Everyone in the group is shouting bullish, but I’m keeping my eye on cash flow in the earnings reports. NVIDIA alone accounts for about 6% of the S&P 500, so index funds and target-date retirement funds can’t avoid it. The S&P was up 2.9% this week; excluding NVIDIA, it might have been closer to 3.8%. One stock held the index back by nearly a percentage point. In other words, a lot of people think they’ve bought a “diversified” broad-market fund, when in reality, the top 10 holdings already determine a large part of the index’s volatility. NVIDIA is the most obvious variable.

Here’s why I focus on cash flow instead of sentiment. In NVIDIA’s latest earnings report, data center revenue accounted for nearly 90% of the total. Gaming, professional visualization, and automotive made up only a small fraction. The more concentrated the revenue mix, the more important it is to see whether operating cash flow is keeping pace with profit growth, whether inventory or receivables are showing any unusual changes, and whether capital expenditures are accelerating. For example, if revenue doubles year over year in a quarter but free cash flow growth clearly lags behind, the market will start asking: How much of these profits is paper prosperity, and how much is actually being realized in cash? The bulls in the group are looking at orders and the narrative. I’m looking further down the income statement: operating cash flow, free cash flow, and days inventory outstanding. It’s not that NVIDIA is a bad company. But when one stock makes up 6% of the index, the details in its earnings report are no longer just its own business—they matter to everyone holding broad-market funds.

Here’s the painful part: every time my stop-loss gets triggered, the market turns around, and now I’m actually afraid to set one. I just look at the revenue mix instead. I used to try mechanical stop-losses, like setting one at -8% or exiting if the stock fell below its 20-day moving average. But several times, a long lower wick during the trading session stopped me out, only for the stock to slowly climb back over the next two or three days and eventually rise even higher than before. It felt like someone was watching my account and taking aim. Later, when I reviewed what happened, I realized the problem wasn’t stop-losses themselves; it was that the criteria I used were too short-term and too emotional. Now I slow things down: first I look at whether the company’s revenue mix is deteriorating, whether major customer orders are weakening, and whether cash flow can cover capital expenditures. If the revenue mix is intact, I don’t react much to short-term volatility. If it changes, I’ll reduce my position without waiting for a stop-loss level. The same habit applies to index holdings: the broad-market funds you buy contain a bunch of holdings you haven’t examined closely, and NVIDIA is the clearest example. You think you’ve bought just one ticker, but in reality, you indirectly own 6% NVIDIA, along with the data center capex cycle, advanced manufacturing capacity, and customer concentration risks behind it.

One more detail: many target-date retirement funds may look conservative, but their equity allocations also include the S&P 500 or total-market indexes, so NVIDIA’s weight gets passed right through. So don’t just look at the word “conservative” in a fund’s name. Look through to its holdings, and you’ll find that its top holdings aren’t much different from those of the index. The convenience of index investing is that it saves you the effort; the tradeoff is that you have to accept that holdings you haven’t examined closely can determine your returns for you. NVIDIA is that stock right now. If you have anything to add, see you in the comments.