Since the launch of ChatGPT three years ago, AI-related stocks are responsible for generating 75 per cent of S&P 500 returns, 80 per cent of earnings growth and 90 per cent of capital spending growth, according to a J.P. Morgan paperDado Ruvic/Reuters
Here’s a glimpse into the near future. The tech sector, drunk on the world-changing potential of artificial intelligence, wildly overinvests in building out the technology, creating a kind of financial black hole that eventually collapses in on itself.
The bubble will burst, the stock market will tank – making Friday’s 3.56-per-cent decline in the Nasdaq Composite Index look like foreshadowing – but the technology now being manically scaled up will remain.
This is not some contrarian prediction pulled out of thin air. It’s just what tends to happen when you take a revolutionary technology and apply infinite funding with zero accountability for long enough.
First comes the boom and then the bust. Then, in the aftermath, do we see what the technology can truly do.
It’s how the dot-com bubble went down. The telecom industry spent way too much money laying fibre-optic cable that was mostly unused for years to come. The bubble burst, and investors big and small lost fortunes. But the groundwork was laid for the internet age, which really did change the world.
The burning question for investors today: When does the AI boom reach its tipping point? That’s arguably the murkiest part of all of this.
There are credible voices on either side of the debate.
On one hand, financial media is full of scary headlines about absurd stock valuations and a dangerous concentration of Big Tech in the U.S. stock market. How much longer can this go on?
OpenAI eyes artificial intelligence data-centre capacity in Canada
Torsten Slok, the widely followed chief economist at private equity giant Apollo Global Management Inc. APO-A-N, recently wrote that the “AI bubble today is bigger than the IT bubble in the 1990s.”
That kind of talk can easily unnerve the average Canadian with money in the stock market. If a weatherman tells you a hurricane is coming, shouldn’t you try to get out of the way?
The problem is, there are lots of different ways to value stocks. Many market strategists insist stocks are not as frothy as they were in the 1990s, by most measures. Pricey, yes. But not unprecedented.
They also say that most of the run-up in the stock market has been supported by earnings growth, which wasn’t the case during the dot-com boom.
Plus, the AI infrastructure build-out is largely supported by cash flows rather than debt. Morgan Stanley estimated that around half of the US$3-trillion expected to be spent on data centres over the next three years will be financed by Big Tech’s internal cash flows.
Opinion: Artificial intelligence: the good, the bad and the ugly environmental costs
Surveying the research of some of the top minds in the business, one gets the sense this has much further to go before any imminent reckoning, despite how crazy speculative the financial markets seem right now.
It brings to mind one of the most famous quotes in investing: “The market can stay irrational longer than you can stay solvent,” British economist John Maynard Keynes said.
A great example of this is the dot-com bubble itself, the excesses of which were evident long before the stock market crashed.
In a speech in December, 1996, then-Federal Reserve Chairman Alan Greenspan warned about the “irrational exuberance” he saw taking hold among investors.
He was right, but it took another three-plus years for the stock market to peak. In that time, the S&P 500 index doubled. The Nasdaq Composite Index quadrupled.
If the peak excesses of the late 1990s are not yet upon us, you can feel the same forces building.
The AI infrastructure spend is the single most important driver in both the global economy and financial markets today. Consider but a few of the signposts on the road to AI dominance:
- Since ChatGPT launched three years ago, AI-related stocks are responsible for generating 75 per cent of S&P 500 returns, 80 per cent of earnings growth and 90 per cent of capital spending growth, according to a J.P. Morgan paper.
- These same stocks accounted for almost all the wealth built by the world’s richest people last year, according to data compiled by Bloomberg.
- Spending on data centres in the U.S. is on the verge of eclipsing spending on the construction of office buildings.
- Data centre-related spending accounted for around half of U.S. GDP growth in the first half of this year, according to tech investor and author Paul Kedrosky.
Spending on data centres soon to overtake
U.S. office building construction
In billions of U.S. dollars, seasonally adjusted annual rate
the globe and mail, Source: j.p. morgan; Census Bureau
Spending on data centres soon to overtake
U.S. office building construction
In billions of U.S. dollars, seasonally adjusted annual rate
A vast overspend is almost baked in at this point. A couple of weeks ago, Meta Platforms Inc. META-Q CEO Mark Zuckerberg downplayed the risk of misspending a “couple of hundred billion dollars” on AI.
“I actually think the risk is higher on the other side – if you build too slowly and then super intelligence is possible in three years,” he said.
The AI hyperscalers can’t shovel money out the door fast enough. Nvidia Corp. NVDA-Q recently announced it would invest as much as US$100-billion in OpenAI, the developer of ChatGPT.
OpenAI, meanwhile, said it would spend US$300-billion on computing power from Oracle Corp. ORCL-N
But Oracle is one of the biggest customers of Nvidia’s AI chips. There’s a circular effect taking hold, with giant sums of money being recycled among the industry’s biggest actors.
The same sort of thing happened through the dot-com bubble. It’s undeniably dangerous.
And yet, you don’t want to be the type of investor who bails on the stock market in 1996. Nor do you want to be the type that goes all in in 1999.
Tickers mentioned in this story
Study and track financial data on any traded entity: click to open the full quote page. Data updated as of 10/10/25 4:00pm EDT.
Meta Platforms Inc
Nvidia Corp
Oracle Corp
Apollo Global Management Inc
Follow related authors and topics
Authors and topics you follow will be added to your personal news feed in Following.