Samsung Electronics and SK hynix are spending at a pace that makes the scale of Korea’s memory-chip race unusually clear. In the first half of 2026, the two companies invested a combined ₩33.4 trillion in research and development, according to regulatory filings reported by Dong-A Ilbo, as the AI-driven memory boom pushes producers to compete not only on capacity but on how quickly they can move to the next generation of chips.
Samsung Electronics spent ₩27.36 trillion on R&D in the first six months, up 51.5% from a year earlier. SK hynix spent ₩6.04 trillion, an increase of 98.4%. For both companies, the increases were at or near record levels. SK hynix’s six-month spending alone is already close to the ₩6.73 trillion it invested during all of last year.
The comparison matters because the current semiconductor upswing is being driven heavily by AI infrastructure. High-bandwidth memory and other advanced memory products have become strategically important components for AI accelerators, turning engineering speed and manufacturing know-how into a direct competitive advantage. The Korean groups are using the favorable part of the cycle to invest rather than simply harvest higher profits.
Why the spending is accelerating now
Samsung’s figure covers the company more broadly, including its Device eXperience businesses as well as semiconductors, so it should not be read as a pure chip-to-chip comparison. Even so, the jump illustrates the amount of capital Korea’s largest technology company is committing to keeping its product pipeline moving while competition from U.S. and Chinese rivals intensifies.
For SK hynix, the increase is especially striking. The company has benefited from its position in high-bandwidth memory, one of the central components of the AI server boom. Nearly doubling first-half R&D spending suggests it is trying to turn that lead into a longer-term technology advantage rather than treating the current demand surge as temporary.
The larger question is what Korea gets from this investment beyond the immediate memory cycle. Sustained R&D at this scale can reinforce the country’s semiconductor ecosystem, but it also raises the stakes around engineering talent, advanced manufacturing equipment and the ability to commercialize new technologies quickly. The next signal to watch is whether record research spending translates into durable leadership as the AI-memory market moves into its next generation.