"It gets late early out there," was Yogi Berra's illogically brilliant take on the late-summer left-field shadows at the old Yankee Stadium.

Also works pretty well as an assessment of today's markets, where constant assurances that the AI theme is "still early" and widespread belief in unceasing economic momentum coexist with some late-cycle characteristics.

S&P 500 operating-earnings growth in the second quarter exceeded 30% and is projected to hit that threshold again in the third, yet the index itself is merely 2% above its June peak from before those stellar results were reported.

The S&P forward P/E has compressed from 23 a year ago to 19 today, which can be read as a signal both that investors are reluctant to pay up for current record profit margins and next year's decelerating growth — and that rising government-bond yields are relevant to the aggregate cost of capital for companies.

Consumer spending has held up fine, because unemployment is low and Americans spend what they make, though real incomes have begun to lag inflation, requiring some dissaving to support consumption.

Corporate-debt spreads have begun to rise in recent weeks from historically tight readings, as technology bond issuance and Treasury volatility combine to test investor risk appetites.

And, of course, the Federal Reserve has transitioned from last year's "insurance" rate cuts to a rate hike, with market expectations of a couple more, which while digestible in a strong economy at least raises the prospect of restraint on an expansion that's carried on for more than half a decade.

Every one of those items can be spun as mid-cycle adjustments, of course, and they may well be. In dynamic markets not governed by a clock, "getting later" is not the same as "almost over." It's tough for an economy to veer quickly into a ditch when the government is running deficits of 6% of GDP while the private sector is investing in new productive assets in multi-trillion-dollar bursts once reserved for emergency government stimulus measures.

And if the valuation compression in the AI-buildout bellwethers is a signal that investors are mindful of an eventual growth plateau, it is also the best argument against the criticism that AI equities are in a bubble. Micron still trading 15% below its June peak and Nvidia sitting at such a steep discount to the market P/E and that of defensive, capex-light Apple, suggest a level of sobriety that – all else equal – could mean that some risk has been drained away.

The extraordinary dissonance in market behavior has made it tough to infer clear macro signals. For weeks, historically anomalous statistics have piled up: Never before have so many new 52-week lows registered with the S&P 500 either up on a given day or so close to a record high.

Treasury yields making new highs has exacerbated the arrhythmic action, as pressure on cyclical parts of the market propels money into the rate-insensitive AI-levered tech giants, which produces all those new lows as the S&P 500 carries higher. In this market moment, very few index rally days are emphatic, no selloffs are fully bereft.

After a June 2 all-time high of 7620, the S&P 500 went sideways in a narrow range for two months and two days until breaking higher to a fresh, if marginal, new record on Aug. 4. Today is two months and two days since that breakout, and the index has merely gone sideways in a narrow range but enters the day about half a percent from another new high.

Yet since the Aug. 4 pop above 7700 for the first time, the median S&P 500 stock, the Russell 2000, the KBW Bank Index and the equal-weighted consumer-discretionary sector are down between 4% and 11%.

There have been times when this kind of unsettled action generates more stress than the broad indexes can absorb, such as in late 2018, when tightening financial conditions and internal market weakness in a mid-term election year caused the S&P 500 to drop nearly 20% into year-end, even with earnings growing 20% and no recession in sight.

That, granted, was a brief if treacherous gut check, giving way to a re-expansion of valuations and a march to new highs through 2019 right up until Covid struck.

Dean Curnutt, founder and CEO of Macro Risk Advisors, articulates the challenge facing risk-aware investors: "The US Treasury market, long referred to as the 'risk-free' asset, is more likely a sponsor of risk than a stabilizing asset in a portfolio…As stock/bond correlation pushes portfolio volatility higher, the correlation among stocks drags it lower. Much lower. The realized correlation among stocks in the SPX over the last 6 months is 4%. This inconceivably low level of correlation is Ozempic for volatility at the index level."

Of course, many people tolerate Ozempic well and are healthier for it. The current market structure has been remarkably generous toward undiscriminating passive index investors. The S&P 500 – by growing more leveraged over time to the strongest secular-growth theme while sidelining disrupted or cyclically vulnerable ones – has protected its owners from deeper drawdowns.

At some point, presumably, the AI-tech dominance of the index will carry too far and the S&P 500's great structural sheltering mechanism will leave it exposed to a bust cycle and a wave of capital destruction. Against that daunting but impossible-to-time prospect, the bond market is now offering a cushion at the best values in years.

This indicator from John Kolovos of Macro Risk Advisors uses several data points to reflect both what investors are saying and what they are doing.

Kolovos on the latest reading: "Bullish sentiment abated last week on the heels of tightening financial conditions and the rise in cross-asset implied volatility. Cyclicals are washed out, and the number of stocks hitting 52 week lows and four week lows, along with the percentage of stocks reaching oversold RSI readings, looks much more like something we would normally see after an 8-10% decline. Yet the S&P 500 has only fallen roughly 3% peak to trough. Historically, that oversold combination has been a good one for forward returns."

-David Snyder, founder of Journey 1 Advisors and the source long known as my Mystery Broker before we identified him late last year, has few qualms about calling this a late-cycle market environment. He has believed since last year that the secular bull market was in its final phase, usually involving a tech-driven momentum blowoff.

In a piece last week, he cited the extraordinary and persistent outperformance of "high-beta" stocks, a phenomenon counter to a generation of finance theory, as something that will eventually face a devastating reversion to the mean.

-As a CNBC journalist, I'm not permitted to invest in individual stocks. And as a late adopter by temperament and disposition, I have interacted very little with AI tools.

So, it was up to a friend who's deeply involved in AI and investing to instruct AI to generate a hypothetical long-short portfolio reflective of my every utterance on TV and each word published under my byline.

The result was quite a bit more conservative than I might have guessed, though the quality and anti-hype attributes make some sense: Long Microsoft, Chevron and Coca-Cola, short Super Micro Computer, Micron and Robinhood. Hope the imaginary AI risk manager had some tight stops on those shorts.

-Reputationally and with plenty of justification, Generation X is not known for its placid acceptance of the "next big thing" that washes over society. Yet in a recent LinkedIn survey, GenX professionals stood out as being most enthusiastic about AI's impact, while their much younger GenZ counterparts are skeptical.

Perhaps it's that GenXers are already established in their careers and not searching for a foothold in a workplace with fewer entry-level jobs. Or maybe this is read by my peers between their late-40s and early-60s as simply another technology transition. In the words of one respondent: "We've lived through every major workplace reinvention of the last three decades…It's just the next chapter." One can hope.

Entering this time of year as a bear has meant fighting uphill, typically.

Below is this year's Dow Jones Industrial Average path set against Ned Davis Research's Cycle Composite chart for 2026. This combines the annual seasonal pattern, the four-year election cycle and the decadal tendencies (years that end in 6).

Right around now is when the tailwinds would start to kick in if historical statistics were all that mattered.

The disclaimers are crucial, though. For one, note the differing scales on the Y-axes – this is mostly about the shape of a year rather than magnitude of moves.

And more to the point, see the profound divergence early this year, when the Iran war and oil surge struck what was supposed to be a first-quarter levitation. It's nice to be aware of the seasonal winds, but as I like to say, such historical patterns are climate, not weather.

SUBSCRIBE TO MIKE SANTOLI'S MARKET MEMO HERE