The Problem With AI Obsession
Trump's 12-out-of-10 deal with China... A pause on the trade war... Looking closer at Magnificent Seven earnings... When capital efficiency dies... Another crack in the AI boom...
It was a 12 out of 10...
That was President Donald Trump's report on his meeting with Chinese President Xi Jinping yesterday in South Korea. As for details, it sounds like both sides got something out of what amounts to a de-escalation agreement between the world's two largest economies.
"We have a deal," Trump told reporters on Air Force One late yesterday following a more than 90-minute meeting between the leaders and other U.S. and Chinese officials. "On a scale from zero to 10, with 10 being the best, I would say the meeting was a 12."
Among the big points...
The White House has agreed to lower the "fentanyl tariff" on China by 10% and to suspend the 24% Liberation Day "reciprocal" tariff on China for one year from now. And China has said it will pause its recently announced tighter export controls on rare earth minerals and related products for a year.
China has also agreed to start buying American soybeans again "starting immediately," Trump said, which it hasn't been doing since the tariff war broke out earlier this year... And Chinese companies will be allowed to buy semiconductors from the likes of Nvidia (NVDA), though it's unclear if the most advanced Blackwell chips will be freely traded.
Of course, this could all change with another Truth Social post in a few months, and the points could be up for negotiation in a year. But for now, U.S.-China trade relations are not as bad as a few weeks ago, which is when Trump threatened an additional 100% tariff on everything from China and dropped stocks from all-time highs...
Trump said he'll be going to China in April, and Xi will come either to Washington, D.C. or Mar-a-Lago after that. We'll see what's on the agenda by then – and what the global markets make of it. In the meantime, though, it looks like investors have moved on to other things...
A mixed start for Magnificent Seven earnings...
As we briefly noted last night, three of the Mag Seven reported earnings after yesterday's close. Alphabet (GOOGL), Meta Platforms (META), and Microsoft (MSFT) all beat Wall Street's consensus expectations on revenue and earnings.
But let's dive in a little deeper today, starting with the terrific-looking headline results.
Alphabet grew its overall revenue 16% year-over-year and reported its first quarter with more than $100 billion in revenue. And it grew net income by 33%.
Microsoft grew its revenue by 18%, thanks to 40% growth in its Azure cloud business. And net income grew 12%, even after the company took a $3.1 billion hit from its investment in OpenAI.
And Meta Platforms giant grew its revenue 26% in the third quarter – marking the social media giant's highest growth rate since the first quarter of 2024.
Now to the more important part – AI spending...
All the headline numbers were solid, but it was what's below the surface that moved the stocks after hours yesterday and today.
As we all know, investors are heavily focused on the AI story, and all eyes were on these companies' capital expenditures ("capex").
All three companies raised their spending forecasts...
Alphabet now expects 2025 capex between $91 billion and $93 billion, up from its previous outlook of $75 billion to $85 billion. The company said it's "investing to meet customer demand" with its Gemini AI service topping 650 million monthly average users.
Microsoft's capex came in at $34.9 billion in the company's fiscal first quarter, above the $30 billion level that Chief Financial Officer Amy Hood forecast in July. And on the company's earnings call, Hood said that capex growth would accelerate in 2026 from 2025, after previously saying capex growth would slow. (Microsoft's capex jumped 45% in the 2025 fiscal year.)
As for Meta, the company said that capex growth "will be notably larger in 2026 than 2025." And in 2025, it's already spending more than it thought. Meta now forecasts capex in a range of $70 billion to $72 billion, higher than its previous outlook of $66 billion to $72 billion.
So more spending is on the way. But here's the interesting part, which could be a telling signal that we've been looking out for...
The market did not like Microsoft and Meta's guidance about accelerating capex – especially with no clear road to profitability in sight from all this AI-related spending. Shares of both were lower today, with Microsoft losing nearly 3% and Meta down by roughly 11%.
The major U.S. stock indexes were lower today, with the Dow Jones Industrial Average down slightly, the small-cap Russell 2000 Index and benchmark S&P 500 Index down almost 1%, and the tech-heavy Nasdaq Composite Index off by about 1.5%.
The Mag Seven are losing their capital efficiency...
Joe Weisenthal, of Bloomberg's "Odd Lots" podcast fame, shared a great graphic on X showing just how much capex has soared for these companies. Alphabet and Microsoft now spend 10 times as much on capex as they did a decade ago... and Meta spends 20 times as much.
These companies became such great investments because of their software-based businesses, enhanced by great "network effects."
As Stansberry Research founder Porter Stansberry wrote in the December 2007 issue of Stansberry's Investment Advisory...
This is the beauty of capital-efficient businesses: As sales and profits grow, capital investments don't. Thus, the amount of money that's available to return to shareholders not only grows in nominal dollars, it also grows as a percentage of sales.
Think of Microsoft's software over the years... It hasn't cost the company any more in expenses to issue new Microsoft Office licenses. So every incremental license sold is almost pure profit.
Today, their capital-efficient models have changed. The Mag Seven are investing more than 60% of their operating cash flow into AI build-outs. They're building costly data centers filled with power-consuming servers that need to be updated and maintained. And the more AI users they get, the more computing power these companies need to supply.
This is the real risk of big tech's AI obsession...
The better news...
These three mega-cap tech companies still generate plenty of free cash flow ("FCF") for now. Our proprietary Stansberry Score still gives all three of these companies "A" grades on capital efficiency.
If you had to pick one, our colleague Whitney Tilson has written time and time again about Alphabet's performance.
He wrote about Alphabet's earnings report in his Whitney Tilson's Daily newsletter this afternoon. As Whitney noted, the company's total costs and expenses grew slightly faster than revenues, so Alphabet's operating margins fell slightly. But they're still at 31%...
With trailing-12-month revenue of $385 billion, it's astounding that a company of this size can keep growing so quickly and profitably.
And he's not concerned about the stock's valuation, writing...
The stock opened around $292 this morning, and its 2025 earnings estimate is $10.53 per share (the $9.92 estimate coming into earnings, plus the $0.61 beat). That means its price-to-earnings (P/E) multiple is 27.7 times.
That's higher than it has been in a while thanks to the huge run-up this year. But it's still only a modest premium to the S&P 500's 22 times to 24 times forward multiple today (depending on whose estimates you use).
My view today is still the same as it has been for more than six years: Alphabet is a great stock for conservative, long-term-oriented investors.
Alphabet was up about 2.5% today. You can read Whitney's full analysis on the company's earnings report – and that of Meta Platforms, which Whitney said, "looks like a buy" – in his free daily here.
The AI story is as important as ever...
As Jim Bianco of Bianco Research shared on X over the weekend, analysts at JPMorgan have separated the S&P 500 into 41 "AI" stocks and 459 "others." And right now, those 41 AI stocks make up 47% of the index's market cap.
Now, heavy concentration is nothing new. And as regular readers have heard many times from us, that's all well and good when markets go up.
JPMorgan's data is perfect evidence of that...
Since the release of ChatGPT in November 2022, its group of 41 AI stocks is up an average of 194%, versus "only" a 27% gain for the rest of the index.
Perhaps more worrying, though, is how those non-AI stocks have only returned about 1% over the past 11 months, Bianco shared. This is a significant "divergence" in the market.
If you don't have exposure to AI in your portfolio, there's a good chance you're missing out on the bull run...
We saw how this divide played just earlier in the week when the S&P 500 reached a new all-time high, despite nearly 400 of the 500 stocks being down on the day. That was the worst breadth ever for an up day in the S&P 500.
And as we've talked about the past few weeks, we're wary of the "circular" financing among AI-heavy businesses... And debt-fueled investments in the style of OpenAI's recent agreement with Oracle (ORCL) are concerning.
That's a point that Stansberry Research senior analyst Brett Eversole talked about last week at our conference in Las Vegas. He warned that such deals – where Oracle could borrow around $100 billion over the next several years to build cloud services for OpenAI – could be "the beginning of the end" of the so-far-healthy AI boom.
If investors increasingly lose faith in the AI story (like they did today with Microsoft and Meta), it could signal a top in the market. That's what happened with Meta stock as recently as 2022. Investors lost faith in the company's heavy investments in the metaverse without a clear business model.
Of course, that sell-off then presented a great buying opportunity for Meta... eventually, and the company reorganized its operations to set the stage for where it is now... Just today, Meta reportedly received a record $125 billion in orders for a $25 billion investment-grade corporate debt sale, indicating the company is having no problem finding funding for AI investments.
But we will keep an eye on any further cracks in the AI boom. This quarter could mark the first one when investors are starting to meaningfully question heavy investment in AI. And if that continues throughout the next couple quarters of increased spending, the AI "leaders" could soon lead the market lower rather than higher.
Alternatively, with continued plans from mega-cap tech companies to spend even more on AI, the broad bull run could continue. After all, sometimes these trends can go on longer than you might think or "make sense."
But the market reaction today to Microsoft and Meta's earnings report tells us that high expectations might be becoming more of a liability for these companies... and their share prices. And that's a risk to consider for the S&P 500's performance in general.
New 52-week highs (as of 10/29/25): Applied Materials (AMAT), Arista Networks (ANET), ASML (ASML), Atmus Filtration Technologies (ATMU), Broadcom (AVGO), BHP Group (BHP), Alpha Architect 1-3 Month Box Fund (BOXX), BWX Technologies (BWXT), Ciena (CIEN), iShares MSCI Emerging Markets ex China Fund (EMXC), EnerSys (ENS), iShares MSCI Spain Fund (EWP), iShares MSCI South Korea Fund (EWY), Comfort Systems USA (FIX), Futu Holdings (FUTU), Alphabet (GOOGL), Hubbell (HUBB), iShares Convertible Bond Fund (ICVT), KraneShares Bosera MSCI China A 50 Connect Index Fund (KBA), Lumentum (LITE), Mueller Industries (MLI), ProShares Ultra Technology (ROM), Taiwan Semiconductor Manufacturing (TSM), ProShares Ultra Semiconductors (USD), Vale (VALE), and Vanguard S&P 500 Fund (VOO).
In today's mailbag, we have a few notes from folks who weren't able to catch the debut of Stansberry Research senior analyst Gabe Marshank's free presentation yesterday. Good news: If you're interested, you can watch a replay at your convenience here.
And we suggest you do...
Over a two-decade-plus Wall Street career, Gabe delivered big returns for the likes of hedge-fund billionaires Steve Cohen, Leon Cooperman, and David Einhorn. And over the past few years, he has turned his attention to making money for everyday investors...
Since joining Stansberry Research, Gabe called the AI boom and the rise of nuclear stocks.
In this new presentation, he shares his next favorite 25x-upside idea and shows how he racked up nine-figure hedge-fund wins by looking where others don't – a strategy that you can deploy immediately.
All the best,
Corey McLaughlin and Nick Koziol
Baltimore, Maryland
October 30, 2025