The State of the Union for Real Estate Investors and Private Lenders

August 12, 2026.

What this article will teach you

●       Why 2.1% GDP growth is masking an unusually concentrated AI-driven investment boom

●       How record corporate profit margins and productivity gains can coexist with weak hiring and a strained consumer

●       Why falling unemployment can actually reflect labor-market weakness rather than strength

●       How oil could create a new inflation problem just as wage growth and real consumption are losing momentum

●       Why the AI buildout is increasingly being financed with debt, leverage, and circular capital flows

●       Why razor-thin credit spreads don’t mean risk has disappeared—they may mean investors are barely charging for it

Register for Our Latest State of the Union Webinar

Want to go deeper into the data? Register for our latest State of the Union webinar at https://justbethebank.com/sotu-elite-officer.

In this presentation, we break down the latest trends in housing, lending, and the broader economy, along with how our Family Office interprets the data and positions capital in today’s market. Registration is free!

Nominal confusion: the growth is real, but highly concentrated

In my July 2025 SOTU, I talked about the idea of “nominal confusion.” If it’s not ringing a bell, don’t worry—it’s not a term you hear all that often anywhere, economics included.

But it’s still the best one I can find to describe something we seem to be seeing everywhere these days: numbers that feel or look like growth, but are inflated, subsidized or structurally distorted.

Case in point: real GDP, which came in at 2.1% annualized for Q1 2026.

That’s respectable. Nothing to write home about—and below the historical average of 3.2%—but still decent.

However, as I’ve said before: headline GDP can lie. Or at least mislead.

Dig into the components and you find that nonresidential fixed investment accounted for 1.42 percentage points, or 67.6%, of all Q1 GDP growth.

That subcategory has literally never represented that much of the quarterly growth. Certainly not in the preceding eight quarters (as you can see in the graphic below), nor at any other point in our 250-year history.

And within that category, one sub-subcategory towered over the rest: information-processing equipment. In plain English, that’s the computers, servers and chips going into AI data centers.

Software followed closely behind, with research and development coming in third.

Now for the math: add those three line items together (0.77 + 0.52 + 0.21) and you get 1.5 percentage points of growth—more than the entire investment category itself.

Meaning, AI/tech spending, represented 111% of the growth of its parent category. That can’t happen, by definition, unless other areas went negative… unless capital was leaving some of the more traditional subcategories.

So when you hear, “The economy grew 2.1% in Q1,” or “1.5% in Q2” (which it did), you need to know there’s a huge K underneath.

What really happened is this: One enormous, concentrated, borrowed bet…grew.

Take that one bet out, and there’s very little underneath it.

That’s the thesis I want you to keep in mind throughout this update: we’re navigating a hyper-focused technology boom, not a broad, well-rounded economic expansion.

The Corporate Credit Cycle: start with capex

If you read my last article, you’ll remember I talked about dispersion and our increasingly K-shaped economy. I said this cycle isn’t one wave—it’s a set of lanes moving at different speeds, with the separation between those lanes getting wider.

And this time, that K-shape isn’t just showing up among households.

It’s showing up inside corporate America too.

A useful way to see what’s happening inside corporate America is through the Corporate Credit Cycle.

Markets move through early-, mid- and late-cycle conditions before eventually entering a downturn or recession. And while no cycle follows the textbook perfectly, the usual indicators give us a useful way to see where we are.

We’ll go a bit out of order and start with the elephant in the room: capex.

Mid-cycle, capex tends to look fairly stable. Companies focus more on optimizing the assets they already have and improving efficiency without adding a lot of new overhead.

Late-cycle looks different. That’s when companies begin running into physical production ceilings and have to spend aggressively on new infrastructure and capacity to meet demand.

So what are we seeing now?

For most of corporate America, today’s numbers look fairly modest. Excluding the Mag 8, S&P 500 capex rose just 9% year-over-year in 2025 and is expected to rise 15% in 2026 —easily within the “stabilizing” bucket for purposes of our Credit Cycle analysis.

For the eight hyperscalers, however, capex isn’t just rising. It’s vertical.

The Mag 8 spent $245 billion on capex in 2024 and $397 billion in 2025. By the end of 2026, that number is projected to hit $724 billion—an 82% year-over-year increase following a 62% jump the year before.

That puts tech capex in a completely different part of the cycle than real-world capex.

And it raises the obvious question: will hyperscaler income keep pace with all that spending?

Highly unlikely.

Their share of S&P income isn’t growing nearly as fast as their share of capex. In fact, capex is roughly double income, and most projections show hyperscaler capex overtaking cash flow later this year.

Translation: they could become cash-flow negative, possibly as soon as this quarter.

This is where the AI-bubble worries come from. History says that when investment races this far ahead of cash flow, the next phase usually isn’t more acceleration.

It’s figuring out whether all that new capacity can actually earn an adequate return.

Corporate profit margins: record highs, and doing more with less

Credit Cycle indicator #2 is corporate profit margins.

Normally, margins increase mid-cycle because demand surges while structural costs remain relatively fixed. By late-cycle, they tend to plateau as tight labor markets push wages higher and capacity constraints drive up resource costs faster than companies can raise prices.

So where are we now?

FactSet has the S&P 500 net profit margin at 14.7% for Q1, the highest they’ve measured since they began tracking it 17 years ago. And that’s up from the prior record of 13.2%…the quarter before.

So what could explain that, especially when consumers don’t seem to be spending as much?

My belief is that we’re increasingly seeing AI show up on the income statement—even in boring sectors like Financials, Utilities and Industrials. We just watched it prop up GDP from the top line. Now you’re seeing AI lift profitability from the cost side, across almost every sector.

And when I say “cost” side, I specifically mean labor input.

According to the BLS numbers, 91.6% of all business-sector output growth is coming from greater productivity—“doing more with less”—rather than expanding the workforce.

That helps explain how you can have record corporate margins…but without the kind of broad hiring boom you’d normally associate with them.

But profit margins are probably the most mean-reverting series in finance. And when they do revert (and they’ve got to, at some point), there’s reason to think that reversal could be even sharper than usual… precisely because they’ve climbed so high.

Capitalism is a self-correcting system, after all. The higher profit margins climb, the more competition they attract—and the more pressure they create for their own reversal.

Look, clearly there’s growth in this economy. It’s not imaginary.

But if we have this much business investment, record profit margins and that level of productivity growth, it’s worth asking:

Why doesn’t it feel more prosperous for the rest of us?

Consumer spending: people aren’t buying more, they’re reallocating

Part of the answer is sitting in consumer spending.

Earlier this cycle, spending regularly outpaced inflation. Not so much anymore; over the last nine months, spending growth has become smaller and less consistent, just as PCE inflation has started accelerating back toward 3% and beyond.

That’s why the minor blip of spending resilience we saw in May—up 0.26% month-over-month—is less encouraging than it looks.

People aren’t confidently buying more stuff. They’re reallocating more of their cash toward necessities… while quietly starving discretionary categories like clothing, electronics and dining out.

And that spending is becoming increasingly “installment-ized.”

According to a recent LendingTree survey, 29% of Buy-Now-Pay-Later users are now putting groceries on installment plans—more than double the 14% from two years ago. Groceries are now the third most commonly purchased category using BNPL.

At the same time, 47% of BNPL users made at least one late payment in the past year, up from 41% a year ago.

So usage is up generally, it’s increasingly being used for essentials, and more people are paying late. Sure sounds like the consumer is feeling stretched.

And it gets worse. This stress can be easy to miss because BNPL doesn’t show up cleanly in traditional credit stats. Economists use the term “phantom debt” for exactly this reason: the official credit card/savings rate/delinquency numbers understate just how stretched consumers really are.

And that pressure could get worse if energy prices move higher.

Oil, labor and the Fed

Oil is definitely the big X factor right now.

Because it’s not just what we pay at the pump; oil is everywhere. It’s the “universal input,” in economic terms.

Now, if the real economy were strong, higher energy prices might not be a huge problem.

But that’s not the backdrop we have at the moment — especially when you look at the labor market.

Just take a gander at the tape:

●       The U.S. lost –23,000 jobs in July.

●       April, May and June were revised down by a combined 177,000 jobs.

●       Now five of the last 12 months have seen negative job growth.

And yet the unemployment rate actually fell over the last two months, down to 4.1% as of July.

Which sounds like good news until you look at why: the labor force shrank by roughly 720,000 people in June, and another 264,000 in July.

This is nominal confusion at its finest: the unemployment rate looks better not because hiring suddenly improved, but because nearly 1,000,000 people left the labor force in just two months.

A smaller denominator means a better-looking ratio…even though fewer total people were actually employed.

Real wages aren’t providing much protection either.

The BLS reported that real average hourly earnings declined in both April and May as inflation outpaced nominal wage growth. Even nominal wage growth (the light blue line) has been trending down.

So households are being asked to absorb an energy shock in the middle of:

●       slower hiring,

●       weaker real and nominal wage momentum,

●       and already-soft growth in real consumption.

That makes this recent PCE spike look a lot less like demand-driven inflation… and more like a tax on demand generally.

So what can we expect from the Fed?

Well, this combination of weak job growth and higher prices puts Warsh and company in a lousy spot.

As of August 10, the market was assigning zero probability to a rate cut before the end of the year, with only a 23% chance the Fed would even hold rates where they are now.

Every other probability was for a hike, including a combined 33% chance of at least 50 basis points—enough to put the policy rate back above 4%.

Leverage: the AI boom is increasingly running on debt

Credit Cycle indicator #3 is leverage.

In a typical cycle, leverage trends lower mid-cycle because surging profits dilute corporate debt burdens and allow companies to fund more of their expansion internally.

Later in the cycle, slowing growth tends to push leverage higher as companies borrow to fund capex, M&A, stock buybacks, etc (often to artificially sustain their financial performance).

So where are we today?

Corporate debt issuance—whether through investment-grade bonds, high-yield bonds or loans—has grown significantly over the last three years.

Meanwhile, equity issuance (newly issued corporate stock) was mostly flat from 2022 through 2025 and hasn’t come close to keeping pace in 2026.

So it’s no surprise that Big Tech is issuing bonds like crazy. Nvidia just completed its largest corporate bond deal ever at $25 billion, with an order book near $85 billion—more than three times oversubscribed.

SpaceX did the same less than two weeks after its IPO.

And it’s not just those two companies. AI-related borrowers had already sold roughly $300 billion of bonds as of May, even before the Nvidia and SpaceX deals. Morgan Stanley expects the total to finish 2026 well above $400 billion.

Retail investors are getting in on the leverage too. They’re no longer just buying shares of the Mag 8; they’re piling into leveraged ETFs, derivative-backed funds that use synthetic swaps to supercharge returns.

Trading inside those ETFs—as measured by notional volume—is running roughly 50% above 2025’s record baseline.

That massive volume forces market makers to constantly buy and sell billions of dollars of the underlying stocks to hedge their books… creating a feedback loop that can push AI stock prices higher and amplify volatility.

Wrapped around all of this is something you’ve probably seen online… what’s become known as the AI infinite money loop.

●       A chipmaker like Nvidia takes an equity stake in an AI lab like OpenAI

●       The lab signs a multi-year deal with a cloud provider like Oracle

●       Oracle spends tens of billions of dollars on Nvidia chips for its data centers.

Rinse and repeat.

For a while, that circular-financing loop was mostly treated like a meme.

But on June 28, the Bank for International Settlements—the central banks’ central bank—published its 2026 annual report. In it, they identified their three biggest threats to the global financial system:

1. Record sovereign debt
2. An AI capex bust
3. And the collapse of those circular financing deals.

Credit spreads: when calm looks like complacency

Credit Cycle indicators #4 and #5 are defaults and credit spreads. As far as I can tell, the market has all but stopped pricing in default risk.

The cleanest gauge is the spread on high-yield junk bonds—the extra yield investors demand to lend to a below-investment-grade company instead of buying a Treasury. Right now, that spread is around 2.7%, a hair off its cycle low.

Most people see spreads this tight and assume they mean risk is low. That’s the trap.

A 2.7% spread isn’t the market telling you defaults won’t happen; it’s the market charging almost nothing for the possibility that they will.

Investors are effectively saying: “we’ll lend to some of the shakier companies out there for a measly 2.7% more than a risk-free Treasury, even though that may be nowhere near enough compensation if things go wrong.”

You can see that same “everything’s fine” attitude straight down the credit ladder.

Late in a cycle, this kind of calm isn’t necessarily safety—it’s often complacency.

Defaults always bottom before they turn. A spread pinned to the floor isn’t a green light; it’s a spring wound as tight as it goes.

Conclusion: mostly late-cycle signals, some wearing a mid-cycle mask

As we wrap up here, let’s zoom out and look at the whole picture:

Capex is wildly K-shaped, but at the top of the K, spending has gone vertical; to my way of thinking, that looks textbook late-cycle.

Profit margins are a bit murkier. On paper, record margins look mid-cycle, but given how much of that strength appears to be coming from productivity and reduced labor input, it’s possible we’re looking at late-cycle wearing a mid-cycle mask.

Leverage is much less ambiguous: it’s through the roof.

And even cash reserves, which technically still look mid-cycle, become harder to interpret. Big Tech is approaching cash-flow negative territory despite borrowing enormous amounts… rather than funding the buildout entirely with its own cash.

M&A tells the same K-shaped story from another angle. The first half of 2026 produced the highest deal value for any first half since 1980, but on the lowest deal volume in six years—fewer deals, much more money.

A handful of giants are buying the future while everyone else largely sits it out. And with fewer AI companies left to acquire, some of that appetite is simply migrating from M&A into capex as those companies build data centers instead.

That leaves credit spreads as one of the only indicators that doesn’t neatly fit the thesis. According to the econ theory textbook, they should already be widening as the cycle ages. They aren’t.

And the historical record doesn’t even match the theory. The high-yield spread index hit its all-time low of 2.41% in June 2007; two months later, the credit market froze. Today, we’re sitting at 2.74%.

The lesson isn’t that 2026 is 2007. It’s that credit spreads don’t necessarily give you a nice, gradual warning as risk builds—they can sit on the floor, tell you nothing, and then gap.

So I don’t look at today’s tight spreads and conclude that we must still be safely mid-cycle. Sometimes the indicator that appears to be saying “everything’s fine” is simply saying nothing at all.

Register for Our Latest State of the Union Webinar

Want to go deeper into the data? Register for our latest State of the Union webinar at https://justbethebank.com/sotu-elite-officer.

In this presentation, we break down the latest trends in housing, lending, and the broader economy, along with how our Family Office interprets the data and positions capital in today’s market. Registration is free!