Samsung Q4 profit set to double on AI memory chip boom

Samsung’s projected Q4 rebound is one of the clearest macro signals yet that AI compute is now the primary growth engine for the memory industry. DRAM and NAND prices rising double digits in a single quarter on the back of AI server demand and disciplined supply suggests hyperscalers and model labs are still in heavy stockpiling mode. That is consistent with a world where frontier training runs and inference clusters are scaling faster than the industry’s ability to bring new capacity online.

For the AGI race, this is effectively the balance sheet view of the scaling hypothesis: as long as memory and GPU vendors can post outsized profits by feeding AI workloads, capital will keep flowing into larger clusters and more capable models. Samsung’s guidance and management commentary about “AI transformation” across devices and operations also hint that the company sees on-device AI as a second wave, complementing cloud training. If that materializes, we get a two-tier ecosystem—massive data centers plus AI-rich endpoints—which is exactly the substrate you’d want for increasingly agentic systems.