AI Dominance Is Squeezing Non-Tech Stocks in 2026

The AI Gravity Well: How Tech Is Consuming Capital Markets

Technology's share of the S&P 500 has climbed past 35% of index market capitalization, according to S&P Dow Jones Indices data — a level that would have seemed implausible when the sector accounted for roughly 15% two decades ago. That single statistic explains more about the current state of equity markets than any earnings report or Federal Reserve statement. The index no longer measures the American economy. It measures the American technology industry, with a small remainder attached.

This is the AI gravity well. Capital, talent, and investor attention are being pulled toward a narrow cluster of companies positioned at the center of the artificial intelligence buildout, and the force is strong enough to distort the valuations of everything outside its orbit. The Wall Street Journal reported on October 2, 2026, that stocks outside tech are expected to struggle, earnings are projected to expand more slowly, and concern about credit risk is building. Each of those three developments deserves separate scrutiny — but they share a common root.

The mechanism is straightforward. Index funds and institutional portfolios must hold technology at its weight, which means every dollar of new passive inflow disproportionately purchases AI-adjacent names. Active managers benchmarked against the S&P 500 face career risk if they underweight the sector, so they don't. The result is a self-reinforcing cycle in which strong performance attracts flows, flows lift valuations, and elevated valuations justify continued outperformance in the eyes of allocators. Nothing about this cycle requires the underlying businesses to be overvalued — only that the market's plumbing directs capital in one direction.

What makes the current moment distinct from prior concentration episodes is the magnitude of capital expenditure involved. Hyperscale computing budgets have become a meaningful driver of aggregate corporate investment, pulling semiconductor capacity, electricity generation, and construction spending toward AI infrastructure. That spending shows up as revenue for suppliers and as costs for everyone else competing for the same resources — power, land, skilled labor, and financing.

Non-Tech Stocks Face a Starved Market

Consensus earnings forecasts from FactSet and Bloomberg tell a two-speed story. Technology sector earnings per share are projected to grow at a rate roughly double that of the median S&P 500 sector over the next twelve months, extending a gap that has widened for several consecutive years. Industrials, consumer staples, healthcare, and materials are expected to post single-digit growth — respectable in isolation, anemic by comparison.

That divergence has consequences beyond relative performance charts. When one sector absorbs the majority of incremental equity inflows, the cost of capital for everyone else rises. A mid-cap industrial company issuing equity to fund expansion now competes for investor attention against AI infrastructure names offering a more compelling narrative. A regional bank raising common equity faces the same problem. The market has not closed to non-tech issuers, but it has repriced them.

The practical effect shows up in valuation multiples. Equal-weighted versions of the S&P 500 have trailed their capitalization-weighted counterparts by a substantial margin over the past several years, a gap that mechanically reflects the outsized contribution of the largest technology names. For investors holding equal-weighted exposure — often as a deliberate diversification strategy — the experience has been one of watching a hedged bet underperform an unhedged one.

There is also a second-order effect on corporate behavior. Non-tech management teams facing compressed multiples and skeptical investors tend to prioritize buybacks and dividends over growth investment. That is rational at the individual company level and potentially corrosive at the aggregate level. Starved of equity capital and reluctant to issue it, non-tech firms may underinvest relative to what their long-term fundamentals would support. The earnings slowdown the Journal describes may therefore be partly self-reinforcing.

Credit Risk in the Shadow of AI Dominance

High-yield credit spreads offer the clearest empirical signal of where this is heading. The ICE BofA U.S. High Yield Index option-adjusted spread has compressed toward the tighter end of its historical range for much of 2026, reflecting strong demand for yield in a market where equity returns have been concentrated in a handful of names. Tight spreads are not inherently alarming — they indicate confidence — but they also mean investors are being paid less to bear credit risk at precisely the moment when the composition of corporate borrowing is shifting.

Concern about credit risk, as the Journal notes, is growing. The sources of that concern are worth separating. First, AI infrastructure investment has been financed increasingly with debt, not just cash flow, meaning the leverage embedded in the technology ecosystem is higher than headline balance sheets suggest. Second, companies outside tech that need to refinance existing obligations face a market where underwriting standards have already tightened, and where lenders have more attractive alternatives in the AI supply chain.

For fixed-income investors, this creates a portfolio construction problem. Broad high-yield indices offer limited compensation for the risk that non-tech issuers face a more difficult funding environment precisely when their earnings are growing more slowly. Credit selection matters more than sector duration in a market where the strongest credits are increasingly clustered in one industry. Spread compression has been broad, but the underlying credit quality dispersion has not narrowed — if anything, it has widened.

Historical Parallels: Market Concentration and Its Consequences

Concentration is not new. In 2000, technology and telecom accounted for roughly a third of S&P 500 market capitalization before the dot-com collapse erased trillions in value. In the 1970s, the "Nifty Fifty" — a group of large-cap growth stocks considered one-decision buys — dominated institutional portfolios before valuations normalized painfully. Each episode shared a common structure: a compelling secular narrative, a mechanical reason for capital to concentrate, and a gradual erosion of the discipline that usually limits position sizes.

What differentiates the current cycle is that the largest AI beneficiaries are, unlike many dot-com era darlings, enormously profitable. Their cash flows are real. Their competitive positions are durable. The risk is not that these businesses are fraudulent — it is that their index weights have grown large enough that the broader market's health depends on their continued strength. That is a systemic risk rather than a company-specific one, and it does not require a bear case on any individual name to be legitimate.

The historical lesson is not that concentration inevitably ends badly. It is that concentration reduces the market's capacity to absorb shocks. When a small number of names drives index performance, a disappointing quarter from one of them can move the entire benchmark. Correlations rise. Diversification, the only free lunch in finance, becomes more expensive to obtain.

What Investors Should Watch in a Two-Speed Market

Several indicators will reveal whether the two-speed dynamic intensifies or begins to normalize.

The first is the earnings growth differential itself. If consensus forecasts for non-tech EPS begin to converge toward technology's growth rate, the case for concentration weakens mechanically. If the gap widens further, investors should expect continued index-level distortion. Watch the revisions data, not just the headline estimates — widening dispersion in analyst revisions often precedes visible changes in relative performance.

The second is credit spreads by sector. If high-yield spreads for technology issuers widen meaningfully while non-tech holds steady, that would signal that the market is finally differentiating between AI-adjacent borrowers and everyone else. The reverse — broad spread widening that spares tech — would confirm that capital is being rationed by sector rather than by credit quality.

The third is breadth. Advance-decline lines, the percentage of S&P 500 constituents trading above their 200-day moving averages, and equal-weighted performance versus cap-weighted performance all measure whether participation is broadening. Sustained narrowing breadth has historically preceded periods of elevated volatility.

For long-term equity investors, the implication is not to abandon technology exposure. It is to recognize that passive index exposure now embeds a substantial concentrated bet that most investors did not explicitly choose. For fixed-income investors, the implication is that index-level credit exposure no longer provides the diversification it once did. Both conclusions point in the same direction: in a two-speed market, the average is not a strategy.

Conclusion: Navigating the New Market Hierarchy

The Wall Street Journal's assessment — that non-tech stocks will struggle, that earnings growth will diverge, and that credit risk concerns will mount — describes a market hierarchy rather than a temporary dislocation. AI dominance has restructured how capital flows through public markets, and the structural forces behind it are not self-correcting in the short term.

Investors who understand that hierarchy can still construct portfolios that meet their objectives. They can tilt away from cap-weighted benchmarks where concentration risk is unacceptable. They can select credit exposure deliberately rather than owning broad indices by default. And they can treat the earnings growth differential as a variable to monitor rather than an assumption to accept.

The gravity well is real. But gravity, unlike fashion, is predictable. That predictability is where opportunity lives.

Source: WSJ.com: Markets