The S&P 500 has rarely leaned so heavily on so few names. As of June 11, 2026, the top 10 holdings of the SPDR S&P 500 ETF (SPY) — the most-tracked proxy for the index — account for roughly 37.3% of the fund’s market value. That share is meaningfully higher than the 27% peak the same cohort touched during the 2000 dot-com bubble, according to Goldman Sachs research.
All 10 of those stocks have a direct line to the AI buildout: chips, cloud capacity, data center networking, or the platforms that are absorbing the spend. With U.S.–Iran peace headlines pushing futures higher into Monday’s open and Brent slipping near $85 a barrel, the index is heading into a fresh week perched on the same narrow shoulders that carried it to a record close above 7,600 earlier this month.
The 10 stocks that decide where the index goes
Here is the snapshot of SPY’s top holdings as published by stockanalysis.com using State Street’s own disclosure data.
| Rank | Ticker | Company | Weight |
|---|---|---|---|
| 1 | NVDA | NVIDIA | 7.84% |
| 2 | AAPL | Apple | 6.83% |
| 3 | MSFT | Microsoft | 4.56% |
| 4 | AMZN | Amazon | 3.71% |
| 5 | GOOGL | Alphabet (Class A) | 3.30% |
| 6 | AVGO | Broadcom | 2.88% |
| 7 | GOOG | Alphabet (Class C) | 2.65% |
| 8 | META | Meta Platforms | 1.96% |
| 9 | TSLA | Tesla | 1.77% |
| 10 | MU | Micron Technology | 1.76% |
| Top 10 total | 37.26% | ||
Notice how few names actually drive the list. Alphabet appears twice because its two share classes (GOOGL and GOOG) are counted separately, and NVIDIA alone carries more weight than the bottom five names combined. The mix is also one-sided thematically: chips and chip enablers (NVDA, AVGO, MU), hyperscale cloud (MSFT, AMZN, GOOGL/GOOG, META), and AI-adjacent platforms (AAPL, TSLA). The S&P 500’s drift is now tightly coupled to whatever the Nasdaq 100 is doing on any given session.
Why this run is unlike past peaks
Concentration in U.S. equities is not new. Goldman Sachs’ equity strategists, who maintain a century-long dataset of S&P 500 weights, count seven episodes in which the top 10 reached “extreme” levels — including the late-1960s Nifty Fifty era and the dot-com bubble. What is different now is the size of the gap.
The 2026 figure is roughly 10 percentage points above the dot-com top. That gap matters because the index’s day-to-day behavior is now bound to a smaller set of earnings prints, capital-expenditure plans, and AI demand signals.
The AI thread runs through every name
Behind the weights are the actual AI revenue prints. Broadcom’s Q2 fiscal 2026 8-K reports semiconductor AI revenue of $10.8 billion, up 143% year-over-year, with the company guiding Q3 AI revenue to about $16 billion — growth above 200% YoY. That single line item is now larger than the total quarterly revenue of most S&P 500 members.
NVIDIA’s index weight is the cleanest expression of the AI capex cycle: data center sales, networking through Mellanox, and the Blackwell ramp into the second half of 2026 sit at the top of every hyperscaler’s order book. Microsoft, Amazon, Alphabet and Meta — the four largest U.S. cloud and platform spenders — are simultaneously NVIDIA’s biggest customers and four of the next five names on this list. Micron rounds out the chip block as the high-bandwidth memory supplier for those same GPUs.
What concentration peaks have meant historically
Concentration at this level matters for three reasons.
1) Index returns are not diversified. A passive S&P 500 buyer is not getting “the market” so much as a heavily weighted AI capex basket sleeved inside 490 other holdings. The bottom half of the index can rally and barely move the tape.
2) Forward returns historically moderate. Goldman’s research notes that in five of the seven prior extreme-concentration episodes the index continued higher after the peak — but the two exceptions, 1973 and 2000, marked the start of prolonged bear markets. The base rate cuts both ways.
3) Factor exposures get distorted. When NVDA alone is 7.84% of the cap-weighted index, S&P 500 returns can no longer be cleanly described as “U.S. large-cap equity.” Quality, growth and momentum factor tilts ride along whether the buyer wants them or not.
What to watch from here
The near-term catalysts are concentrated, just like the index. NVIDIA reports next in late August, Broadcom in early September, and Micron’s Q3 FY26 print drops June 24 — each of those releases can move the index by itself. The Fed’s stress-test results on June 24 add a secondary cross-current via the financials, but the dominant signal remains AI capex.
For investors, the practical takeaway is to know what they actually own. A 60/40 portfolio that holds the S&P 500 on the equity side is — today — closer to a 24% AI-capex sleeve than to the diversified U.S. equity allocation textbooks describe. That is neither a buy nor a sell signal. It is a position to size with eyes open.
Sources
- SPDR S&P 500 ETF (SPY) Holdings — State Street disclosure, as of June 11, 2026.
- Goldman Sachs Asset Management — Is the S&P too concentrated?
- Broadcom Q2 FY26 8-K (SEC filing)
- TheStreet — S&P 500 above 7,600 for the first time (June 2, 2026)
- TheStreet — Stock Market Today (June 15, 2026): Futures jump on U.S.–Iran peace deal
Disclosure: This article was produced with AI assistance and reviewed before publication. It is for informational purposes only and is not investment advice.