We touched on concentration in Chapter 1. Here we need to sit with it a little longer, because when you understand how index construction works, the concentration story takes on a different dimension.

In 1990, the ten largest S&P 500 companies, names like IBM, Exxon, General Electric, and Philip Morris, accounted for roughly 19% of the index. Their share of total index earnings was also about 19%. Weight and fundamentals were in alignment.

By 2015, not much had changed. The top ten still represented about 19% of the index and about 19% of earnings. The index was working as designed.

Then something shifted. A more durable change than anything seen since the index was created in 1957.

By the end of 2025, the top ten companies hold 40.7% of the index’s total weight. But they are expected to generate only about 32% of its earnings. The market capitalisation of the largest companies has run significantly ahead of their proportional profit contribution.

The last time concentration was even remotely close to this was at the peak of the dot-com bubble in early 2000, when the top ten reached about 27% of the index. Today’s concentration exceeds that peak by more than 13%age points.

This creates three specific dynamics that every Indian investor in a US index fund needs to understand.

The first is the passive feedback loop. Every time any investor anywhere in the world adds money to an S&P 500 index fund, that money flows into all 503 companies in proportion to their existing weights. More money to NVIDIA. More money to Apple. More money to Microsoft. Not because those companies are doing anything new that day. Just because they are big and getting bigger. The index, in a subtle way, amplifies the success of whatever was already successful.

The second is correlated risk. The top ten S&P 500 companies in 1990 were spread across oil, industrial manufacturing, consumer goods, banking, and technology. A crisis in one sector did not automatically damage the others. 

Today’s top ten are overwhelmingly linked by a single theme: artificial intelligence and the technology infrastructure around it. 

NVIDIA makes the chips. Microsoft runs the cloud. Alphabet and Meta build the applications. Amazon provides the infrastructure. If something changes the consensus view on AI, whether it is a regulatory shift, a breakthrough from a competitor, or simply slower-than-expected adoption, there are far fewer uncorrelated positions in the index to cushion the blow.

The third is what it means for you personally. When you buy a “diversified US market fund” today, you are not getting a neutral cross-section of the American economy.

You are making a concentrated bet on AI adoption, on the continued dominance of about a dozen technology companies, and on market participants continuing to assign premium multiples to those companies’ future earnings. That bet has paid off spectacularly well over the past three years. Whether it continues to do so is the central question for US equity investors in 2026 and beyond.

None of this means you should not own the S&P 500. The top 10% of S&P 500 companies by weight generate 60% of the index’s net income. These are genuinely exceptional businesses. The information technology sector in the S&P 500 trades at about 30 times forward earnings today, compared to 50 times at the peak of the dot-com bubble in 2000. This is elevated but not irrational.

But you should own it, understanding what it actually is.