Report: Samsung execs worried company could lose money on smartphones for the first time
Summary
Samsung's MX (mobile experience) division faces a potential first-ever net loss on smartphones in 2026, despite strong Galaxy S26 sales, due to skyrocketing prices of DRAM and NAND memory. These components, particularly LPDDR5x, are increasingly critical for AI applications, with Nvidia's upcoming Vera AI CPU and rack-scale platforms consuming vast amounts. Memory and storage costs have roughly doubled, now accounting for over a third of a budget phone's bill of materials and more than 20 percent for premium devices. While Samsung's MX division struggles, its semiconductor division reported a record $38 billion profit in Q1 2026. Despite efforts by Samsung, Micron, and SK Hynix to boost production, DRAM supply is projected to fall 40 percent short of demand by 2027, leading to price increases across Motorola's Moto G series and Samsung's Galaxy A, Z Flip, Z Fold, and Tab S11 devices.
Key takeaway
For entrepreneurs in consumer electronics or AI hardware, you should anticipate sustained high memory and storage component costs through 2027. This necessitates re-evaluating product pricing strategies and bill of materials to maintain profitability, especially for devices with significant memory footprints. Consider diversifying supply chains or exploring alternative memory technologies to mitigate future supply shocks.
Key insights
Soaring memory costs driven by AI demand threaten smartphone profitability and increase device prices.
Principles
- AI compute demand drives memory market dynamics.
- Component costs disproportionately impact budget devices.
In practice
- Monitor LPDDR5x memory pricing trends closely.
- Evaluate BOM structures for AI-driven cost shifts.
Topics
- Samsung MX
- Smartphone Profitability
- DRAM and NAND Pricing
- AI Memory Demand
- Bill of Materials
Best for: Entrepreneur, Product Manager, Executive, Investor
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by AI - Ars Technica.