Semiconductor Cycles Explained: How to Tell a Boom from a Bust

Why Are Semiconductors a 'Cyclical Industry'?

The heart of it is the time lag between demand and supply. For products like memory chips, even when demand rises it takes years to build a new plant and reach mass production. As a result, the following pattern repeats.

1. Demand grows and prices rise, so companies race to ramp up capital investment.
2. A few years later, all that added capacity floods the market at once.
3. Oversupply pushes prices down and inventory piles up, so investment is cut back.
4. Over time supply falls short again, and prices start rising once more.

Because demand shifts relatively quickly while supply follows only belatedly, prices and profits swing far to one side and then reverse in a pendulum-like motion. When you add in the business cycles of downstream industries such as PCs, smartphones, servers, and AI, the amplitude grows even larger.

Why is the swing bigger than in other manufacturing? First, memory is a commodity with standardized specifications, so it's hard for companies to differentiate their products, and the overall supply-demand balance ends up setting the price. Second, a chip plant is a capital-intensive facility that costs trillions of won just to build, so once it's up there's a strong incentive to keep it running rather than stop it. That means production doesn't easily shrink even when supply is already overflowing, so price declines tend to run deep. Third, when several companies decide to expand at roughly the same time on the same reasoning, the added volume pours out all at once and the glut is amplified. When these three factors overlap, the troughs of boom and bust become unusually deep.

How Do You Judge a Boom from a Bust?

Instead of making flat predictions, you should watch the direction of the indicators that practitioners commonly track together. Confidence rises when several indicators, not just one, point the same way.

• Memory prices: The trend in contract prices for DRAM and NAND. A turn upward is read as improving conditions, while a sustained decline is read as a sign of oversupply.
• Inventory: Whether inventory at manufacturers and their customers (set makers) is building up or being drawn down. Falling inventory is often cited as a leading signal of recovery.
• Utilization rate: How much the plants are running. In a downturn, companies defend prices by cutting production (lowering utilization) to reduce supply.
• Capital expenditure (CAPEX): Companies' plans to expand or scale back investment. Excessive expansion can plant the seeds of the next oversupply.
• Downstream demand: The business conditions of the buyers who actually purchase chips, such as PC and mobile replacement cycles and data center and AI investment.

These indicators tend to move in sequence, with lags between them. Typically downstream demand revives first, then customers' inventory shrinks, then prices firm up at a bottom, and finally corporate earnings and capital spending recover. Conversely, when a downturn arrives, inventory builds first and prices wobble, followed by production cuts and reduced investment. That's why, rather than jumping to conclusions from just one or two indicators, it's important to read the sequence and direction of several indicators together.

Specific price levels and earnings figures shift greatly over time, so it's safest to verify them directly against the latest disclosures and brokerage reports. Even for the same indicator, the reporting agency and methodology can differ, so it's wise not to treat any single source's numbers as an absolute benchmark.

Why Samsung Electronics and SK Hynix Are Sensitive to the Cycle

Memory chips make up a large share of both companies' revenue. Because memory is standardized in specification, its structure is close to one where price directly drives earnings. When prices rise, profits swell quickly; when they fall, profits can plunge just as fast. On top of that, the heavy fixed-cost burden from large-scale facility investment means that even small changes in utilization and selling price are magnified in profits. As a result, earnings swing widely with the direction of the industry, and this is frequently cited as an example of cycle sensitivity.

Here there's a point to keep in mind. The fact that earnings swing greatly means the stock price also reacts sensitively to expectations about the industry, but the stock price does not simply track earnings. The market often prices in a recovery before earnings actually arrive, and cools off in advance near a peak. In other words, by the time the indicators look good the stock may have already moved, and when the indicators look bad it may already be passing the bottom. So the purpose of understanding the cycle is not to nail the timing, but to grasp the context of which phase of the cycle we are closer to now. The above is merely an example meant to explain industry structure, and is not a recommendation to buy or sell any particular stock.

Common Misconceptions

• "Once prices start rising, it's definitely a boom." A temporary rebound and a genuine trend reversal are different. You need to see whether inventory, utilization, and demand are improving together.
• "Cycles come on a fixed schedule." The length or amplitude of past cycles does not guarantee a repeat. Technological shifts and changes in demand make each cycle play out differently.
• "You can answer with a number how long the boom will last." Fixing a specific date or target price is nothing more than a prediction. What matters is the attitude of continually observing which indicators are moving in which direction.

Frequently Asked Questions (FAQ)

Can you know the bottom or the peak of a cycle in advance?

It's hard to time it precisely. That said, when signals such as falling inventory, production cuts, and a slowing pace of price declines overlap, it's often interpreted as approaching a recovery phase. This is a probabilistic judgment at best, not a certainty.

Do foundries and system semiconductors also ride the cycle?

There is a difference in degree. Memory, a commodity product, has the greatest volatility, while segments with a large share of customized products tend to be relatively milder. Even so, they are still affected by downstream demand.

Will AI demand make the cycle disappear?

The nature of demand may change, but as long as the time lag in supply exists, periodic ups and downs themselves are unlikely to vanish. New demand may alter the amplitude or the shape of the cycle, but it's wise to always treat the phrase "this time is different" with caution. In the end, the right posture toward the cycle is less about nailing a specific number and more about steadily reading the signals the indicators send and building your own basis for judgment.

This article is for informational purposes only and is not a recommendation to buy or sell any particular stock. Investment decisions and their consequences rest with the investor.