Let’s be honest, you can’t open a browser or turn on the news without hearing about AI. It’s the talk of the town, the engine of the future, and the source of a massive, multi-billion dollar investment boom. And at the heart of it all are the "hyperscalers"—you know, the handful of tech giants that basically run the cloud and are building the massive infrastructure AI needs to function.
It’s an exciting time. But as someone who has spent years in the insurance world, my brain is wired to look for the "what if." What's the risk hiding in plain sight?
Well, it turns out there’s a big one, and it’s a bit of a financial magic trick. Some of the world's biggest companies are funding this AI gold rush using a method that keeps massive amounts of debt off their main books. Officials are starting to call it "shadow borrowing," and for insurers and private credit funds, it’s a risk we seriously need to talk about.
So, What Exactly Is This "Shadow Borrowing"?
It sounds a little sneaky, right? And in a way, it is.
Think of it like this: Imagine your friend wants to buy a flashy, expensive sports car. Instead of taking out a huge car loan that would show up on their credit report and make other lenders nervous, they find a creative workaround. Maybe they have a rich uncle set up a separate company that buys the car, and your friend just makes "lease" payments to that company. On paper, your friend's personal debt looks low. But in reality, they're still on the hook for that very expensive car.
That’s a simplified version of what’s happening here. These tech hyperscalers are using what the finance world calls “off-balance sheet arrangements.” Instead of taking on traditional debt that shows up clearly on their financial statements, they’re using other structures to finance their AI build-out. This could involve things like complex lease agreements for data centers or setting up special entities just to hold the debt for new server farms.
The result? The company gets the cash it needs to build, but the full extent of its financial obligations isn't immediately obvious to everyone. And according to folks at the Bank for International Settlements (BIS), this practice is becoming more and more common.
Why This Should Matter to You, Especially in Insurance
Okay, so a few tech giants are being clever with their accounting. Why should we, in the insurance and private credit space, lose any sleep over it?
Because we’re the ones lending them the money.
Insurers and private credit funds are major players in financing these kinds of large-scale corporate projects. We buy the bonds, we fund the loans, we provide the capital that makes this AI revolution possible. And our entire business model is built on one thing: accurately assessing risk.
Here's the problem: if a company's true debt is partially hidden, how can we accurately price the risk of lending to them? It’s like trying to decide whether to give someone a mortgage without seeing that they have three other secret credit cards maxed out. You're making a decision based on incomplete information.
This creates a couple of pretty significant dangers for our industry:
-
Underestimated Risk: We might be looking at a hyperscaler’s balance sheet and thinking, "Wow, they look incredibly stable with very little debt." Based on that, we offer them favorable lending terms. But if we don't see the billions tied up in these off-balance sheet deals, we've fundamentally miscalculated the risk.
-
Concentration Risk on Steroids: Let’s face it, the world of AI infrastructure is dominated by just a few massive companies. A huge amount of capital from the private credit world—including from insurers—is flowing to the same handful of players. If one of these giants were to stumble, the ripple effect would be enormous. The fact that their debt structures are becoming less transparent just adds another layer of uncertainty to an already concentrated bet.
It's All About the Data Centers
So, where is all this money going? Mostly, it’s being poured into the physical backbone of AI: data centers.
AI models, especially the large language models everyone is talking about, are incredibly power-hungry. They require thousands upon thousands of specialized computer chips housed in massive, climate-controlled buildings. Building and equipping these data centers costs a fortune.
To fund this, a hyperscaler might sell one of its existing data centers to a group of investors (which often includes private credit funds and insurers) and then immediately lease it back. This is called a "sale-and-leaseback" arrangement.
The tech company gets a huge infusion of cash to invest in new AI chips, and the investors get a steady stream of lease payments from a high-quality tenant. It seems like a win-win. But that lease obligation is a form of debt, and it doesn't always show up in the same way a traditional loan would. When you’re talking about deals worth billions of dollars, that’s a pretty big detail to have tucked away in the footnotes.
Are We Walking Into a Familiar Trap?
This isn't the first time the financial world has seen problems with off-balance sheet financing. If you were around during the 2008 financial crisis, you might remember the chaos caused by "Special Purpose Vehicles" (SPVs) that held risky mortgage debt off the big banks' books.
Now, I'm not saying this is the same thing or that we're headed for another 2008. The underlying assets—data centers for blue-chip tech companies—are very different from subprime mortgages.
But the principle is similar. Whenever complexity and a lack of transparency creep into finance, risk has a way of hiding in the shadows. And when the market eventually shines a light into those shadows, it can lead to some nasty surprises.
For us as insurers and investors, the key takeaway isn't to panic or run away from AI. That would be like trying to ignore the invention of the internet. The key is to be smarter and more demanding. It means our risk analysis and due diligence have to go a level deeper. We can't just take the main balance sheet at face value anymore. We have to ask the tough questions about lease obligations, special financing arrangements, and the true, all-in debt picture of the companies we're funding.
The AI boom is real, and it’s reshaping our world. But as the ones writing the checks, it's our job to make sure we understand exactly what we're paying for—and all the risks that come with it.
