Published on: 2026-08-19
Updated on: 2026-08-19
Portfolios built around the artificial intelligence rally develop a structural imbalance: the holdings that produced the strongest returns become the largest source of concentration risk. When AI stocks fall as a group, taking semiconductors, data-centre infrastructure and mega-cap growth down together, diversification becomes an allocation decision rather than a theoretical one.
Several ETFs reduce technology exposure across six routes: defensive sectors, valuation factors, concentration, volatility control, non-equity assets and direct hedging. Some change which stocks a portfolio owns. Others reduce how much equity risk it carries.
XLP and XLV cut dependence on AI-related earnings but stay fully exposed to equity risk.
VTV, QUAL and COWZ address valuation and business quality, not a technology hedge.
RSP, USMV and JEPI change concentration, volatility or return structure within equities.
BIL and GLD move outside equities, while SQQQ is a leveraged directional trade.
An AI stock correction can originate from three places, and the right ETF depends on which is dominant.
Semiconductors, data-centre infrastructure and mega-cap growth reprice sharply while the wider economy holds up. Sector rotation is the mechanism, not economic weakness.
Earnings hold up, but investors stop paying premium multiples for growth expected far into the future. The de-rating hits the most expensive names hardest.
Technology weakness spreads into a general equity selloff, often alongside higher yields, tighter credit or slowing growth.
Grouped by the job each fund performs rather than ranked best to worst.
| ETF | Approach | Main risk it reduces | Gross expense ratio |
|---|---|---|---|
| XLP | Defensive sector | AI earnings dependence | 0.08% |
| XLV | Defensive sector | Tech earnings dependence | 0.08% |
| VTV | Value | Growth valuation compression | 0.03% |
| QUAL | Quality | Weak earnings and leverage | 0.15% |
| COWZ | Free cash flow | Cash-flow valuation risk | 0.49% |
| RSP | Equal weight | Mega-cap concentration | 0.20% |
| USMV | Minimum volatility | Equity volatility | 0.15% |
| JEPI | Equity plus options | Volatility | 0.35% |
| BIL | Short Treasuries | Total equity exposure | 0.14% |
| GLD | Gold | Earnings dependence | 0.40% |
| XLU | Utilities | Cyclical earnings risk | 0.08% |
| SQQQ | Inverse leverage | Nasdaq decline | 0.99% |
Gross ratios exclude fee waivers and can change. Verify with the issuer.
XLP shifts exposure toward recurring consumer demand: household products, beverages and food retail. XLV moves it toward pharmaceuticals, medical devices, insurers and healthcare services. Neither depends on data-centre investment or semiconductor demand holding its present pace.
The limitation is structural. They are still equities. A broad liquidation pulls defensive share prices down with the wider market, and healthcare adds regulatory, reimbursement and pipeline risks unrelated to technology.
Utilities earn a defensive reputation from stable electricity and water demand. Their share prices are a separate question. These are capital-intensive businesses carrying substantial debt, so a correction driven by rising yields can pressure XLU at the same time.
Data-centre expansion has also tied part of the utility growth story to AI electricity demand. XLU therefore behaves differently depending on whether AI stocks are falling on valuation, on rates, or on doubts about infrastructure spending.
These three become relevant when the weakness comes from valuation rather than a collapse in AI earnings.
VTV tilts toward large-cap value, mechanically cutting weight in the highest-multiple names. COWZ ranks the Russell 1000 by free cash flow yield and weights the top hundred by the cash they generate, with financials excluded by design.
QUAL screens for profitability, earnings stability and lower leverage, though it is the partial exception here: its index is sector-neutral to a technology-heavy parent, so it improves business quality without materially reducing technology weight or the multiple paid for it.
VTV and COWZ therefore reduce reliance on the market paying premium prices for long-duration growth, while QUAL reduces reliance on those earnings proving fragile. Neither job is the same as safety, and factor strategies lag whenever growth leadership resumes.
Market-cap weighting carries a built-in consequence: the biggest winners automatically become the largest positions. When a small group of mega-cap stocks leads the index, their influence over returns rises with their market value.
RSP equal-weights the same constituent list, limiting the influence of any single mega-cap. Our RSP vs SPY comparison covers how each structure behaves across market regimes.
The purpose is narrow. RSP addresses dependence on a handful of dominant stocks, not equity-market risk in aggregate.
USMV runs a minimum-volatility optimisation across US stocks, then pulls sector weights back toward its parent index. Because that parent is dominated by mega-cap technology, the optimisation cannot fully escape it. A minimum-volatility label is not shorthand for a technology-free portfolio.
JEPI pairs an actively managed, lower-beta equity portfolio with an options-linked income sleeve, targeting a smoother ride while surrendering part of the upside.
Both alter how an equity portfolio behaves. Neither removes equity exposure.
Everything above sits inside the equity market. BIL does not. It holds very short-dated US Treasury bills, so its price barely responds to equity swings and its return tracks short-term rates. Other short-duration Treasury funds carry different rate sensitivities.
Moving from a Nasdaq-heavy fund into XLP changes the type of equities held. Moving part of the allocation into BIL reduces the quantity of equity risk itself.
GLD replaces dependence on corporate earnings with exposure to gold. Gold pays no yield, so rising real interest rates raise the cost of holding it, and it can fall exactly when investors expect a defensive asset to hold up.
SQQQ serves a different purpose from the other eleven. It seeks leveraged inverse exposure to the Nasdaq-100 on a daily basis, and that reset means longer holding-period returns depend on the sequence of market moves, not only the final direction of the index.
In a market that repeatedly falls and recovers, compounding can produce an outcome far removed from the index return multiplied by the stated leverage, even when the directional call proves correct. XLP, VTV, RSP and BIL change portfolio exposure. SQQQ expresses a view.
The twelve are not interchangeable, and treating them as one category of protection is the common mistake. A defensive label describes what a fund holds, not which decline it survives.
Eight of the twelve change the composition of equity risk without changing its quantity. BIL and GLD reduce how much is carried at all. SQQQ profits from a fall, making it a trade rather than an allocation. Choosing starts with the reason AI stocks are falling, not the label on the fund.
No. Adding a fund with different earnings drivers lowers the share of a portfolio tied to AI while the original holding stays in place. Selling removes the exposure outright, with different timing and cost consequences.
Several do. RSP holds every S&P 500 constituent, and USMV and QUAL are sector-neutral to a technology-heavy parent index. Only XLP, XLV and XLU exclude the sector by construction, and BIL and GLD are not equities.
No. Short-dated funds such as BIL hold their value because their price barely responds to rate moves. Longer-dated bond ETFs can fall alongside equities when a selloff is driven by rising yields rather than slowing growth.
There is no fixed limit, but the tracking objective applies to a single session. Across weeks in a volatile market, the compounded return can diverge sharply from the index move. Issuers describe these as short-term instruments.
Rate-driven declines are hardest to offset with equities. Long-duration growth, utilities and gold can come under pressure at once, since higher yields lift discount rates and the cost of holding non-yielding assets. Short-dated Treasury exposure is least affected.
EBC’s ETF range spans defensive sectors, value and short-duration bond exposure. Compare what is available before deciding which route fits your reading of the risk.
