How Semiconductor Stocks Became the World's Thermometer

Broadcom's shares rose sharply, exciting the market, reportedly on expectations for custom AI chips and networking equipment. This is not an assessment of those particular catalysts. The more interesting question is why semiconductor stocks have come to be treated as a mirror of the global economy.

Semiconductors have long followed a distinctive silicon cycle. Rising demand prompts enormous factory investment, but construction takes years. By completion, demand may have passed, producing repeated shortages and gluts. These large swings made chip stocks objects of attention as leading economic indicators. From PCs to smartphones to cars, semiconductors have reflected what might sell next.

A distinguishing feature of the present AI boom is that demand really exists. Major cloud companies are expanding the computing resources needed for generative AI. This prevents a simple comparison with some past bubbles. Yet actual demand and the future demand priced into shares need not move at the same speed. Announced investment and eventual returns are separated by a lag. Markets try to bridge it by anticipating expectations.

History offers many cases of expectations for new technological infrastructure driving prices ahead of demand. Nineteenth-century railway stocks and excess fibre investment in the late 1990s funded infrastructure eventually needed, but share prices priced in that eventuality too early. A technology's direction being right and its share price being right at a particular moment are different questions. Keeping that distinction helps separate excitement from composure.

Sector indices such as the Philadelphia Semiconductor Index, or SOX, are often quoted as barometers of wider market sentiment, reflecting confidence in the industry's leading role. But a thermometer does not create heat. When semiconductor stocks surge, do we see the temperature of the real economy or of participants' expectations? The two are often hard to distinguish.

Research into crowd psychology repeatedly notes how difficult it is for people inside an excitement to recognize their own excitement. Overheating that seems obvious later is experienced as 'this time is different'. There is little present reason to doubt AI's technological importance. Precisely for that reason, confidence in the technology should be quietly separated from confidence in today's share price.

Calling semiconductor stocks the world's thermometer is probably reasonable. But thermometers can sometimes confuse the temperature of what they measure with their own. Broadcom's rise shows that the AI narrative is not over, no more and no less. Judgment remains in each reader's hands.