Back in the late 1990s, I was on the corporate bond syndicate desk at Citi (NYSE: C) when a telecom company called WorldCom issued over $60 billion in bonds.
At the time, it was the largest issuance for a non-sovereign issuer.
This aggressive borrowing was justified by the argument that bandwidth demand was exploding and internet traffic was doubling. The thinking was that if you built the biggest infrastructure, the economics would sort itself out later.
Because scale would fix everything.
At the time, that felt like airtight logic. And for a while, it actually worked.
WorldCom used its debt to buy competitors, lay fiber and lock in capacity ahead of demand. And Wall Street loved it. Institutional investors happily scooped up these bonds and the stock kept climbing higher.
With all this capital rushing into the market, it laid the groundwork for the dotcom bubble.
But the deal also locked WorldCom into a future it could only survive if everything went right.
That’s an uncomfortable parallel to what I’m seeing in today’s market.
Different Technology, Familiar Behavior
What most people probably remember about WorldCom is the accounting fraud that grabbed headlines. The company hid rising costs by booking everyday expenses as long-term investments, making the business look far more profitable than it really was.
But fraud only bought the company time. What made WorldCom truly vulnerable was leverage.
The $60 billion of outstanding debt wasn’t just about financing growth. It marked the point when the business changed from earning its way forward to borrowing its way forward.
That distinction is easy to ignore when growth is strong. But once momentum dies down, leverage stops amplifying success and starts narrowing the margin for error.
In other words, fraud explained how WorldCom fell apart. But debt explains why it couldn’t survive once momentum turned.
By the time WorldCom collapsed in 2002, it had more than $40 billion in debt on its balance sheet and filed for what was then the largest bankruptcy in U.S. history.
I’m not bringing this up to tell a war story. I’m bringing it up because I’m starting to see the same behavior happening again.
Of course, there’s a different technology driving the story today. And both the companies and the numbers involved are much bigger.
But the same instinct exists.
I see companies making rational decisions based on optimistic assumptions. And they’re financing those assumptions with borrowed money.
Today’s stock market looks healthy at first glance. Indexes are near record highs and volatility is low.
But as we’ve talked about before, it’s a barbell-shaped market.
A small group of companies tied to artificial intelligence now accounts for the vast majority of its performance. By some estimates, around 80% of U.S. equity gains over the past year came from AI-related stocks.
And seven companies now represent more than a third of the S&P 500’s total value.
That concentration has happened because capital has decided that AI is the only future that matters.
I don’t dispute the importance of AI. And I’m on record that I believe the U.S. must win the race to artificial superintelligence (ASI).
But I am concerned that the market isn’t pricing AI rationally.
Because what’s driving this cycle isn’t just earnings. It’s spending. And a growing share of that spending is being financed with debt.
Today’s tech giants are in the middle of the largest infrastructure buildout since the telecom boom. As you can see from this chart, annual AI spending by some of these companies rivals the GDP of entire countries.
Source: 9fin.com
Amazon plans to spend more than $100 billion a year on AI-related data centers and cloud infrastructure. Meta is also anticipating hundreds of billions in cumulative capital expenditures over the next several years to support AI workloads. Microsoft continues to expand its data center footprint at a pace that would have been unthinkable a decade ago.
Some of that spending is funded by cash flow. But a growing share of it isn’t.
Because these same companies have been issuing bonds at an accelerating pace to finance AI infrastructure. JPMorgan estimates that AI-linked investment-grade debt tied to AI could reach $1.5 trillion by 2030.
On its own, that isn’t a problem.
Like I said before, debt doesn’t matter much when growth is accelerating. It only starts to matter when expectations change.
Right now, companies are building compute ahead of demand on the assumption that AI is guaranteed to scale.
It’s a bet that makes sense on paper. Spend now, get big fast and assume that profits will follow.
This logic should also sound familiar to anyone who was around in the late 1990s. Back then, telecom companies laid fiber ahead of demand because traffic was “guaranteed” to arrive.
Here’s the thing…
Telecom demand didn’t collapse. It simply arrived more slowly than capital markets expected. Yet it was enough to wipe out equity holders and leave creditors fighting over scraps.
That same risk exists today.
Many companies already use generative AI. But a large percentage of them report little to no measurable impact on profits so far. Productivity gains are real, but they’re uneven. And monetization is still being figured out.
Meanwhile, capital spending continues as if the outcome is settled.
That could lead to a very dangerous outcome.
Here’s My Take
Today’s market is built on the extraordinary technological promise of AI.
I believe AI will change the world. However, I’m also cognizant that adding leverage to the system adds more risk to investors.
If AI delivers at the scale and speed investors expect, the companies leading this buildout will justify every dollar they’ve borrowed and spent. And whoever reaches artificial superintelligence first will enjoy a competitive advantage unlike anything we’ve seen before.
But if returns arrive slowly, balance sheets will eventually start to matter again.
And the difference between companies earning their way forward and borrowing their way forward will become impossible to ignore.
Regards,
Ian King
Chief Strategist, Banyan Hill Publishing
Editor’s Note: We’d love to hear from you!
If you want to share your thoughts or suggestions about the Daily Disruptor, or if there are any specific topics you’d like us to cover, just send an email to dailydisruptor@banyanhill.com.
Don’t worry, we won’t reveal your full name in the event we publish a response. So feel free to comment away!
