The AI Spending Boom Is Starting to Show Cracks
The Weekend Edition is pulled from the daily Stansberry Digest.
You couldn't write the story any worse...
A Wall Street Journal report on Tuesday painted a sour picture for OpenAI, the creator of ChatGPT.
It noted that OpenAI "recently missed its own targets for new users and revenue." And insiders say the company's chief financial officer is concerned that it has overcommitted to spending on data centers...
Chief Financial Officer Sarah Friar has told other company leaders that she is worried the company might not be able to pay for future computing contracts if revenue doesn't grow fast enough, according to people familiar with the matter.
Board directors have also more closely examined the company's data-center deals in recent months and questioned Chief Executive Sam Altman's efforts to secure even more computing power despite the business slowdown, the people said.
Altman and Friar didn't deny those facts in a statement to the Journal, but said, "We are totally aligned on buying as much compute as we can."
Every so often, a report about a single company – and in this case, a private one – can significantly move the market. This one punished shares of a handful of businesses with ties to OpenAI...
Oracle (ORCL) and Broadcom (AVGO) quickly fell about 4% on the day, continuing the recent pullback. Advanced Micro Devices (AMD) shares fell around 3%, and Nvidia (NVDA) lost more than 1%.
Real estate and infrastructure firms with data-center exposure fell, too.
We've been warning about the risks of this sort of "circular" scenario over the past year...
In short, companies like Nvidia and AMD have invested in OpenAI to be the firm's primary chip suppliers. In return, OpenAI has essentially promised future spending (some $600 billion) for these and other companies to develop the data centers and infrastructure needed to grow AI's footprint.
This dynamic has fueled buzz about OpenAI's business – without the revenue to justify it. ChatGPT is no doubt a useful, breakthrough product, but is OpenAI's business profitable? Not so much. Yet, it has played a large part in the rising expectations for all AI businesses over the past few years.
We wrote about OpenAI and its web of deals in a February Stansberry Digest issue. And we shared the take from our Stansberry's Investment Advisory editor Whitney Tilson...
These circular relationships can keep a company like OpenAI going for a long time – but can collapse quickly. And I think we've reached a tipping point.
This story could turn out to be nothing more than a bump in the road. But if things are as bad at OpenAI as this latest report suggests, and if the company pulls out the rug on most or all of its spending promises, look out for the companies that have been known as the "AI leaders" in the past few years.
Hype Isn't Carrying AI Stocks Anymore
The Magnificent Seven results are coming in...
On Wednesday, four of the Magnificent Seven stocks released their quarterly earnings reports.
The four companies – Amazon (AMZN), Alphabet (GOOGL), Meta Platforms (META), and Microsoft (MSFT) – have a combined $12 trillion market cap and make up 20% of the S&P 500 Index.
On the surface, all four companies are thriving... And based on the market's reaction, you might think all is well. All of the major U.S. stock indexes recovered from the brief sell-off on Tuesday and closed higher by Thursday.
But the Mag Seven story is more mixed when you start to dig into the details of these companies.
That's why on Thursday, Meta shares lost more than 8% and Microsoft fell almost 4%. Amazon rose only slightly. Just Alphabet gained big, with shares up roughly 10% to a new all-time high.
Each company beat Wall Street's earnings and revenue expectations...
They're all still growing at high rates, too. Meta led the way, growing revenue by 33% year over year in the quarter. Alphabet, Microsoft, and Amazon weren't too far behind – growing revenue by 22%, 18%, and 17%, respectively.
While everyone's headline numbers were good, Wall Street was also focused on the companies' forecasts for spending, AI included. And none of these companies disappointed with its guidance...
- Microsoft sees $190 billion in capital expenditures ("capex") this year, up 60% from 2025.
- Meta raised its capex forecast for 2026 to a range of $125 billion to $145 billion, up from the previous range of $115 billion to $135 billion.
- Amazon maintained its forecast of $200 billion in capex this year.
- Alphabet now sees its full-year capex coming in at $180 billion to $190 billion, up from its prior outlook of $175 billion to $185 billion.
Altogether, these companies now plan to spend $710 billion on capex this year – just about double what they spent last year.
The Hidden Trade-Off Behind AI Growth
This is "good" for the economy – on the surface...
On Thursday morning, the Bureau of Labor Statistics released its latest estimate of America's gross domestic product ("GDP"). The data shows that investment in Information Processing Equipment – a component that tracks AI investment – contributed about 0.8 percentage points to GDP growth in the first quarter.
The entire economy grew 2% in the quarter, meaning AI investment accounted for more than 40% of all economic growth.
As long as the hyperscalers keep spending, they put a "floor" under economic growth.
Capital efficiency is a different story...
As we wrote in February, big tech companies' AI investments have made them a lot less capital efficient over the past few years...
Only one of the four big companies – Microsoft – is expected to be able to cover its capex with the cash it generates from its operations this year. That means the rest are going to have to come up with the money another way – like with new debt or share offerings.
And that's still the case today.
In the first quarter, Amazon only generated about $1 billion in free cash flow ("FCF") – down 95% from the same quarter last year. And with its newly increased capex guidance, Wall Street now expects Meta to run at an FCF deficit this year.
That would mark the company's first year of negative FCF since it went public in 2012.
These companies generated huge amounts of FCF because their Internet-based business models allowed them to scale so effectively. And their stocks have commanded premium valuations as a result.
Now, they're losing the capital efficiency that made them such good investments over the past 15 years.
The market is now demanding AI revenue, not just spending – or enough growth elsewhere to justify those spending plans.
So the AI spending train is still chugging along... But the companies fueling it are running out of time to show the payoff.
All the best,
Corey McLaughlin
Editor's note: Earnings season is starting to test the market's biggest AI winners. And with expectations already stretched, even small surprises will trigger sharp moves. According to Marc Chaikin, founder of our corporate affiliate Chaikin Analytics, a critical shift is already underway. It could change where the biggest gains come from next... and highlight a new group of companies with far greater upside potential.