threat to private credit lenders. In normal circumstances commercial lending (banks, credit unions) are conservative in their lending, they vet their customers before making loans. But in the case of private lending those standards appear to be far lower, there does not appear to be any rigorous vetting going on with their customer base.
Private credit has issued over $2T in loans and the amount of these loans that are at risk is probably very high. The valuations of the businesses have risk in getting marked down, and the revenues behind those firms are at risk as well. There is a high probability that many of their loans will get a substantial haircut and many others will just be declared worthless through bankruptcy.
Finally, there is the AI sector of the US economy. Every participant in this space has a valuation that should be considered overvalued. There are the owners of the data centers, the chip manufacturers, the neoclouds which are resellers of data processing/inference, and the AI operators themselves. They all are at risk, and their current valuations reflect an overly optimistic projection on their business growth and profit margins.
The data center owners have extremely high costs in terms of buildout, maintenance, and utilities such as water and electricity. They also have pushback from local communities. Trying to earn back even the initial investment in these enormous data centers will be a challenge, much less making their ongoing operations profitable.
The chip manufacturers like NVIDIA, Intel, Micron, Broadcom, AMD and others have massive market valuations on the prospect their chips will be in constant demand going forward. But this could easily be a faulty assumption if the AI operators are unable to find their way to profitability. At some point, as the losses continue year after year, the ability to continue to buy advanced microchips at very high prices will begin to slow and it might slow down rapidly. That would make the current super high valuations of these stocks (and the stock market writ large) unsustainable.
The neoclouds may carry the most risk, these are the resellers of cloud computing. Companies in this space like Coreweave, Iren and Nebius are counting on the AI industry to continue on with huge overpaying of services and massive issuances of new debt. If the AI business begins to slow then the value of the resold computing space could easily fall off a cliff. These firms all carry the risk of bankruptcy if AI slows even by a little.
Then there are the AI operators we all know like Grok (SpaceX), Copilot (Microsoft), Claude (Anthropic), ChatGPT (OpenAI), Gemini (Alphabet), Muse (Meta) and others. These are the hyperscalers with the deepest pockets but also the biggest debts. While many of them are protected by other business models they have all made very substantial investments (bets) into the AI space.
The belief is that whoever wins the battle for AI supremacy will be the most valuable company in the history of mankind. But what might be happening in real life is something much similar to the Dotcom market crash of 2001, where many of the early internet companies failed and went bankrupt.
The AI business is a trap for all of the overvalued companies at once. The AI firms themselves, assuming that companies, governments and individuals plan to collectively spend trillions of dollars on AI in the next few years might be greatly surprised by how little money those entities are actually willing to spend.
This will put them all into jeopardy, but along the way it will put commercial real estate investments into danger, as the data centers will lose value, in some cases may even lose the need for their existence. Private equity investments into AI-related service companies could make them irrelevent and not able to pay their debts. And private credit has a massive exposure to the AI trade, in all sorts of physical buildout, loans to speculative AI firms, startups, companies that sell cloud compute, chipmakers, all along the AI pathway is a variety of danger. Without the creation of many paying customers for AI services the private credit investments will move closer to some degree of loss.
All told the United States economy has trillions of dollars of investments that are marked higher than they should be. How bad is the situation? Could there be $12T in valuations that in reality are closer to $3T in reality? No one knows. But this is not conjecture, the overvaluations are real and the clock is ticking for when the moment comes when the actual value, the real value where another party is willing to pay in real dollars, and that day may not be too far away.