Portfolio Reality Check

Your Index Fund Is an AI Bet

You think you own "the market." You own seven companies and a rounding error. Here's how much of your retirement is riding on AI without your permission.

The Hidden Bet

The Most Popular Investment in America Has a Secret

Roughly $8 trillion sits in S&P 500 index funds. The pitch for these funds is simple: buy the whole market, diversify across 500 companies, and let compound growth do the work. That pitch made sense for decades.

It stopped being accurate around 2023. The S&P 500 is a market-cap weightedCompanies with higher stock market valuations get a larger share of the index. Bigger companies matter more. index. That means the biggest companies get the biggest slice. And the biggest companies right now are all making the same bet.

35.7%
Share of the S&P 500 held by the top 7 stocks
Apple, Microsoft, NVIDIA, Amazon, Alphabet, Meta, and Broadcom. All seven derive a significant portion of their revenue or growth thesis from artificial intelligence. As of February 2026.
Thirty-five point seven percent represents the combined market capitalization weight of the seven largest companies in the S&P 500, all of which have significant AI exposure.

One out of every three dollars in your S&P 500 index fund belongs to seven companies. All seven are AI plays. If AI delivers, your index fund wins. If AI disappoints, your "diversified" portfolio takes a concentrated hit.

This is not diversification. This is a directional bet with extra steps.

The Numbers

How Concentrated Is Your Index Fund?

Here is what you actually own when you buy an S&P 500 index fund. These weightings are from SPDR's published holdings as of early 2026.

Top 10 Holdings in the S&P 500
Percentage of total index value, February 2026
This visualization shows the percentage allocations of the top 10 S&P 500 companies: Apple 7.2%, Microsoft 6.5%, NVIDIA 6.1%, Amazon 5.2%, Alphabet 4.8%, Meta 3.9%, Broadcom 2.8%, Tesla 2.5%, Berkshire 2.2%, Magnificent 7 together accounting for approximately 35% of index value.
Apple
7.2%
Microsoft
6.5%
NVIDIA
5.9%
Amazon
4.2%
Alphabet
4.0%
Meta
2.9%
Broadcom
2.0%
Tesla
1.8%
Eli Lilly
1.6%
JPMorgan
1.5%

The top 10 stocks account for 37.6% of the entire index. The bottom 400 stocks account for about 28%. You own more Apple than you own the smallest 200 companies combined.

And this is the worst the concentration has been since the index began. In 2000, at the peak of the dot-com bubble, the top 10 stocks were 27% of the S&P 500. We are 10 points past that today.

The AI Thread

Every Top Holding Is an AI Bet

It would be one thing if the top seven companies operated in different industries. They don't. Each one has tied its growth story to artificial intelligence.

AI Revenue Exposure by Top S&P 500 Holding
Estimated share of revenue directly tied to AI products, infrastructure, or AI-driven ad targeting
This table shows estimated AI revenue exposure for major S&P 500 companies, ranging from Microsoft at approximately 30% AI revenue to Apple at approximately 5% AI revenue.
Company S&P Weight Primary AI Role AI Revenue Share
NVIDIA 5.9% Sells the GPUs that train and run AI models 88%
Microsoft 6.5% Azure AI cloud, Copilot products, OpenAI partnership 32%
Alphabet 4.0% AI-powered search and ad targeting, Google Cloud AI 28%
Amazon 4.2% AWS AI services, Alexa, AI-driven logistics 22%
Meta 2.9% AI recommendation engine drives 100% of ad revenue 40%
Apple 7.2% Apple Intelligence, AI-driven services growth 8%
Broadcom 2.0% Custom AI chips, networking for AI data centers 35%

NVIDIA is the most exposed. 88% of its revenue comes from data center GPUs bought for AI workloads. But the exposure runs through every name. Meta's ad targeting algorithm is AI. Google's search relevance is AI. Amazon's recommendation engine and logistics optimization are AI. Microsoft is selling AI access to every enterprise on Earth.

Apple is the least AI-dependent of the group, but even Apple's growth narrative relies on Apple Intelligence driving iPhone upgrade cycles. If that story falters, the stock that weighs 7.2% of your index takes a hit.

When the top seven stocks all depend on the same thesis, concentration and correlation compound each other. Your index fund doesn't own seven different bets. It owns one bet expressed seven different ways.
Your Exposure

Calculate Your AI Bet

Pick the fund you hold. The calculator shows how much of your money is riding on AI-linked companies.

35.7%
of your money is in the Magnificent 7 AI stocks
If you have $500,000 in SPY, roughly $178,500 is riding on seven companies that all need AI to justify their valuations.
In the 2022 tech drawdown, these 7 stocks fell an average of 38%. A $500K portfolio would have lost $67,800 from AI exposure alone.
The Dot-Com Parallel

This Has Happened Before

In March 2000, the top 10 stocks in the S&P 500 were 27% of the index. Microsoft, Cisco, GE, Intel, and Walmart led the list. The thesis then was the internet would change everything. That thesis was correct. The valuations were wrong.

March 2000

Top 10 concentration: 27%. The internet was real. The stocks were overpriced. Microsoft fell 65% from its 2000 peak and took 16 years to recover. Cisco fell 80% and never recovered. Intel is still below its 2000 price. GE no longer exists as a single company.

February 2026

Top 10 concentration: 37.6%. AI is real. Are the stocks overpriced? NVIDIA trades at 35x forward earnings. Microsoft at 32x. These are lower multiples than Cisco's 150x in 2000, but significantly above the S&P 500's long-term average of 17x.

The point is not that AI is a bubble. AI's commercial impact may be larger than the internet's. The point is that concentration creates fragility regardless of whether the underlying thesis is correct. When 36% of your index depends on one theme, a recalibration of that theme hurts everywhere at once.

In 2000, the "diversified" S&P 500 fell 49% peak to trough. The top-heavy structure meant that an internet bust dragged down a fund designed to represent the entire economy.

What To Do About It

Three Ways to Think About Your Exposure

Knowing the problem is step one. Here are three frameworks for the portfolio decision.

01
Accept It Intentionally
If you believe AI will deliver on its commercial promise over the next decade, the S&P 500's concentration is a feature. You're getting heavy AI exposure through profitable, cash-rich companies at lower multiples than the dot-com era. The risk you accept: if AI deployment disappoints, you have no hedge. Make this choice explicitly, not by default.
02
Dilute It With Equal Weight
The equal-weight S&P 500 (RSP) holds the same 500 companies but gives each one 0.2% of the fund. The top 7 stocks drop from 35.7% to 1.4%. You still own the AI companies, but they can't dominate your returns. The tradeoff: equal-weight has underperformed cap-weight during the AI run-up. If AI megacaps keep winning, RSP will lag. If concentration unwinds, RSP outperforms.
03
Hedge With Non-Correlated Assets
Keep your S&P 500 fund but pair it with assets that don't move with AI stocks. International equities (VXUS) have 8% AI overlap vs. 36%. Bonds provide ballast. Commodities and REITs respond to different forces. A 60/40 split between SPY and a non-AI-correlated basket reduces your AI concentration from 36% to about 21%.

There is no option where AI disappears from your portfolio. Even the international index holds TSMC, Samsung, and ASML. But there is a spectrum between 36% AI exposure and 8% AI exposure. Choosing where you sit on that spectrum should be a conscious decision.

35.7%
More than a third of your "diversified" index fund is riding on seven companies that all need AI to work. Before you buy more, decide if that's the bet you want to make.

How I Built This

The index weightings in this essay come from published fund holdings. AI revenue estimates are derived from company filings, analyst reports, and segment disclosures. Here are the key assumptions.

S&P 500 Weightings Source
SPDR S&P 500 ETF Holdings
Published daily by State Street. Weightings fluctuate with price changes. The figures in this essay reflect early February 2026 holdings and will shift as prices move. A 10% decline in NVIDIA alone would reduce the Magnificent 7 share by roughly 2 percentage points.
AI Revenue Exposure Estimates
Segment filings + analyst estimates
NVIDIA's 88% is based on its Data Center segment as a share of total revenue. Microsoft's 32% uses Intelligent Cloud as a proxy. Meta's 40% reflects that all ad revenue flows through AI recommendation systems. These are approximations with a ±5-10% margin depending on how "AI revenue" is defined. The structural argument holds across the range.
Dot-Com Concentration Data
27% top-10 weight at March 2000 peak
Source: S&P Dow Jones Indices historical data. The exact figure varies by measurement date. Some sources cite 25-28% depending on whether the measurement uses month-end or intraday peak values. The directional comparison (lower then, higher now) holds across all measurement methods.
Equal-Weight Fund Data
Invesco S&P 500 Equal Weight ETF (RSP)
RSP holds all 500 S&P components at approximately equal weight, rebalanced quarterly. Each stock is roughly 0.2% of the fund. Performance comparisons use total return including dividends.
Jesse Walker
Jesse Walker has been an individual investor for 30 years. Before that, he was a poker professional, which is where he learned that the best decision and the best outcome aren't always the same thing. He writes about financially navigating the uncertainties of AI.