Chips are becoming the new centre of American power

First came the PC boom. Then came the dot-com bubble. Now comes AI. Each wave left its mark on the market, but this time feels different. Chips are no longer just components inside computers. They are the infrastructure behind cloud computing, data centres, defence systems, electric vehicles and artificial intelligence.

That does not mean the sector cannot be overvalued. It probably can be. But the bigger story is harder to dismiss. Nearly one-fifth of the US stock market is now tied to semiconductors. The market is telling us where it thinks economic power is moving.

For decades, semiconductors were important but hidden. They sat inside machines, phones, cars and factories. They made other things faster, smaller and more powerful, but they were rarely the main story. Oil was strategic. Steel was industrial strength. Banks controlled the flow of capital. Chips were the quiet layer underneath the visible economy.

That has changed.

Today, the chip is no longer just an input. It is the bottleneck, the prize and the battleground. It determines how fast AI models can be trained, how much computing power a company can deploy and how much strategic capacity a country can build. The most valuable companies in the world now depend, directly or indirectly, on access to advanced semiconductors.

That is why the chart matters. It does not show GDP, employment or production. It shows market belief. It shows how much of the US stock market is now priced around the companies that make the digital economy possible.

The early 1980s were the first signal. The IBM PC boom lifted semiconductors because personal computing was beginning to move into homes and offices. But the shift was still contained. The market noticed, then moved on.

The dot-com bubble was different. By 2000, investors had convinced themselves that the internet would change everything. They were right about the direction, but wrong about the timing, the valuations and many of the companies. The internet did change everything, but not before the market had badly overpaid for the first version of the dream.

That is the obvious warning today. When any sector rises this much, the bubble question cannot be ignored. AI enthusiasm has pushed expectations to extraordinary levels. Investors are not just pricing in growth. They are pricing in dominance, scarcity and a seemingly endless need for more computing power. Some of that may prove excessive. Markets often take a real story and stretch it until it becomes dangerous. Railways were real. The internet was real. Housing demand was real. That did not stop investors from overpaying.

But dismissing this as just another bubble is too easy. The rise since 2010 does not look like a single speculative spike. It looks like the market gradually realising that almost every important technological shift points back to the same place. Smartphones needed chips. Cloud computing needed chips. Data centres needed chips. Electric vehicles needed chips. Defence systems needed chips. AI needs them at an entirely different scale. That is the difference between 2000 and now. The dot-com era was mostly about what the internet might become. The semiconductor era is about what the world is already trying to build.

There is also a geopolitical story inside the market data. Semiconductors sit at the intersection of corporate profit, national security and industrial policy. The US wants to reshore capacity. China wants to reduce dependence. Taiwan remains central to the global supply chain. Chips are no longer merely commercial products. They are instruments of sovereignty.

In the old economy, power meant owning oil fields, factories, shipping routes and banks. In the new economy, it increasingly means controlling computation. And computation begins with chips. The market may be too excited. The valuations may be stretched. Some investors may learn that even world-changing technologies can become terrible investments at the wrong price.

But the deeper message is clear. The future may look weightless, digital and intelligent. Yet it rests on something very physical: factories, supply chains, electricity, minerals, cooling systems and silicon. The cloud was never really a cloud. AI is not magic. The digital economy has a physical core. And right now, the market is pointing straight at it.