Samsung shifts internal mobile DRAM deals to quarterly amid price surge

Samsung's device solutions and mobile experience teams agreed to continue mobile DRAM on a three-month pricing cadence to secure volumes ahead of Galaxy S26, but the change exposes the phone division to shorter-term market volatility as memory and processor costs climb.

Samsung’s semiconductor unit and its smartphone division have agreed to continue mobile DRAM supplies on a three-month pricing cadence rather than locking in annual prices, according to industry sources. The device solutions team has set baseline volume commitments for next year to try to prevent shortages before the launch of Galaxy S26. The arrangement keeps supply flexibility but leaves the mobile experience group exposed to rapid market swings because Samsung is not exempting its own divisions from fluctuations in DRAM and NAND prices.

The figures in the report underline the pressure on both teams. Mobile DRAM prices have more than doubled, with 12 GB LPDDR5X modules reaching around $70 in late November, up from about $33 in January. Processors, already the most significant expense in any smartphone, have increased by 25.5% compared to the previous year, raising the dx division’s Q3 procurement bill to 10.9 trillion Korean won from 8.7 trillion won. Together, processor and memory now account for about 35% of the total device cost, a 5% increase from historical levels, adding cost pressure for the mobile experience team as they finalize device pricing and configurations.

The internal tension also reflects a strategic reallocation inside Samsung. The device solutions division is pursuing higher margins from hbm used in Artificial Intelligence accelerators, reallocating wafers and capacity away from standard mobile chips. That shift helps the semiconductor unit capture more profitable enterprise demand but reduces internal supply for phones just as mobile memory prices climb. The report notes that buyers who signed longer DRAM contracts earlier may regret those arrangements if spot prices continue to rise, while quarterly windows transfer short-term market risk back to the phone business.

55

Impact Score

Introducing Mistral 3: open artificial intelligence models

Mistral 3 is a family of open, multimodal and multilingual Artificial Intelligence models that includes three Ministral edge models and a sparse Mistral Large 3 trained with 41B active and 675B total parameters, released under the Apache 2.0 license.

NVIDIA and Mistral Artificial Intelligence partner to accelerate new family of open models

NVIDIA and Mistral Artificial Intelligence announced a partnership to optimize the Mistral 3 family of open-source multilingual, multimodal models across NVIDIA supercomputing and edge platforms. The collaboration highlights Mistral Large 3, a mixture-of-experts model designed to improve efficiency and accuracy for enterprise artificial intelligence deployments starting Tuesday, Dec. 2.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.