“There are two times in a man’s life when he should not speculate: when he can’t afford it, and when he can.” Mark Twain
Artificial intelligence is destined to revolutionise our society, yet the off-balance sheet funding into AI ventures could be inflating a bubble.
This situation reflects a classic self-deception that humanity succumbs to during every economic boom.
We convince ourselves we can hide risk by moving it somewhere else. But risk is like gravity. You can shift the weight around, disguise it or rename it, but it never goes away. It only gets harder to notice.
That’s the core of what’s happening in AI right now.
The biggest companies in the world are racing to build data centres that cost as much as small nations.
They don’t want those numbers on their balance sheets because big numbers scare investors. So they rent instead of owning. They use special-purpose entities (SPV).
They let private-credit funds borrow on their behalf. On paper, the company looks lighter. In reality, the risk is simply moved a few feet to the left.
It’s the financial equivalent of shoving clutter into a closet before guests arrive.
Look at Meta. A $27–30 billion data centre that Meta doesn’t “own,” financed by a partner who technically does, except Meta guarantees the value for sixteen years.
Tech and AI firms are using Special Purpose Vehicles (SPVs) to shift $120 billion in data centre capital expenditures off their balance sheets, according to the Financial Times.
The debt doesn’t show up where you’d expect, but the obligation is still theirs. The liability still exists. It’s just wearing a different coat.
And this isn’t one company. It’s the emerging trend in the whole industry.
In the old days, if you wanted to build a factory, you borrowed money, built it, and the debt sat on your books.
Today, Big Tech has found a loophole called the Special Purpose Vehicle (SPV).
The pattern is simple: AI demand is exploding. The cost of building AI is even bigger. Balance sheets can’t absorb it, so the costs are pushed off-balance-sheet.
Meanwhile, the assets backing this boom, GPUs, age like fruit. A chip that costs millions today can be obsolete in twelve months.
Yet companies borrow against them for five years. That works as long as everyone believes tomorrow will be better than today. It stops working the moment growth slows, even a little.
You can build AI capacity in under two years. Power grids take a decade to upgrade. Capital is racing ahead of infrastructure that cannot keep up, and the wall only becomes visible when you hit it.
And underneath all this, a circular economy emerges. Company A funds Company B, which buys from Company C, which invests back in Company A. Everyone reports growth. But no new money enters the system.
It’s like neighbours selling each other cars they can’t afford, then taking out loans against the cars to keep the game going.
It feels prosperous, right up until someone tries to sell to an actual outsider and discovers the market isn’t real.
If this sounds familiar, it should. Every bubble is different in content but identical in structure: optimism → leverage → complexity → surprise.
The surprise is never the thing itself; it’s the realisation we were looking at it the whole time and chose not to see it.
The feeling this moment should create is not panic, just humility. A recognition that we have built an AI ecosystem on layers of financing that depend on everything going right, forever. That’s never happened in history.
The risk isn’t that Big Tech collapses tomorrow. They won’t. The strongest firms are profitable, diversified, and adaptive. Some will turn AI into durable earnings.
The real vulnerability sits around them: the funds, vehicles, partners, and lenders whose balance sheets can’t absorb a pause, a delay, or a disappointment.
Every great technology creates greater societal value. Most companies chasing it don’t capture that value.
AI will change everything, but most investors won’t win.
We’ve seen this before.
Cars reshaped the world. Two thousand automakers tried. A few survived.
Radio and aviation promised a new era in the 1920s. Investors rushed in. The stories were right. The timing wasn’t. After the crash, almost all the money was gone.
The mistake is confusing a breakthrough with a business. When uncertainty feels exciting instead of dangerous, capital piles in. That’s when losses are made, not avoided.
Investors want more, hate missing out, follow the crowd, and believe there’s always a sure thing. These habits show up most at market extremes—and they usually end badly.
The real question isn’t whether AI wins.
We believe that companies like Alphabet, a fund holding, are successfully integrating AI deeply within their existing ecosystem. This strategic positioning will better enable them to monetise their AI in the future.
You don’t need a conspiracy theory to see this. Just a willingness to look at the obligations companies don’t want you to look at.
Stop focusing on what companies report. Start asking what they’ve promised, such as achieving Artificial General Intelligence (AGI) in the near future or at least turning into profitability.
We need to start paying attention beyond what they’ve promised.
- Look at the leases
- The guarantees
- The long-dated commitments are backed by fast-ageing assets
- The returns depend on everything going right
- The round-trip revenue
- The power bottlenecks, and
- The financing that sits just outside the balance sheet.
As Warren Buffett once said, “The less prudence with which others conduct their affairs, the greater the prudence with which we should conduct our own affairs.”
Financial engineering can hide risk, but it cannot destroy it. It just moves it from the balance sheet to the footnotes.
If you understand the debt we pretend not to see, you understand the future better than the people creating it.
The AI boom is an undeniably revolutionary technology, comparable to the advent of railroads and the internet.
In fact, this technology is demonstrating one of the fastest rates of adaptability ever observed. But the argument for the financial boom relies on the premise that circular revenue is real and that depreciating assets can support long-term debt. Historically, that is rarely true.
Awet Ghirmay, Head of Investment Research
If you want to learn more about this topic, please find the link here.
More information about our fund can be found here: Hargreaves Lansdown or AJ Bell.
Risk warnings
The value of investments and the income derived from them may fall as well as rise. You may not get back what you invest. This communication is for general information only and is not intended to be individual advice. You are recommended to seek competent professional advice before taking any action. All statements concerning the tax treatment of products and their benefits are based on our understanding of current tax law and HM Revenue and Customs practice. Levels and bases of tax relief are subject to change. This blog is based on my own observations and opinions.
