Do Nvidia, Alphabet, Meta Drops Hurt Your Index Fund?

NVDAYTD-18.9%SPYYTD+7.9%QQQYTD+7.7%
Analyze NVDA →

Today’s big story: Could Anthropic Stock Fall 25%?

At Wednesday’s close on Nov 18, Nvidia stood 36.9% below its October high. Alphabet was 29.8% below its May high. Meta Platforms was 20.7% below its September high.

Most coverage explains each fall on its own. This article looks at the full record instead. Since 2000, three or more of six big tech stocks fell 20% together in 12 separate episodes. SPY, the S&P 500 ETF, was higher 12 months later in 8 of the 11 that have a full year of data.

That sounds reassuring, but any day since 2000 gave about the same odds. SPY was higher a year later on 75.4% of all trading days. So falling giants have not, by themselves, been a signal either way. What split the good years from the bad was how long the leaders stayed down.

circuit board
Image from Pixabay

How Narrow Is Today’s Fall?

Only three of the six are down 20% or more. Microsoft is 8.5% below its high. Amazon is 5.7% below, and Apple just 1.1%. SPY closed 6.3% below its Oct 6 high.

So the index has not followed the three stocks into a 20% fall. QQQ, the Nasdaq-100 ETF, sits in between, at 13.2% below its high.

The S&P 500 is only 6.3% off its high while three of its biggest names are down 20% or more. If you would rather your money leaned less on a few tech giants, steady income is another way to stay invested. The Trefis Income Solution pays a dividend yield about 3x the S&P 500’s, and it lost less than the S&P 500 in all 9 big selloffs since 2018. See the metrics.

Nvidia reported results after the close on Tuesday, Nov 17. Revenue was $115.8 billion for the quarter ended Oct 25, up 103% from a year earlier. The stock still fell 2.9% on Wednesday, to $150.96.

What Happened In The Year After Each Episode?

The table lists every finished episode since 2000 and what followed each start.

Episode start Days with 3+ down SPY a year later QQQ a year later Starting stocks a year later
Mar 14, 2000 481 -13.9% -59.3% -41.1%
Apr 15, 2005 8 +12.7% +19.9% +89.4%
Feb 7, 2006 90 +15.6% +9.4% +19.9%
Jan 16, 2008 372 -38.4% -36.8% -50.6%
May 20, 2010 52 +24.9% +30.9% +40.4%
Aug 19, 2011 2 +26.2% +36.8% +32.3%
Feb 5, 2016 7 +21.8% +28.2% +147.8%
Oct 29, 2018 90 +15.3% +20.7% +19.2%
Mar 9, 2020 20 +41.2% +61.1% +71%
Jan 21, 2022 240 -8.6% -18% -34.3%
Mar 18, 2025 33 +17.6% +25% +39.6%

Source: Trefis analysis of Intrinio split-adjusted daily closes. The six stocks are Apple, Microsoft, Nvidia, Amazon, Alphabet and Meta. Alphabet counts from Aug 2005 and Meta from May 2013, once each had a year of trading, so the early episodes drew on four or five names. A stock counts as down when it closes 20% or more below its high of the prior 252 trading days. An episode starts after 126 trading days with no such cluster. Returns are price only, without dividends. Starting stocks are the names down on the first day, held in equal amounts.

SPY’s median gain over the 12 months was 15.6%. QQQ was higher in 8 of 11 cases, with a median gain of 20.7%.

The stocks that started each episode did even better in the median case. They were higher a year later in 8 of 11 episodes, with a median gain of 32.3%. Their range was wide, from a 50.6% loss in 2008 to a 147.8% gain in 2016.

SPY’s lowest close in each 12-month window was a median 11.2% below the start. The worst was 44.9% below, in November 2008.

Could Investors Tell The Bad Years Apart At The Start?

The three losing years were the long episodes. Three or more stocks were down 20% on 481 trading days in the 2000 episode. The 2008 episode had 372 such days and 2022 had 240.

All eight other episodes had 90 such days or fewer. SPY rose between 12.7% and 41.2% after each of them. But an episode’s length is only known after it ends.

At the starts, SPY was between 3.0% and 18.9% below its high. The bad starts sat inside that range, at 7.1% in 2000, 12.5% in 2008 and 8.3% in 2022.

We also reran the test with other cut-offs for a stock to count as down. At 15% below its high, SPY was higher a year later in 11 of 15 episodes. At 25%, it was 6 of 9. No cut-off beat the any-day rate by a clear margin.

Is 2026 A Short Episode Or A Long One?

By the record’s rule, today’s cluster is part of an episode that began on Mar 27, 2026. Four stocks were down 20% that day, and five on Mar 30. Microsoft, Alphabet and Meta were down 20% together again on Jul 23 and 24.

The current trio first appeared on Nov 5. It has held on four trading days, most recently Wednesday. In all, 2026 has had 8 such days so far. SPY was 20.7% higher six months after the March start.

By its count of days, 2026 still looks like a short episode. The other reading is that three waves in eight months may not be one brief drop. We think the repeat matters, because the long episodes ran 240 days or more.

Nvidia’s own past shows a deep fall can get deeper. After it first closed 33% below its 2021 high, it fell another 49.7% before its October 2022 low. After it first closed 31% below its 2018 high, it fell another 36.3%. Those were different markets and a smaller company, so they are context, not a forecast.

What Does The Record Mean For Your Money?

Our read: three tech giants falling together has not, on its own, told investors where SPY goes next. The odds after these episodes were about the same as on any day. The real risk was a long episode, and the start did not reveal one.

This sits inside the larger AI story. If you hold an index fund, it helps to know how much of it rides on these names. We look at that in Is Your VOO Fund Secretly An AI Bet?

Even a good year could bring a deep dip first. In 2020, SPY’s lowest close was 18.7% below the start, though it ended the 12 months 41.2% higher.

Could You Hold On Through The Next Fall?

Holding on is easier when your portfolio falls less than the market. For every dollar its benchmark lost, the Trefis High Quality (HQ) Portfolio lost about 81 cents on average, holding 30 quality stocks in equal weight.

See How HQ Held Up

For questions about this article or underlying research, email help@trefis.com.