Jun 11, 2026
Many investors buy ETFs because they want diversification. That is usually a smart starting point: ETFs are low-cost, simple to use, and one fund can give an investor exposure to hundreds of companies.
But in today's AI-led market, more ETFs do not automatically mean more diversification. Several funds with different labels can still point back to the same underlying stocks.
A portfolio can look diversified on the surface while being heavily exposed to a small group of companies underneath. This is especially true when broad market funds, Nasdaq funds, technology funds, semiconductor funds, and AI-themed funds all hold some of the same large stocks.
AI may well be one of the most important technology shifts of this generation. The issue is not owning AI exposure. The issue is unknowingly owning more of it than intended - often far more than the fund labels suggest.
ETF names can make a portfolio feel more balanced than it really is. A broad market ETF, a Nasdaq ETF, a technology ETF, and a semiconductor ETF all sound like different allocations. But recent public holdings data show how the same names can appear again and again.
| ETF / fund | Investor intent | Repeated holdings examples | Why overlap matters |
|---|---|---|---|
| VOO / S&P 500 | Broad U.S. market | Nvidia 7.84%; Apple 6.44%; Microsoft 4.89%; Amazon 4.19%; Alphabet ~3.62% | Even a broad fund is heavily influenced by the largest companies. |
| QQQ / Nasdaq 100 | Growth / innovation | Nvidia 8.55%; Apple 7.41%; Microsoft 5.09%; Amazon 4.67%; Micron 3.81%; AMD 3.44%; Broadcom 3.21%; Meta 2.94% | Many mega-cap names already sit inside broad index funds. |
| XLK / Technology | Tech sector exposure | Nvidia 13.56%; Apple 11.18%; Microsoft 8.50%; Micron 7.06%; Broadcom 5.42%; AMD 5.04% | This can add another layer of exposure to the same technology leaders. |
| SMH / Semiconductor | AI chip exposure | Nvidia 15.94%; TSMC 9.62%; Micron 8.17%; AMD 7.41%; Broadcom 7.30% | A semiconductor fund can deepen exposure to the same AI infrastructure theme. |
Source note: Holdings are examples from public fund pages and third-party holdings pages around late May / June 1, 2026. Holdings change over time and should not be treated as fixed weights.
The AI boom has encouraged a very understandable pattern. An investor starts with an S&P 500 ETF for core exposure, then adds QQQ for more growth. A technology ETF follows, based on the belief that software, cloud computing, and digital infrastructure will keep compounding. Finally comes a semiconductor ETF, because Nvidia, AMD, Broadcom, Micron, and Taiwan Semiconductor sit at the center of the AI buildout.
None of those choices is irrational on its own. The problem appears when they are combined without checking the look-through exposure.
The investor may feel they have built four different portfolio sleeves. Under the surface, however, the portfolio may be much narrower: U.S. large-cap growth, AI infrastructure, semiconductors, cloud platforms, and mega-cap technology.
Consider a simplified example. Sarah is a long-term investor. She does not think of herself as a trader, and she is not trying to make a large single-stock bet. Her portfolio looks diversified at first glance:
| Sleeve | Allocation | Nvidia weight in sleeve | Portfolio contribution |
|---|---|---|---|
| S&P 500 ETF | 50% | 7.84% | 3.92% |
| Nasdaq 100 ETF | 20% | 8.55% | 1.71% |
| Technology ETF | 15% | 13.56% | 2.03% |
| Semiconductor ETF | 10% | 15.94% | 1.59% |
| Direct Nvidia shares | 5% | 100.00% | 5.00% |
| Estimated total Nvidia exposure | 100% | Look-through total | 14.25% |
Source note: This is an illustrative calculation using recent public holdings weights. ETF holdings move over time, so the exact number will change. The point is the mechanism: the same stock can appear inside several funds plus a direct position.
Sarah thinks she owns a 5% Nvidia position. Counting what sits inside her ETFs, the real figure is closer to 14% of the portfolio. Her allocation may still be appropriate - but the risk is larger than her ticker list suggests.
The same exercise can be repeated for Apple, Microsoft, Amazon, Alphabet, Meta, Broadcom, AMD, and Micron. Some of these companies may appear inside broad market ETFs, Nasdaq ETFs, sector ETFs, factor ETFs, and thematic AI funds. The exposure is not always obvious from the ETF names.
Another reason this risk is easy to miss is that not every AI-linked company is classified as "technology." Amazon is usually treated as consumer discretionary. Alphabet and Meta are commonly classified under communication services. Nvidia, Apple, Microsoft, Broadcom, AMD, and Micron sit closer to the technology and semiconductor side of the market.
So an investor can look at a sector chart and still underestimate how much of the portfolio depends on the same broad theme: AI demand, cloud spending, data centers, chips, digital advertising, and mega-cap platform businesses.
ETF overlap is really an exposure problem, not just a ticker problem.
Most broad index ETFs are market-cap weighted. This means the largest companies receive the largest weights. That structure has real advantages: it is simple, low-cost, and hard for many active strategies to beat over long periods.
But market-cap weighting also means that when a small number of companies become very large, they begin to dominate broad indexes. Recent market data showed the 10 largest S&P 500 companies making up roughly 43% of the index's market value, compared with about 29% in 2020 and 19% in 1990.
None of this proves the market is in a bubble. Many of today's largest companies are highly profitable, cash-generative businesses. The practical takeaway is narrower: a "broad" market ETF may be less balanced than its name implies. If the largest AI-linked companies fall together, the index has fewer offsetting exposures to cushion the move.
Valuation is another reason overlap matters right now. A high valuation is a poor crash-timing tool and says nothing about next week. What it changes is the risk/reward balance: when investors pay a high price for future earnings, there is less room for disappointment.
One useful long-term valuation measure is the Shiller P/E ratio, also called the CAPE ratio. It compares stock prices with inflation-adjusted average earnings over the prior 10 years. Because it smooths one-year earnings swings, it can give investors a broader sense of whether the overall market is cheap, normal, or expensive compared with history.
As of May 2026, several public data sources put the S&P 500 CAPE ratio in the 40 to 43 range, depending on the source and date - well above its long-run average near 17, and within reach of the late-1999 dot-com peak. A reading that high will not tell anyone when a correction starts, but it does signal that a great deal of good news is already priced in.
That matters for ETF overlap because expensive markets and concentrated markets can reinforce each other. If the same AI and mega-cap technology companies carry a large share of index performance, and those companies are already priced for strong growth, a disappointment in earnings, margins, AI spending, or interest rates can affect several "different" ETFs at once.
| Valuation / risk measure | Recent or historical context | What it means for ETF investors |
|---|---|---|
| Shiller P/E / CAPE | Recent May 2026 readings were roughly in the 40-43 range, depending on source and date. | The U.S. market is expensive compared with most of its history, so future returns may be less forgiving if expectations disappoint. |
| Long-run CAPE average | Long-run mean is around 17 and median around 16 in commonly cited historical datasets. | Current valuation levels are far above normal historical levels, even if they are not a precise market-timing signal. |
| Dot-com comparison | The CAPE ratio peaked around 44 in late 1999, close to today's range. | The comparison is not perfect, but it is a useful reminder that real technology revolutions can still become risky when valuations and concentration rise together. |
| Concentration + valuation | The largest S&P 500 companies now represent an unusually large share of the index. | When the same expensive mega-cap stocks appear across several ETFs, downside can be more correlated than investors expect. |
Source note: CAPE readings vary slightly by provider and date. The table uses rounded public readings from May 2026 to make the valuation point without implying false precision. CAPE is a long-term valuation indicator, not a short-term crash forecast.
| Market example | What happened | Why it matters for ETF overlap |
|---|---|---|
| S&P 500 concentration in 2026 | The 10 largest companies were reported to represent about 43.2% of the S&P 500's market value. | A broad index can still depend heavily on a small group of companies. |
| 2022 growth and tech selloff | QQQ lost about 32.6% in 2022 as rates rose and growth valuations compressed. | Different growth-oriented funds can fall together when they share the same style and holdings. |
| Dot-com bubble, 2000-2002 | The Nasdaq peaked at 5,048 in March 2000 and fell to 1,139.90 by October 2002, a decline of about 77%. | A technology revolution can be real while stock prices still go through painful drawdowns. |
Source note: Historical returns and market concentration figures are included for context only. Past performance does not predict future results.
ETF overlap often feels harmless during a bull market. In fact, it can feel like a benefit. If Nvidia, Microsoft, Apple, Amazon, Alphabet, Meta, Broadcom, and other mega-cap technology names are rising, a portfolio with repeated exposure to those companies may perform very well.
But strong performance can hide concentration risk. The real test comes when the same stocks begin to fall together.
We saw a version of this in 2022. Many growth and technology-heavy portfolios suffered large drawdowns as interest rates rose and investors became less willing to pay high valuations for future growth. QQQ lost about 32.6% that year. For a deeper example of how that period affected concentrated growth portfolios, see Guardfolio's breakdown of the 2022 market crash.
The dot-com bubble offers a longer-term warning. The internet was real and went on to reshape the world, yet investors who were overexposed to technology stocks at extreme valuations still took severe losses. The Nasdaq fell about 77% from its 2000 peak to its 2002 low.
AI is not a rerun of the dot-com bubble. Today's leaders are far stronger businesses than most 2000-era internet names. The harder truth is simpler: even a genuine technology shift can turn into a dangerous exposure when too much capital rides on the same handful of winners continuing to win.
Investors do not need to avoid ETFs. They simply need to inspect them properly.
A useful overlap review should answer a few practical questions:
Which companies do I own across all ETFs and accounts?
What is my total exposure to my top 10 underlying holdings?
How much of my portfolio is tied to AI, technology, semiconductors, or U.S. mega-cap growth?
Do my ETFs actually behave differently, or do they mostly rise and fall together?
What would happen if Nvidia, Apple, Microsoft, Amazon, Alphabet, Meta, and Broadcom corrected at the same time?
Investors can start by checking fund holdings manually, but it becomes difficult once a portfolio includes several ETFs, direct stocks, and multiple accounts. An ETF overlap checker can help reveal when different funds are quietly holding the same companies and increasing duplicate exposure.
The key is to move from ticker-level thinking to exposure-level thinking. Ticker-level thinking says: "I own VOO, QQQ, XLK, and SMH." Exposure-level thinking asks: "How much do I actually own of Nvidia, Apple, Microsoft, Amazon, Alphabet, Broadcom, AMD, Micron, semiconductors, AI infrastructure, and U.S. mega-cap growth?"
This is where portfolio risk management becomes useful. Instead of only looking at fund tickers, investors can look through ETFs, combine exposure across accounts, and monitor concentration, overlap, sector exposure, allocation drift, drawdowns, and correlation across the whole portfolio.
This is not an argument against technology or AI. Some investors will deliberately want more exposure to AI, semiconductors, or large-cap growth, and that can be a valid choice.
The aim is simply to make the exposure visible. Once investors can see what they actually own beneath the tickers, they can judge whether the concentration fits their goals, time horizon, and risk tolerance.
ETF overlap is one of the easiest risks to miss in modern portfolios. It matters even more now because the AI boom has pushed many investors toward the same group of mega-cap technology stocks, sometimes without them realizing how much exposure they already have.
A portfolio with five ETFs can still be concentrated; a portfolio with two can be well diversified. What decides it is not the number of tickers - it is what sits underneath them.
The lesson from 2022 is that growth and technology exposure can fall quickly when market conditions change. The lesson from 2000 is that even a real technology revolution can still become a painful investment cycle if valuations and concentration go too far.
Before adding another ETF, investors should ask one simple question:
Am I truly diversifying, or am I buying the same exposure again with a different label?
That question can help investors avoid hidden concentration risk before the next market correction, not after it.
Disclaimer: This article is for educational purposes only and should not be considered investment, financial, tax, or legal advice. Investors should do their own research or consult a qualified professional before making investment decisions.
Vanguard S&P 500 ETF / VOO fund profile: https://investor.vanguard.com/investment-products/etfs/profile/voo
VOO holdings snapshot used for table examples: https://stockanalysis.com/etf/voo/holdings/
Invesco QQQ holdings and risk language: https://www.invesco.com/qqq-etf/en/about.html
State Street XLK holdings as of June 1, 2026: https://www.ssga.com/us/en/intermediary/etfs/state-street-technology-select-sector-spdr-etf-xlk
VanEck SMH holdings as of June 1, 2026: https://www.vaneck.com/us/en/investments/semiconductor-etf-smh/
QQQ 2022 annual return reference: https://finance.yahoo.com/quote/QQQ/performance/
Goldman Sachs dot-com bubble history: https://www.goldmansachs.com/our-firm/history/moments/2000-dot-com-bubble
S&P 500 concentration context: https://www.wsj.com/livecoverage/stock-market-today-dow-sp-500-nasdaq-06-01-2026/card/one-thing-i-m-watching-the-ai-fication-of-the-s-p-500-1vfPrCA1xUjyZJjo2YjU
Robert Shiller / Yale online stock market data and CAPE ratio: https://www.econ.yale.edu/~shiller/data.htm
Multpl Shiller PE Ratio current and historical mean/median/max: https://www.multpl.com/shiller-pe
Many investors buy ETFs because they want diversification. That is usually a smart starting point: ETFs are low-cost, simple to use, and one fund can give an investor exposure to hundreds of companies.
But in today's AI-led market, more ETFs do not automatically mean more diversification. Several funds with different labels can still point back to the same underlying stocks.
A portfolio can look diversified on the surface while being heavily exposed to a small group of companies underneath. This is especially true when broad market funds, Nasdaq funds, technology funds, semiconductor funds, and AI-themed funds all hold some of the same large stocks.
AI may well be one of the most important technology shifts of this generation. The issue is not owning AI exposure. The issue is unknowingly owning more of it than intended - often far more than the fund labels suggest.
ETF names can make a portfolio feel more balanced than it really is. A broad market ETF, a Nasdaq ETF, a technology ETF, and a semiconductor ETF all sound like different allocations. But recent public holdings data show how the same names can appear again and again.
| ETF / fund | Investor intent | Repeated holdings examples | Why overlap matters |
|---|---|---|---|
| VOO / S&P 500 | Broad U.S. market | Nvidia 7.84%; Apple 6.44%; Microsoft 4.89%; Amazon 4.19%; Alphabet ~3.62% | Even a broad fund is heavily influenced by the largest companies. |
| QQQ / Nasdaq 100 | Growth / innovation | Nvidia 8.55%; Apple 7.41%; Microsoft 5.09%; Amazon 4.67%; Micron 3.81%; AMD 3.44%; Broadcom 3.21%; Meta 2.94% | Many mega-cap names already sit inside broad index funds. |
| XLK / Technology | Tech sector exposure | Nvidia 13.56%; Apple 11.18%; Microsoft 8.50%; Micron 7.06%; Broadcom 5.42%; AMD 5.04% | This can add another layer of exposure to the same technology leaders. |
| SMH / Semiconductor | AI chip exposure | Nvidia 15.94%; TSMC 9.62%; Micron 8.17%; AMD 7.41%; Broadcom 7.30% | A semiconductor fund can deepen exposure to the same AI infrastructure theme. |
Source note: Holdings are examples from public fund pages and third-party holdings pages around late May / June 1, 2026. Holdings change over time and should not be treated as fixed weights.
The AI boom has encouraged a very understandable pattern. An investor starts with an S&P 500 ETF for core exposure, then adds QQQ for more growth. A technology ETF follows, based on the belief that software, cloud computing, and digital infrastructure will keep compounding. Finally comes a semiconductor ETF, because Nvidia, AMD, Broadcom, Micron, and Taiwan Semiconductor sit at the center of the AI buildout.
None of those choices is irrational on its own. The problem appears when they are combined without checking the look-through exposure.
The investor may feel they have built four different portfolio sleeves. Under the surface, however, the portfolio may be much narrower: U.S. large-cap growth, AI infrastructure, semiconductors, cloud platforms, and mega-cap technology.
Consider a simplified example. Sarah is a long-term investor. She does not think of herself as a trader, and she is not trying to make a large single-stock bet. Her portfolio looks diversified at first glance:
| Sleeve | Allocation | Nvidia weight in sleeve | Portfolio contribution |
|---|---|---|---|
| S&P 500 ETF | 50% | 7.84% | 3.92% |
| Nasdaq 100 ETF | 20% | 8.55% | 1.71% |
| Technology ETF | 15% | 13.56% | 2.03% |
| Semiconductor ETF | 10% | 15.94% | 1.59% |
| Direct Nvidia shares | 5% | 100.00% | 5.00% |
| Estimated total Nvidia exposure | 100% | Look-through total | 14.25% |
Source note: This is an illustrative calculation using recent public holdings weights. ETF holdings move over time, so the exact number will change. The point is the mechanism: the same stock can appear inside several funds plus a direct position.
Sarah thinks she owns a 5% Nvidia position. Counting what sits inside her ETFs, the real figure is closer to 14% of the portfolio. Her allocation may still be appropriate - but the risk is larger than her ticker list suggests.
The same exercise can be repeated for Apple, Microsoft, Amazon, Alphabet, Meta, Broadcom, AMD, and Micron. Some of these companies may appear inside broad market ETFs, Nasdaq ETFs, sector ETFs, factor ETFs, and thematic AI funds. The exposure is not always obvious from the ETF names.
Another reason this risk is easy to miss is that not every AI-linked company is classified as "technology." Amazon is usually treated as consumer discretionary. Alphabet and Meta are commonly classified under communication services. Nvidia, Apple, Microsoft, Broadcom, AMD, and Micron sit closer to the technology and semiconductor side of the market.
So an investor can look at a sector chart and still underestimate how much of the portfolio depends on the same broad theme: AI demand, cloud spending, data centers, chips, digital advertising, and mega-cap platform businesses.
ETF overlap is really an exposure problem, not just a ticker problem.
Most broad index ETFs are market-cap weighted. This means the largest companies receive the largest weights. That structure has real advantages: it is simple, low-cost, and hard for many active strategies to beat over long periods.
But market-cap weighting also means that when a small number of companies become very large, they begin to dominate broad indexes. Recent market data showed the 10 largest S&P 500 companies making up roughly 43% of the index's market value, compared with about 29% in 2020 and 19% in 1990.
None of this proves the market is in a bubble. Many of today's largest companies are highly profitable, cash-generative businesses. The practical takeaway is narrower: a "broad" market ETF may be less balanced than its name implies. If the largest AI-linked companies fall together, the index has fewer offsetting exposures to cushion the move.
Valuation is another reason overlap matters right now. A high valuation is a poor crash-timing tool and says nothing about next week. What it changes is the risk/reward balance: when investors pay a high price for future earnings, there is less room for disappointment.
One useful long-term valuation measure is the Shiller P/E ratio, also called the CAPE ratio. It compares stock prices with inflation-adjusted average earnings over the prior 10 years. Because it smooths one-year earnings swings, it can give investors a broader sense of whether the overall market is cheap, normal, or expensive compared with history.
As of May 2026, several public data sources put the S&P 500 CAPE ratio in the 40 to 43 range, depending on the source and date - well above its long-run average near 17, and within reach of the late-1999 dot-com peak. A reading that high will not tell anyone when a correction starts, but it does signal that a great deal of good news is already priced in.
That matters for ETF overlap because expensive markets and concentrated markets can reinforce each other. If the same AI and mega-cap technology companies carry a large share of index performance, and those companies are already priced for strong growth, a disappointment in earnings, margins, AI spending, or interest rates can affect several "different" ETFs at once.
| Valuation / risk measure | Recent or historical context | What it means for ETF investors |
|---|---|---|
| Shiller P/E / CAPE | Recent May 2026 readings were roughly in the 40-43 range, depending on source and date. | The U.S. market is expensive compared with most of its history, so future returns may be less forgiving if expectations disappoint. |
| Long-run CAPE average | Long-run mean is around 17 and median around 16 in commonly cited historical datasets. | Current valuation levels are far above normal historical levels, even if they are not a precise market-timing signal. |
| Dot-com comparison | The CAPE ratio peaked around 44 in late 1999, close to today's range. | The comparison is not perfect, but it is a useful reminder that real technology revolutions can still become risky when valuations and concentration rise together. |
| Concentration + valuation | The largest S&P 500 companies now represent an unusually large share of the index. | When the same expensive mega-cap stocks appear across several ETFs, downside can be more correlated than investors expect. |
Source note: CAPE readings vary slightly by provider and date. The table uses rounded public readings from May 2026 to make the valuation point without implying false precision. CAPE is a long-term valuation indicator, not a short-term crash forecast.
| Market example | What happened | Why it matters for ETF overlap |
|---|---|---|
| S&P 500 concentration in 2026 | The 10 largest companies were reported to represent about 43.2% of the S&P 500's market value. | A broad index can still depend heavily on a small group of companies. |
| 2022 growth and tech selloff | QQQ lost about 32.6% in 2022 as rates rose and growth valuations compressed. | Different growth-oriented funds can fall together when they share the same style and holdings. |
| Dot-com bubble, 2000-2002 | The Nasdaq peaked at 5,048 in March 2000 and fell to 1,139.90 by October 2002, a decline of about 77%. | A technology revolution can be real while stock prices still go through painful drawdowns. |
Source note: Historical returns and market concentration figures are included for context only. Past performance does not predict future results.
ETF overlap often feels harmless during a bull market. In fact, it can feel like a benefit. If Nvidia, Microsoft, Apple, Amazon, Alphabet, Meta, Broadcom, and other mega-cap technology names are rising, a portfolio with repeated exposure to those companies may perform very well.
But strong performance can hide concentration risk. The real test comes when the same stocks begin to fall together.
We saw a version of this in 2022. Many growth and technology-heavy portfolios suffered large drawdowns as interest rates rose and investors became less willing to pay high valuations for future growth. QQQ lost about 32.6% that year. For a deeper example of how that period affected concentrated growth portfolios, see Guardfolio's breakdown of the 2022 market crash.
The dot-com bubble offers a longer-term warning. The internet was real and went on to reshape the world, yet investors who were overexposed to technology stocks at extreme valuations still took severe losses. The Nasdaq fell about 77% from its 2000 peak to its 2002 low.
AI is not a rerun of the dot-com bubble. Today's leaders are far stronger businesses than most 2000-era internet names. The harder truth is simpler: even a genuine technology shift can turn into a dangerous exposure when too much capital rides on the same handful of winners continuing to win.
Investors do not need to avoid ETFs. They simply need to inspect them properly.
A useful overlap review should answer a few practical questions:
Which companies do I own across all ETFs and accounts?
What is my total exposure to my top 10 underlying holdings?
How much of my portfolio is tied to AI, technology, semiconductors, or U.S. mega-cap growth?
Do my ETFs actually behave differently, or do they mostly rise and fall together?
What would happen if Nvidia, Apple, Microsoft, Amazon, Alphabet, Meta, and Broadcom corrected at the same time?
Investors can start by checking fund holdings manually, but it becomes difficult once a portfolio includes several ETFs, direct stocks, and multiple accounts. An ETF overlap checker can help reveal when different funds are quietly holding the same companies and increasing duplicate exposure.
The key is to move from ticker-level thinking to exposure-level thinking. Ticker-level thinking says: "I own VOO, QQQ, XLK, and SMH." Exposure-level thinking asks: "How much do I actually own of Nvidia, Apple, Microsoft, Amazon, Alphabet, Broadcom, AMD, Micron, semiconductors, AI infrastructure, and U.S. mega-cap growth?"
This is where portfolio risk management becomes useful. Instead of only looking at fund tickers, investors can look through ETFs, combine exposure across accounts, and monitor concentration, overlap, sector exposure, allocation drift, drawdowns, and correlation across the whole portfolio.
This is not an argument against technology or AI. Some investors will deliberately want more exposure to AI, semiconductors, or large-cap growth, and that can be a valid choice.
The aim is simply to make the exposure visible. Once investors can see what they actually own beneath the tickers, they can judge whether the concentration fits their goals, time horizon, and risk tolerance.
ETF overlap is one of the easiest risks to miss in modern portfolios. It matters even more now because the AI boom has pushed many investors toward the same group of mega-cap technology stocks, sometimes without them realizing how much exposure they already have.
A portfolio with five ETFs can still be concentrated; a portfolio with two can be well diversified. What decides it is not the number of tickers - it is what sits underneath them.
The lesson from 2022 is that growth and technology exposure can fall quickly when market conditions change. The lesson from 2000 is that even a real technology revolution can still become a painful investment cycle if valuations and concentration go too far.
Before adding another ETF, investors should ask one simple question:
Am I truly diversifying, or am I buying the same exposure again with a different label?
That question can help investors avoid hidden concentration risk before the next market correction, not after it.
Disclaimer: This article is for educational purposes only and should not be considered investment, financial, tax, or legal advice. Investors should do their own research or consult a qualified professional before making investment decisions.
Vanguard S&P 500 ETF / VOO fund profile: https://investor.vanguard.com/investment-products/etfs/profile/voo
VOO holdings snapshot used for table examples: https://stockanalysis.com/etf/voo/holdings/
Invesco QQQ holdings and risk language: https://www.invesco.com/qqq-etf/en/about.html
State Street XLK holdings as of June 1, 2026: https://www.ssga.com/us/en/intermediary/etfs/state-street-technology-select-sector-spdr-etf-xlk
VanEck SMH holdings as of June 1, 2026: https://www.vaneck.com/us/en/investments/semiconductor-etf-smh/
QQQ 2022 annual return reference: https://finance.yahoo.com/quote/QQQ/performance/
Goldman Sachs dot-com bubble history: https://www.goldmansachs.com/our-firm/history/moments/2000-dot-com-bubble
S&P 500 concentration context: https://www.wsj.com/livecoverage/stock-market-today-dow-sp-500-nasdaq-06-01-2026/card/one-thing-i-m-watching-the-ai-fication-of-the-s-p-500-1vfPrCA1xUjyZJjo2YjU
Robert Shiller / Yale online stock market data and CAPE ratio: https://www.econ.yale.edu/~shiller/data.htm
Multpl Shiller PE Ratio current and historical mean/median/max: https://www.multpl.com/shiller-pe