AI Demand Elasticity May Rewrite Storage Cycles, Expansion Not Necessarily Leading to Profit Collapse

On July 18, analyst Jukan published a report indicating that a recent market analysis suggests that the recent pullback in storage chip stocks is influenced not only by deleveraging factors but also by the market beginning to trade on expectations of supply expansion in 2028. Although the industry generally expects that the tight supply-demand situation for high-end storage products like HBM will persist until 2027, research institutions and sell-side analysts widely believe that as manufacturers like Samsung Electronics and SK Hynix ramp up production, the supply-demand gap is expected to start easing in 2028. Traditional wisdom holds that storage stocks typically peak about two quarters ahead of storage prices, but this analysis suggests that the market may not be confined to this pattern and could reflect expectations of future supply increases even earlier.

However, all pessimistic expectations are based on the same assumption—that new production capacity will be released in concentrated amounts in 2028, leading to a significant drop in storage prices. Historically, since the 1980s, price crashes in the storage industry have almost always stemmed from rapid supply expansion rather than demand decline, with demand in the PC, smartphone, and cloud computing eras consistently growing. This AI cycle may fundamentally differ from previous ones: the demand for computing power and storage in AI applications exhibits higher price elasticity, meaning that when prices drop, the rate of demand growth may exceed the impact of the price decline, thereby weakening the traditional storage cycle's destructive power.

The analysis cites a paper by Zhang and Zhang published in 2026 titled "The Economics of Digital Intelligence Capital," which states that the price elasticity of demand for AI Tokens is approximately 1.42, meaning that for every 1% drop in price, demand can increase by about 1.42%. The paper argues that when API prices are significantly reduced, developers will not only increase their usage but will also adopt more compute-intensive inference architectures, leading to convex growth in Token consumption. For example, if the price of DRAM drops by 30%, traditional cycles often see significant reductions in revenue and profit; Samsung Electronics' operating profit fell by 52.8% year-on-year in 2019. However, in the new demand environment driven by AI, if sales increase by about 42% and process upgrades drive costs down by about 15%, industry revenue may remain stable, with profit declines expected to narrow to about 15%. This difference may determine whether storage manufacturers should continue to value themselves according to traditional cycle stock price-earnings ratios of 5 to 6 times.

Disclaimer: This content is provided for general branding and informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online events, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets or to use any services. Crypto assets are highly volatile and may result in loss. WEEX services and online events may not be available in all regions and are subject to applicable laws, regulations, and eligibility requirements. You are responsible for ensuring that your use of WEEX services complies with local laws and for carefully assessing the risks before participating in any crypto-related activities.