The $MU CEO's statement that memory supply might not catch up with demand until 2028 doesn't mean memory prices won't fall by then. I think an important detail is being overlooked here: Memory prices don't need to equalize with demand.

For example, let's say there are 100 units of supply and 120 units of demand in the market. When manufacturers increase capacity and supply to 110, there's still a supply deficit. However, the shortage drops from 20 units to 10 units. Even this can put downward pressure on prices.

So, memory prices can fall even as demand continues to increase. It doesn't need supply to completely catch up with demand; it's enough for supply to grow faster than demand.

It's also important to remember the HBM side here. Due to AI investments, HBM demand is increasing very rapidly, and manufacturers are allocating more capacity to these high-margin products. Therefore, it's also important to know how much of this new capacity will actually alleviate the market.

Another important point is this: A decrease in memory prices doesn't necessarily mean a decrease in Micron's revenue and profit. Because the amount of bits the company sells can grow much faster. For example, if the price drops by 20% while sales volume increases by 40%, the company's revenue will still rise. Moreover, the picture can change even more as the share of higher-value products like HBM in total sales increases. LTAs are also very important.

Therefore, I don't interpret the Micron CEO's statement as "memory prices will remain high until 2028." Rather, he's saying that AI-driven demand is so strong that, despite billions of dollars in new investments by manufacturers, closing the supply gap could take years.

I think the real thing to watch for MU isn't when supply will fully catch up with demand, but rather the speed at which the supply gap is closing, bit growth, the HBM mix, and the company's pricing power.

Because a decrease in memory prices is not the same as Micron's investment thesis being undermined.