FEATURED

Your Index Fund Is Now a Chip Fund

Here is a number that should reframe how you think about your 401(k).

Semiconductor stocks now account for a record 19.7% of the S&P 500, quadrupling since 2020. Not tech broadly. Not software. Semiconductors specifically. Nearly one dollar out of every five in the index.

That concentration has now blown past the tech bubble peak. Ten years ago, the semiconductor weight in the S&P 500 was 2%. Today it sits at 19.7%, more than double what it was at the dot-com peak. The math behind that number is almost entirely a function of a handful of names, led by NVIDIA, Broadcom, TSMC, AMD, and more recently memory stocks like Micron and SanDisk.

A passive investor who thinks they own the whole market actually owns a portfolio where one out of every five dollars is riding the semiconductor cycle. That is a meaningful exposure most people do not realize they have.

And the feedback loop is self-reinforcing. As semiconductor shares outperform, their weighting within the index rises, prompting passive funds to allocate even more capital to the same companies. That additional demand supports further price gains and pushes index weights higher still. ETFs attracted more than $1 trillion in inflows year-to-date through late June 2026, running roughly 45% above the record pace from the prior year. Every dollar of that flows into chips by default.

The Warning Signal Almost Nobody Is Discussing

Bank of America’s proprietary Bubble Risk Indicator, scored on a zero-to-one scale where one signals extreme bubble-like price action, registered 0.91 for the PHLX Semiconductor Sector and 0.82 for the Technology Select Sector as of July 1, 2026. The Nasdaq 100 sits at 0.69 on the same scale. Bank of America stopped short of declaring a full bubble but urged investors toward greater valuation discipline after an extended semiconductor rally.

The valuation warnings are stacking up elsewhere too. The S&P 500’s price-to-sales ratio sits at 3.22, well above its long-term historical average of 1.84. The Buffett Indicator, which compares total U.S. stock market capitalization to GDP, currently stands at 231.8%, a level that signals the market is significantly overvalued by historical standards.

On the technical side, the SOX index’s 200-day moving average currently sits near 10,700. The index peaked above 14,650 in June before pulling back sharply on July 1, the first trading day of the second half, when chip stocks fell between 3% and 9% in a single session. The iShares Semiconductor ETF (SOXX), which had gained nearly 89% in the first half of 2026, was down over 5% on that open.

That last point matters. One sector dropping 5% on a single day is one thing. One sector that represents nearly 20% of the S&P 500 dropping 5% is a different problem entirely.

The iShares Semiconductor ETF (SOXX) gained roughly 89% in H1 2026, approximately 8 to 9 times the S&P 500’s gain over the same period. When one sector moves at that multiple relative to the index, the question is not whether a pullback is possible. The question is what happens to everything else when it arrives.

The Other Side of the Argument

This is not a simple bubble call. The earnings are real.

The AI buildout is not theoretical. The four largest hyperscalers combined are now expected to spend close to $700 billion on capital expenditures in 2026, a more than 60% increase from already historic 2025 levels. Approximately 75% of that, or around $450 billion, is tied directly to AI infrastructure: servers, GPUs, data centers, and the chips that power all of it. That capital has to land somewhere, and it lands at the chip designers and equipment makers.

Not everyone frames the concentration as purely a risk. As one market participant put it, the folks selling the picks and shovels are in incredibly good stead, while those buying them still have to prove that the billions they are spending is worth it. That distinction between chip suppliers and their hyperscaler customers is increasingly the pivotal debate on Wall Street.

Slight tangent, but it matters: AI capex consensus estimates have proven too low for two years running. At the start of both 2024 and 2025, consensus implied capex growth of roughly 20% for the year. In reality, it exceeded 50% in both years. The hyperscalers have consistently surprised to the upside. That track record is part of why the market keeps giving the sector the benefit of the doubt.

Goldman Sachs has noted that the AI capex boom has lifted semiconductor profit margins sharply, but rising depreciation expenses from hyperscalers will start to pressure some of that profitability over the next few years. That is the timeline the market is not fully weighing.

Trade Framework

For long-term holders: The concentration risk is not a reason to exit. It is a reason to understand what you actually own and size your semiconductor-specific positions accordingly, rather than assuming your index fund is balanced. It is not.

For active traders: The gap between the SOX and equal-weight S&P 500 exposure has widened dramatically. Any rotation broadening out of the largest chip names and into equal-weight or small/mid-cap exposure offers a directional trade if that broadening accelerates. July 1 saw small and mid-cap names holding up better than chip stocks, which is worth watching as a potential early signal.

For options traders: Implied volatility across the semiconductor complex has compressed significantly despite elevated bubble risk readings. Put spreads on SOXX or individual large-cap chip names, with strikes below current support, offer defined-risk downside hedging at relatively modest cost given where vol sits. This is not a directional bear bet. It is portfolio insurance at a reasonable price, and Bank of America has flagged a similar relative-value hedging approach for exactly this environment.

The Magnificent Seven now account for roughly 30% of the S&P 500’s total market value. Semiconductors alone account for 19.7%. The party, as one analyst described it, is concentrated in physical infrastructure: data centers, energy, semi-cap equipment, semiconductors, networking, and construction. That is accurate. But parties that concentrate in one room have a specific kind of ending.

The question heading into Q3 earnings is whether the sector can deliver numbers that justify a record 19.7% index weight. If it does, the trend extends. If it does not, the unwinding touches every passive investor who assumed they owned the whole market.

For informational and educational purposes only. Not investment advice. Trading involves risk, including loss of principal.