SK hynix Invests in European RISC-V Designer Semidynamics to Tackle AI Memory Bottleneck

SK hynix has invested in Semidynamics, a European chip designer, to address the 'memory wall,' a critical bottleneck in AI systems, as memory costs in AI data centers are sharply rising.

Semidynamics specializes in RISC-V technology designed to maximize the efficiency of high-bandwidth memory (HBM), which helps SK hynix solidify its market-leading position and add value to its products.

This investment also acts as a geopolitical hedge, allowing SK hynix to navigate US-China trade tensions by collaborating with a European partner on the open standard RISC-V architecture.

000660.KS

SK hynix's recent investment in Semidynamics is a strategic move to solve one of AI's biggest challenges.

AI systems are incredibly data-hungry, creating a 'memory wall' where processors wait idly for data from memory. This bottleneck is becoming more severe as AI models grow. In fact, hyperscalers are now expected to spend nearly 30% of their AI data center budget on memory in 2026, a huge jump from just 8% a couple of years ago. This surge in spending highlights the urgent need for solutions that use expensive, high-performance memory like HBM more efficiently.

This is where SK hynix's investment comes in. As the leading supplier of HBM, SK hynix has its capacity sold out for the near future. To maintain its leadership and justify premium prices, it needs to help customers get more value from each memory chip. Semidynamics offers a perfect solution with its specialized RISC-V CPU cores. Their technology is designed specifically to reduce processor stalls by managing hundreds of memory requests simultaneously, effectively tearing down the memory wall and maximizing HBM utilization.

The partnership is also a savvy geopolitical move. Amid rising tech tensions between the U.S. and China, collaborating with a European company on an open-source standard like RISC-V provides a stable, lower-friction path for innovation. Furthermore, with strong support from the EU's Chips Act and Spain's PERTE Chip program, Semidynamics is a de-risked partner, giving SK hynix a solid foothold in Europe's growing semiconductor ecosystem.

In essence, this isn't just an investment in a startup; it's SK hynix moving up the value chain. By backing a partner that can unlock the full potential of its HBM, the company is defending its market share, diversifying its partnerships, and positioning itself to capture more value from the AI revolution.

HBM (High Bandwidth Memory): A type of high-performance memory used in AI accelerators and supercomputers, offering much faster data transfer speeds than conventional DRAM.
RISC-V: An open-source instruction set architecture (ISA) for computer processors. Being open, it allows anyone to design, manufacture, and sell RISC-V chips and software without licensing fees.
Memory Wall: A term describing the growing gap between processor speed and the speed at which data can be retrieved from memory. This forces the processor to wait, wasting valuable computing cycles.

SK hynix Invests in European RISC-V Designer Semidynamics to Tackle AI Memory Bottleneck

SK hynix has invested in Semidynamics, a European chip designer, to address the 'memory wall,' a critical bottleneck in AI systems, as memory costs in AI data centers are sharply rising.

This investment also acts as a geopolitical hedge, allowing SK hynix to navigate US-China trade tensions by collaborating with a European partner on the open standard RISC-V architecture.

000660.KS

SK hynix's recent investment in Semidynamics is a strategic move to solve one of AI's biggest challenges.

HBM (High Bandwidth Memory): A type of high-performance memory used in AI accelerators and supercomputers, offering much faster data transfer speeds than conventional DRAM.
RISC-V: An open-source instruction set architecture (ISA) for computer processors. Being open, it allows anyone to design, manufacture, and sell RISC-V chips and software without licensing fees.
Memory Wall: A term describing the growing gap between processor speed and the speed at which data can be retrieved from memory. This forces the processor to wait, wasting valuable computing cycles.