Rising automotive memory prices are starting to ripple through China’s auto market. According to IThome, citing CCTV News, more than 10 Chinese new-energy vehicle makers have recently raised vehicle prices or scaled back discounts, typically by RMB 2,000 to RMB 6,000. The latest round of price adjustments has been driven by higher component costs, with automotive-grade memory chip prices rising around 180% over the past three months, the report notes.
Against this backdrop, IT Home notes a clear divergence in pricing between internal-combustion-engine (ICE) vehicles and new-energy vehicles. While many new-energy models are raising prices due to rising component costs, the ICE vehicle market continues to deepen discounts. However, the report points out that steep price cuts have failed to effectively boost sales. ICE vehicle retail sales fell 37% year over year to about 530,000 units in April, highlighting the significant shift underway in the market.
The trend reflects the ongoing transition in the auto market. According to TrendForce, global NEV sales will reach 23.35 million units in 2026, a 14% YoY increase that will continue to outpace the growth of ICE vehicles. However, rising upstream costs are starting to weigh on demand. Upward pricing pressure in the auto market is expected to be stronger this year than in 2025.
Multiple Forces Drive Tightness in Automotive Memory Supply
The sharp rise in automotive memory prices reflects broader shifts in both the AI and automotive industries. According to China Automotive News, surging AI demand is a key driver behind rising automotive memory prices. The report notes that the world’s three largest memory suppliers—Samsung, SK hynix, and Micron—have allocated more than 80% of their advanced-node capacity to AI servers.
At the same time, the rapid advancement of vehicle intelligence is also driving memory prices higher. China Automotive News points out that as regulations are gradually relaxed in various countries, L2 and L3 driver-assistance systems are becoming more widespread, significantly increasing memory requirements in vehicles. While an L2-equipped vehicle may require only 8GB of memory to meet basic needs, enabling urban navigation-on-autopilot (NOA) functions and on-device large language models can push memory requirements to more than 300GB, an increase of over 35 times.
Adding to these pressures are structural constraints within the automotive memory supply chain. As the report highlights, automotive-grade memory must meet stringent reliability and safety requirements, resulting in certification cycles of 18 to 24 months and making capacity expansion more challenging.
