The stocks of Samsung Electronics and SK Hynix have failed to join the broader semiconductor rally, even as the industry's AI-driven upcycle lifts chipmakers elsewhere — and a new demand category, DRAM for robots, is now being discussed as a possible turning point for Korea's two memory leaders.

The contrast is stark. While the global semiconductor sector has climbed on the back of AI accelerator demand, the two companies that together dominate the world's DRAM and NAND output have not shared in the upside. Investor attention has shifted to whether the next leg of memory demand will come not from servers or smartphones, but from humanoid robots and autonomous machines.

Why have Samsung and SK Hynix lagged the rally?

The two Korean suppliers sit at the center of the global memory market, yet their shares have decoupled from a semiconductor rally that has rewarded companies tied more directly to AI compute. The question raised by market watchers, as framed in a BigGo Finance report, is blunt: can "robot DRAM" become the breakthrough that reconnects Samsung and SK Hynix to the industry's momentum?

Memory demand has historically moved in cycles driven by PCs, then smartphones, then datacenter servers. Each wave rewarded the suppliers positioned for it. The current rally has been powered by AI, and the parts of the memory stack capturing that upside are narrow. A broader catalyst — one that pulls standard DRAM volumes up with it — has so far been missing for the Korean duo.

What is robot DRAM, and why does it matter?

Robots equipped with onboard AI need memory close to the point of sensing and decision-making. Unlike a datacenter GPU fed by high-bandwidth modules, a mobile robot must carry its memory with it, under strict power and thermal limits. That creates a distinct demand profile: high capacity, low power, and resilience in physically demanding environments.

If humanoid robots and industrial automation scale from pilots into volume deployment, each unit would carry its own memory complement. Aggregated across millions of machines, that could add a new demand pillar for DRAM makers — one where Samsung's and SK Hynix's manufacturing scale, honed over decades of supplying phones and PCs, translates directly.

Can robotics demand move the needle for memory giants?

The honest answer is that nobody knows yet, and the source material presents robot DRAM as an open question rather than a forecast. Samsung and SK Hynix produce the overwhelming majority of the world's DRAM across leading-edge nodes; any new category large enough to tighten supply would affect pricing across their whole portfolio, not just a niche product line.

The skepticism is warranted. Robotics remains a young market measured in unit volumes that are small next to smartphones. For robot DRAM to matter commercially, robot shipments would need to grow by orders of magnitude. What the robot-DRAM thesis offers today is a plausible narrative for where the next mass-market memory socket could appear — the same position smartphones occupied in the late 2000s.

What should investors and buyers watch next?

For anyone tracking the Korean memory makers, the signals to monitor are concrete:

• Any statement from Samsung or SK Hynix identifying robotics as a named target segment in their DRAM roadmap
• Design wins in which memory suppliers are qualified by humanoid-robot or autonomous-machine manufacturers
• Commentary from management on whether robot demand factors into capacity or pricing guidance
• Evidence that robot unit volumes are climbing toward levels that register in memory demand statistics

Until those signals appear, robot DRAM stays a thesis, not a market. The Korean duo's underperformance against the semiconductor rally reflects the absence, so far, of a demand story investors can underwrite.

The deciding factor will be whether robotics demand grows into a volume category big enough to tighten the DRAM market — and until then, Samsung and SK Hynix will keep trading on memory-cycle fundamentals rather than AI-rally momentum.