Nothing Else Matters

If you judged 2026 solely by the S&P 500’s year-to-date performance⁽¹⁾, you might conclude that this is shaping up to be another uneventful year, a continuation of the durable bull market. That would certainly be a reasonable view, but an incomplete one.

Beneath the surface, the market has been going through a massive reshuffling.

The companies that carried the market for the better part of three years (the famed Magnificent Seven), have stumbled. As a group, they are down ~4% this year, even as the broader market moved higher.

The headline index performance also masks volatility in its constituents. As Schwab shows below, the average stock in the S&P 500, Russell 2000 and the Nasdaq is down ~20%, ~30% and ~40% respectively, from YTD highs.

Source: Charles Schwab. As of 7/21/2026

Day to day, keeping up with sector leadership is nearly impossible.  

Semiconductor stocks routinely swing several percentage points in a day. Software companies alternate between market leaders and laggards within the same week. The VIX, a gauge of market volatility, remains relatively subdued, yet single-stock volatility is near the highs of the year.

These hundreds of seemingly unrelated stories are really all part of one conversation: Artificial Intelligence.

AI is driving corporate capital spending, reshaping earnings expectations, influencing credit markets and boosting the real economy. We’re way past the question of whether AI matters; the question now is who gets paid. And this is the question investors seem to answer differently every few weeks.

Last year, the answer appeared obvious.

The hyperscalers (Microsoft, Amazon, Alphabet and Meta) owned the cloud infrastructure and had both the financial resources and strategic imperative to build the computational foundation of the AI economy. Investors applauded every increase in capital spending to bring AI to the masses and enterprise. The larger the data-center budget, the more confidence the market seemed to have that these companies would dominate the next decade.

Massive investment was viewed as insurance against obsolescence and a claim on future profits. Today, those same spending announcements are greeted differently.

Free cash flow that once funded buybacks and dividends is increasingly being redirected toward data centers, semiconductors, electrical and cooling infrastructure and long-term power contracts. Balance sheets are becoming more leveraged. The hyperscalers have become meaningful issuers in the corporate bond market as they finance one of the largest capital spending cycles in decades.

That change in perception explains much of the market’s action this year.

If hyperscalers were determined to spend hundreds of billions of dollars building infrastructure, someone else was bound to benefit. Semiconductor companies suddenly looked like the more attractive expression of the same investment theme. One company’s capital expenditure is another’s revenue.

Earlier in the AI cycle, hyperscalers accounted for much of the incremental S&P 500 earnings growth. Today, semiconductor companies are expected to contribute nearly half of that growth.  

Source: JP Morgan. As of July 20, 2026.

This view, though, is tenuous as well.

Some mornings the market becomes convinced that demand for computing power is effectively limitless. Semiconductor stocks rally sharply while software companies fall on the assumption that increasingly capable AI models will commoditize much of what enterprise software does today.

Then a new open-source model arrives, more efficient and cheaper to run, and suddenly the view shifts. What if we are overbuilding data centers? What if compute becomes abundant? Could it be that the long-term winners are not the companies selling the infrastructure at all, but those building applications on top of it?

Nothing fundamental has changed overnight, except for the bet on the expected winners.

But this churn has produced something positive.

For the past three years, one of the most common criticisms of the market was that it was too reliant on the continued strength in just seven companies. This year, the criticism has fallen apart. Industrials, Materials and Utilities are all in an uptrend. The equal-weight version of the S&P 500 index is outperforming the main cap-weighted index.

The rally has broadened, and it is encouraging, but the underlying theme has not changed. Utilities are benefiting because AI requires electricity. Industrial companies are benefiting because AI requires transformers, generators and cooling equipment.

By most accounts, we are still in the early innings. The Census Bureau estimates that only one in five businesses are currently using AI in “any business function”…a fairly generous definition of “use”.⁽²⁾

Yet investment in computing infrastructure has exploded. Private investment in software and computer equipment has been one of the strongest engines of investment growth, even as overall fixed investment outside of the AI sector has softened.

Source: Danske Bank Research. March, 2026.

The economy is building AI infrastructure years before most businesses fully understand how they will use it. Demand is growing however, and whatever compute capacity currently exists, it is fully utilized. Cloud providers continue to highlight the energy and materials bottlenecks that prevent them from bringing more capacity online.

We do know from history that new technologies that create extraordinary economic value can also produce periods of overinvestment and disappointing shareholder returns. Railroads transformed America, but many railroad investors went bankrupt. The internet reshaped nearly every industry, yet countless internet companies disappeared along the way.

The implication of which outcome is more likely this time around extends far beyond Silicon Valley. It reaches into the bond market, where technology companies increasingly finance AI infrastructure through debt issuance. It reaches into industrial production, utilities and commercial construction. Most importantly, it reaches into household balance sheets. After one of the strongest bull markets in modern history, equities now represent the highest share of household net worth.

Source: Bloomberg. As of July 10, 2026.

The performance of financial markets increasingly matters to the real economy through the wealth effect. Rising equity prices support consumer spending, encourage business confidence and make capital easier to raise.

A reassessment of AI economics would influence financing conditions, corporate investment decisions and ultimately, economic growth itself. That is ultimately why every earnings call, every capital spending announcement and every new data-center project is now so scrutinized. The fundamental question is whether the largest investment boom in a generation will earn a positive return.

Markets are exceptionally good at recognizing that something important is happening. But identifying, in real time, where AI’s profits will ultimately land is challenging (and exciting). Until that answer becomes clearer, market leadership will continue to rotate. Single stock volatility will remain elevated beneath a seemingly calm index and every quarterly earnings report will be interpreted through the same lens.

Enjoy your weekend and summer.

Alex

(1) S&P 500 +8.3% YTD, as of 7/24/2026. Source: FactSet

(2) https://www.census.gov/library/stories/2026/05/ai-use-businesses.html