Buying an S&P 500 index fund can feel like a straightforward way to diversify. After all, many American workers dollar-cost average into this sort of fund with every paycheck via a 401(k) or IRA without even thinking about it.
Steve Sosnick, chief strategist at Interactive Brokers, argues that investors should look more closely before assuming that broad-market exposure offers enough diversification to offset a portfolio weighed down by technology holdings trading near all-time highs.
In his view, an S&P 500 fund or a similar broad-market index fund can leave an investor exposed to the AI trade by about 40% to 45%. Adding individual AI-related stocks or semiconductor names on top can concentrate a portfolio far more than its owner realizes.
Why is the S&P 500 so AI-heavy right now?
This silent-but-potentially-risky overweighting occurs because the S&P 500 — and the popular index funds that track it — are weighted by market capitalization, which means larger companies (by market value) make up more of the index.
In fact, as of early September, the top seven S&P 500 companies made up over 34% of the index’s value. What do they all have in common? They’re all heavily involved in the AI boom:
- Nvidia: (NVDA)
- Microsoft: (MSFT)
- Apple: (AAPL)
- Amazon: (AMZN)
- Meta Platforms: (META)
- Broadcom: (AVGO)
- Alphabet Class A: (GOOGL)
- Alphabet Class C: (GOOG)
That does not mean an S&P 500 fund is inherently unsuitable, nor does it mean technology cannot keep rising. Sosnick’s point is narrower and more practical: Investors should measure the overlap between a broad index fund and their other holdings, then decide whether that combined exposure fits their risk tolerance, especially while long-term bond yields are rising.
The immediate takeaway, according to Sosnick, is to review positions by their shared economic exposure rather than by the number of ticker symbols in the account.
The distinction matters because a successful buy-the-dip strategy can also become a habit that ignores changes in business fundamentals, valuation, and financing conditions. Sosnick sees signs that those conditions deserve more attention now.
Also read: Dow Jones vs. S&P 500: Which index actually represents the market?
Why an S&P 500 fund can overlap with AI stocks
Diversification means spreading risk across investments that do not all depend on the same outcome. A portfolio can look diversified on paper because it owns an index fund, individual stocks, and perhaps a sector fund. It may still be making one large bet if many of those holdings depend on continued enthusiasm for AI spending, semiconductor demand, or the largest technology companies.
Sosnick’s concern begins with market capitalization weighting, the approach used by many S&P 500 index funds. In a market-cap-weighted fund, the largest companies receive the largest allocations. That structure can be useful for investors seeking broad index exposure, but it also means that large technology companies can have an outsized influence on the fund’s returns.
“Even if you’re putting money into an S&P 500 mutual fund or index fund, which is pretty generic, whether you’re doing ETFs or mutual funds, et cetera, you’re about 40% or 45% exposed to the AI trade.”
—Steve Sosnick, when asked whether taking risk off the table means rotating out of big technology stocks
Sosnick’s estimate is a warning about portfolio construction, not a forecast for the next market move. An investor who owns a broad S&P 500 fund, NVIDIA, Micron Technology, and a semiconductor-focused leveraged ETF may own several different securities, but each can be sensitive to a reversal in the same AI trade.
The correct question is not simply whether each holding has performed well. It is whether all of them could decline for the same reason.
TheStreet
How rising long-term bond yields pressure stock valuations
The second part of Sosnick’s caution is interest rates. Long-term bond yields are the returns investors can receive from lending money over longer periods. When those yields rise, investors either demand a higher expected return from stocks, or reduce exposure to stocks altogether. That can pressure stock valuations, particularly for companies whose expected cash flows lie further in the future.
Sosnick describes the basic valuation logic in plain terms: A stock can be viewed as the present value of its future cash flows or earnings. A higher interest rate reduces that present value because future dollars are discounted more heavily.
The mechanism does not determine the price of every stock on every day, but it explains why a sustained rise in long-term rates can be difficult for richly valued growth stocks.
“When long rates go up, that pressures valuations, if you’re actually thinking fundamentally. A stock in theory is the present value of its future cash flows or earnings, depending on how you wanna do the model. But the higher the interest rate, the lower the present value, the more you have to discount those future earnings.”
—Steve Sosnick, when asked what being nimble in the current market looks like
Sosnick pointed to the 10-year Treasury yield near 5% and the 30-year yield above 5% as important market thresholds. He said investors may have to rethink assumptions if the 10-year moves through 5%, because fundamental valuation models would need to be reconsidered in a more persistently high-rate environment. He did not say that a move through that level guarantees a selloff, and he also noted that it is possible that bond traders could halt at that level.
For a long-term, buy-and-hold investor, the useful takeaway is not to attempt to predict every shift in bond yields. It is to recognize that an investment portfolio heavily tilted toward the AI trade may have two related vulnerabilities: a technology disappointment and a higher discount rate applied to future earnings. Reviewing those exposures together is more useful than treating them as separate risks.
Related: S&P 500 investors are quietly making a huge shift
How to tell differentiate a buyable dip from a deteriorating business
Sosnick does not reject buying dips. He says Interactive Brokers customers have often been successful with the approach over the past 10 to 15 years. His warning is that a strategy with a long record of working can become reflexive. A price decline is not, by itself, evidence that a stock has become a bargain.
The first test is whether the business fundamentals have changed markedly. Fundamentals are the conditions that support a company’s business, such as its earnings outlook, competitive position, and ability to generate cash.
If a stock falls because of a short-lived bout of market nervousness while the underlying business remains sound, Sosnick sees a potential opportunity. If the business outlook has worsened, on the other hand, the lower price may reflect a legitimate reason for investors to sell.
“From a longer-term point of view, I think you want to look at situations where the market might really dislike something, but the fundamentals still remain solid. On the other hand, if those fundamentals appear to have changed, that’s not a buying opportunity. There’s a reason why people are selling.”
—Steve Sosnick, when asked how investors can distinguish a buying opportunity from a dip to avoid
His Salesforce example illustrates the difference between changing sentiment and a broken business. Sosnick said investors moved from loving Salesforce (and software as a service at large) to hating them, while neither extreme sentiment necessarily described Salesforce — as a company — accurately.
He also cited Microsoft, where investors focused on companies receiving Microsoft’s spending, including semiconductor businesses, rather than on why Microsoft was making the investment. A stock can fall out of favor without its core business becoming worse.
Timing still matters. Sosnick said a trader may act quickly during a short-term dislocation, while a longer selloff can require waiting for the dust to settle. That is an argument for matching an action to an investor’s time horizon. A short-term trader should use a pre-defined plan (including an exit strategy, like a stop-loss order) for a temporary price move. A long-term, buy-and-hold investor needs a clearer view of whether the company’s business case remains intact.
Why having an exit plan matters before buying a stock
Sosnick separates a trade from an investment because each requires a different time frame, even if both involve the same stock. He recalled a friend who bought a stock, watched it rise, and then asked when to sell. Sosnick’s first question was the price target. The friend had not set one.
For investors using trading-style tactics, Sosnick recommends defining a buy level and a sell level before opening a position. He also recommends setting a stop level, a preplanned price at which an investor exits to limit their loss. The maximum acceptable loss belongs to the exit decision: It defines the loss an investor is prepared to accept before selling the position.
Related: Bank of America takes heat for stark S&P 500 call
A long-term, buy-and-hold investor may set wider limits and hold a position for a longer period than an active trader. Sosnick nevertheless argues that investors should keep monitoring investments rather than treating an original thesis as permanent. A preplanned process can prevent the common mistake of buying first and inventing the selling rules only after the price has moved.
This framework also helps explain the activity Sosnick sees among Interactive Brokers’ active customers. He said customers bought Nvidia before earnings and took profits after the stock moved higher. He said Microsoft appeared on the sell side after a 15% pop during the summer. In Sosnick’s view, those customers were treating many individual names as trades, while they treated VOO purchases during significant declines — such as the tariff tantrum and the aftermath of the start of the Iran War — more like investments.
Where value stocks can fit in a concentrated portfolio
Reducing overlap does not require abandoning stocks. Sosnick says investors may want to look beyond the most popular growth names and consider lower-beta companies. (Beta is a measure of how sensitive a stock tends to be to broad market moves, so a lower-beta stock has historically tended to move less than the market.)
He said value stocks have performed well relative to growth stocks over the past year or two, a trend he believes many investors have overlooked. He highlighted dividend-paying companies supported by free cash flow, meaning cash a company has left after running and investing in its business. His preference is for dividends that businesses can afford from that cash rather than dividends supported by borrowing.
Sosnick did not offer buy, sell, or hold recommendations on individual stocks. Instead, he suggested starting with sectors that tend to be more stable and value-oriented, including industrials, basic materials, and consumer staples. He was less favorable toward consumer discretionary companies as a starting point for this screen.
For valuation, he mentioned the PEG ratio, which compares a stock’s P/E ratio with its earnings growth rate. Sosnick said a value investor should not want to overpay for a company with a PEG ratio of one, and should avoid paying a huge premium to own companies whose P/E ratios exceed the market average.
It’s important to note, however, that these are just Sosnick’s screening ideas — not guarantees that a stock is cheap or will outperform.
The takeaway for S&P 500 investors with AI holdings
Sosnick’s market outlook is cautious. He said the S&P 500 may trend sideways to lower, and he called a 10% correction overdue, while declining to predict an immediate 20% bear market. That forecast is an opinion, and he also acknowledged the market’s powerful tendency to attract dip-buying after pullbacks.
The more durable decision procedure does not depend on accepting his forecast. First, list an S&P 500 fund alongside every individual stock and sector fund. Next, identify how much of the portfolio relies on the AI trade, semiconductors, and the largest technology companies. Then decide whether the combined exposure, including any leverage, is appropriate if rates remain elevated or the technology trade becomes more volatile.
Long-term, buy-and-hold investors can use that review to rebalance toward exposures they actually want, rather than assuming the total number of holdings in a broad-market fund equals diversification. Active traders can add defined buy levels, sell levels, and stop-loss levels before placing a trade.
In both cases, the discipline is the same: Distinguish a temporary decline from a change in the underlying business, and keep enough liquidity available for near-term needs.
All this being said, Sosnick still believes a broad index fund can remain a useful core holding. His warning is that an investor should understand the risks already inside that core before adding more of the same theme. When the S&P 500, individual technology holdings, semiconductors, and leveraged products all point toward the AI trade, even a generic-looking portfolio may be less diversified than it appears.