Investors Pulled $7 Billion From Three Cash Funds In A Month. Where Did It Go?

Each of the four is earning far more on each dollar of sales than its five-year average. That's the good news, and the risk.

Read across the rows and three different stories show up. Nvidiais the chip designer that looks least stretched for its growth: 24 times earnings for a company whose sales rose 83% in a year, and it turned 42 cents of every sales dollar into free cash flow. AMDis the opposite: 98 times earnings, because its profits are only now catching up with its price. Micronlooks cheapest of all at 9 times earnings, but its operating margin is 75% against a five-year average of 20%. Memory makers have often looked cheapest right at the top of the cycle, when profits are biggest. A low P/E on peak margins isn't the same as a low price.

So the money going into chip funds didn't buy "chips" in general. It bought a few companies earning margins far above their own five-year averages, at prices that already assume a lot of those margins last.

How Far Have They Already Fallen?

Here's where they stood at Friday's close, after a partial rebound that day (SMH +3.6%, AMD +11%). The first column is Thursday's move; the third is how far each sits below its 52-week high, using Friday's close against the high in our price data:

Some of last month's new chip-fund money took a double-digit hit in a single day.

Two of these had a spectacular year before this. Micron is still nearly three times its price of a year ago, and AMD is up more than half, so a slide of 31% to 43% gives back only part of the run. Broadcom is the one that was already sliding: it was down over 12 months even before Thursday. Nvidia, the biggest of them, hit its high just a month ago, on October 6.

How Low Can Chip Stocks Go?

This is the number to set beside any rebound talk. Here's the worst fall of each in the last three big selloffs, peak to bottom, and when it got back to the old high, from the Trefis price cache:

In 2022 the chip fund fell nearly twice as far as the S&P 500, and most of its biggest names fell two to almost three times as far.

Two things stand out. First, 2020 was quick for everyone: chips and the index all got back to their highs within months. Second, 2008 and 2022 were not. After 2008, Nvidia took until 2016 and AMD until 2017 to get back to their old highs; the chip fund took until mid-2013. After 2022, it took Micron, AMD and TSMC until early 2024. Nobody knows which kind this is. But the table tells you that a chip position can go from up a lot to down by half, and stay there for a year or more.

The index-level record is gentler. In what happened the last 21 times the Nasdaq fell 5% in a day , the S&P 500 was higher a year later 20 times, median about +25%, but three months later only 12 times. Buying the dip in the index has usually worked; it often hasn't felt good for the first few months, and chip stocks swing far more than the index.

What's The Bigger Number?

Here's the part that will still matter in 2027: how much chip you already own before you add. Nvidia alone was 7.30% of SPY and Broadcom 2.33% as of Nov 4, per State Street, nearly 10% in two chip names. Together they are worth about $5.9 trillion on Friday's prices. On $100,000 in SPY, that's about $9,600 riding on two chip companies before you buy a single chip fund.

So a dollar moved from T-bills into SOXX isn't a dollar moved from "safe" to "stocks." It's a dollar moved from the steadiest corner of the market to the corner your index fund may already lean toward.

What Should You Do This Week?

1. Set your cash floor first.Keep the money you'll need in the next year or two in cash or T-bills, whatever the market does. In 2008, 2020 and 2022 the T-bill fund BIL never fell more than about 1%. That floor isn't idle money; it's what keeps a bad quarter from forcing you to sell. Only cash above the floor is cash you can invest.

2. Set your chip cap.Pick the most of your stock money you're willing to have in semiconductors, counting what's inside your index funds; say 15%. Add up Nvidia, Broadcom and the rest inside your funds, plus any chip funds or stocks you own directly, and compare. If you're near the cap, a chip fund isn't the place for new money.

3. If you add, add in steps, sized for the 2022 row.Before you buy, look at the 2022 column and ask whether you could hold through another fall that size. Then split new money into a few set amounts over the coming weeks, or set the prices at which you'll buy. A further 10% fall then becomes a plan, not a shock.

What this replaces: emptying the T-bill fund into whatever fell most, or freezing because it fell. The risk: if chips rebound fast, a capped, step-by-step buyer owns less of the rebound than an all-in buyer. If they fall further, the cap and the floor are what let you keep buying.

What Are The Pros Watching?

Three things, and none of them is next week's flow report. Margins: every name in our first table is earning well above its five-year average, so the first sign of margins slipping matters more than any single day's price. Whether money keeps leaving cash funds even as chips fall, which would mean the buying is a plan, or stops, which would mean it was a reaction. And Nvidia's quarterly report later this month, the biggest single test for the chip story. Each is a reason to check your floor and your cap. None is a reason to change them on the day.

What's The Bigger Idea?

We call it Fear is a planning mistake. The fear of missing out that pulls money out of T-bills, and the fear that freezes it after a 10% chip day, are the same mistake: deciding with no plan in front of you. A cash floor and a sector cap take the decision out of the moment. They fit the bigger 60/40 question we started in Is 60/40 Dead?: how many kinds of trouble can your money survive?

The cash went to stocks, a lot of it to chips, just before chips fell hardest.About $7.1 billion left BIL, IQMM and SHV in the month to Nov 4 while SOXX and SMH took in $9.8 billion. The big names inside earn margins well above their five-year averages, sit 22% to 43% below their highs, and at their worst fell 48% to 88% in past selloffs. Before you add, set a cash floor of one to two years' needs and a chip cap that counts your index funds, then add in steps.

What Should You Ask Next?

• How much Nvidia do you really own? The chip stocks inside your index funds.

• Bonds now pay more than stocks. Is 60/40 back?

Follow this story with us: we'll keep watching where the money moves and tell you what changed, and what didn't, for a long-run plan.

Does Your Plan Say How Much Cash To Keep?

Money moves fast on days like Thursday; a plan shouldn't. If you have $1 million or more invested, have a Trefis expert partner advisor set your cash floor and sector limits for a second opinion in a free 15-minute call, or let Invest With Trefis build and run the plan for you.

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