By Matt Paine, President, Paine Group
I'll start with what I'm not saying. I'm not saying artificial intelligence is a fad. It's already changing how my own team works, and it will reshape whole industries. The technology is real.
My argument is narrower. The price of owning that future has gotten ahead of the future itself, and too many portfolios are riding on it. Meanwhile, one of the oldest wealth-building assets there is has been sitting in the penalty box for four years. For anyone willing to think past the headlines, that combination is an opportunity.
Most investors own the AI trade whether they meant to or not
If you own an S&P 500 index fund, you probably think you own a slice of the whole American economy. You mostly own a few technology companies.
As of mid-2026, the ten largest companies made up about 37.5% of the S&P 500, the highest share since at least the 1970s. A decade earlier it was under 18%. At the peak of the dot-com bubble in 2000, Morgan Stanley researchers put it at roughly 26%. And the weight has run ahead of the results: in 2025 the top ten carried roughly 41% of the index's value while producing about 32% of its earnings.
That isn't diversification. It's one big bet on one theme, wrapped in a product that feels safe.
The spending has outrun the cash
The bull case for AI stocks rests on one assumption: the enormous sums being spent on data centers and chips will earn an attractive return. That assumption is starting to look expensive.
Amazon, Alphabet, Meta, Microsoft, and Oracle went into 2026 planning to spend roughly $660 to $690 billion on capital projects this year, close to double their 2025 levels, and estimates have kept climbing since. J.P. Morgan analysts noted that this single year of spending would likely match everything those companies spent since ChatGPT launched in late 2022. Amazon alone now expects about $220 billion in 2026 and says it will spend meaningfully more in 2027.
Here's the part that should give investors pause. For years these companies paid for growth out of pocket. One analysis estimates that capital spending at the four largest U.S. cloud providers will consume about 92% of their operating cash flow in 2026, up from 41% in 2023. So the money is increasingly borrowed. Meta is planning a record $97.5 billion fundraise for AI. Microsoft is reportedly preparing a $25 billion bond sale, its first U.S. investment-grade issue in more than nine years.
When the strongest balance sheets in the world start borrowing to keep up with each other, that's not proof of a bubble. But it is the behavior you see late in a cycle.
The market is starting to notice
October made this concrete. The Nasdaq 100 fell more than 5.8% for the month, and Reuters reported that over $2 trillion in market value was on track to disappear as investors questioned whether the spending would pay off fast enough. Alphabet posted its biggest one-day percentage drop on record. Several companies beat their earnings estimates and still sold off, because investors had moved from asking "Is AI demand real?" to asking "Does it pay?"
That shift matters. When good news stops lifting prices, the easy money has usually been made.
Real estate: the asset everyone gave up on
Now look at the other side of the ledger.
Commercial real estate has been through one of its longest repricing periods in recent memory. Green Street's Commercial Property Price Index, a standard benchmark for the institutional market, was still about 14% below its 2022 peak in mid-2026, even after rising 4.7% over the twelve months through September. Office has been hit hardest, at roughly 35% below peak. Everyone has heard that story, and it has pulled down sentiment for the whole asset class.
That's the opening. Tech is priced for things to go right. A lot of real estate has already been priced for things to go wrong.
Three forces make the setup stronger from here.
Supply is drying up. High interest rates and construction costs have choked off new building. U.S. apartment starts fell to about 55,000 units in the first quarter of 2026, a 73% drop from the 2022 peak and the lowest quarterly total since 2011. Construction input costs are still running 40% to 45% above pre-pandemic levels. Buildings that already exist in good locations become more valuable when nobody can afford to build competitors.
Inflation hasn't gone away. Consumer prices rose 3.7% over the year through September, above forecasts. Well-located real estate with the ability to reset rents has historically been one of the few assets whose income can rise with prices. A stock trading at a high multiple has no such link.
You can see what you own. Real estate generates income from physical space that people need. You can walk it, inspect it, insure it, and improve it. There's no pile of borrowed money behind somebody else's promise of future returns. For a lot of investors, that clarity is worth something after a decade of abstraction.
"Quality" is doing a lot of work in that sentence
I want to be careful here, because "buy real estate" is as lazy as "buy tech." The asset class is not one thing. Commodity office in a market with falling demand can still lose money for years.
When I say quality, I mean a few specific things:
Locations with real, lasting barriers to new supply: geography, zoning, scarce land, or all three.
Demand drivers that don't depend on a single employer or industry.
Assets bought at or below what it would cost to build them today.
Debt that's conservative enough to survive higher rates for longer.
Operators who know the local market personally, not from a spreadsheet three states away.
Those filters eliminate a lot of what's for sale. What's left tends to be the kind of asset that holds up across cycles.
The honest risks
A contrarian argument that skips the risks is just a sales pitch, so here they are.
Rates cut both ways. The Federal Reserve raised its benchmark rate in September to a range of 3.75% to 4.00% and held there in October, with two officials voting for another increase. The 10-year Treasury yield traded around 5% in October. Higher borrowing costs put pressure on property values, and Green Street has warned that prices could slip if yields stay elevated.
Real estate is also illiquid. You can't sell a building in a few seconds the way you can sell a stock, and you shouldn't own it with money you might need next year.
And the AI trade could keep running. Concentrated markets can stay concentrated for a long time, and some forecasters expect the top ten's share to go higher still. Calling the top of anything is a good way to look foolish.
So I'm not arguing that anyone should sell everything and buy apartment buildings. I'm arguing that if most of your wealth is tied to a single technology theme through funds you think of as diversified, now is a sensible moment to ask whether that's a choice you made on purpose. The point of owning hard assets isn't to win the AI debate. It's to make sure you're fine however that debate turns out.
The bottom line
Every cycle has its version of "this time is different." Sometimes the technology really is different. The valuations rarely are.
AI will change the world. That doesn't mean today's prices for AI companies will pay off. Quality real estate, bought carefully and financed conservatively, is unlikely to make headlines. It's built to still be standing, and still earning, when the headlines move on.
If you'd like to talk through how real assets might fit your plans, we're always glad to have the conversation. Contact Paine Group.
This article reflects the views of the author as of November 2026 and is provided for general informational purposes only. It is not investment, legal, or tax advice, and it is not an offer to sell or a solicitation of an offer to buy any security or investment product. Past performance does not guarantee future results. All investments involve risk, including the possible loss of principal. Consult your own financial, tax, and legal advisors before making investment decisions.
