How did the rise of AI data centers—by redirecting Samsung, SK Hynix, and other memory manufacturers’ capacity toward advanced chips, reducing the supply and increasing the prices of conventional DRAM and NAND flash used in smartphones—end the era of falling memory and screen cos

AI data-center demand turned memory from a routine cost deflator into a supply constraint: manufacturers could earn more by prioritizing high-bandwidth and other AI-oriented memory, leaving less capacity for phone-grade DRAM and NAND. That raised handset input costs just as budget-phone makers had little room to raise retail prices. 56

• Why the old cost model broke: For years, falling prices for memory and displays helped smartphone vendors add storage, RAM, and better screens without materially raising prices. AI servers changed the allocation of manufacturing capacity: Samsung, SK Hynix, and peers diverted output toward higher-margin memory for AI workloads, while hyperscalers sought to lock in supply. Conventional memory became scarcer and more expensive. 56

• Why low-end phones were exposed: In a sub-$200 handset, components represent a very large share of the selling price and gross profit per device is thin. A meaningful rise in DRAM, NAND, or display costs cannot easily be passed on: raising price risks losing highly price-sensitive buyers, while holding price converts the increase almost directly into lower gross margin. Vendors can respond only by accepting lower profit, reducing specifications, or cutting marketing and other costs.

• Why Xiaomi was especially affected: Xiaomi’s volume concentration in lower-priced devices made it unusually sensitive to this cost inflation. In Q2 2026, its smartphone revenue fell 7.5% to 42.1 billion yuan and smartphone gross margin dropped to 8.5%, from 11.5% a year earlier, illustrating how component inflation overwhelmed its ability to protect handset profitability. 1

• The reported Q2 outcome: Xiaomi reported revenue of 108.9 billion yuan, down 6.1% year on year, and adjusted net profit of 6.2 billion yuan, down 42.6%; its gross profit fell 17.2% to 21.6 billion yuan. 5 The results missed the estimates in your question and marked a third straight quarterly decline in profit; the immediate driver was margin compression from more expensive memory and other components, compounded by weaker smartphone demand. 1

• Why Apple is better insulated: Apple is not immune—its management has acknowledged that memory prices pressure profitability. 5 But it sells far higher-priced devices, earns much larger dollar gross profit per phone, has stronger consumer pricing power and supplier leverage, and has a substantial services business. It can therefore absorb some higher bill-of-materials cost, spread it across a much larger gross-profit pool, or modestly raise prices without the same demand risk as a budget-focused Android vendor.

• Important distinction: Xiaomi’s group result was not caused solely by memory prices; smartphone demand, the company’s product mix, and costs elsewhere in the business also mattered. But the AI-driven component squeeze was particularly damaging because it struck the company’s core, low-margin handset segment. Xiaomi said it expected the pace of memory-price increases to slow in the second half of 2026. 1