Memory chipmaker SK Hynix has made a “strategic investment” in Spanish RISC-V chip company Semidynamics to advance “memory-centric AI inference architecture.”

The size of the investment has not been disclosed, but in a statement, Semidynamics said the funding represents the “growing importance of tight architectural alignment between processors and advanced memory technologies.”

The partnership will see the two companies collaborate on the optimization of Semidynamics’ architecture with next-generation memory technologies to support the growing demands being placed on infrastructure by AI inference workloads.

The investment will also support future tape-outs and system-level development, including rack platform buildout, the company said. Semidynamics recently completed its first tape-out with TSMC using the chipmaker’s 3nm process node technology.

“SK Hynix’s investment is a direct reflection of where AI infrastructure is heading, systems where memory architecture is as strategically important as compute,” said Roger Espasa, founder and CEO of Semidynamics. “We built Semidynamics around that thesis, and this partnership strengthens our position as we bring our inference platform to market at a moment when the industry has recognized that token economics are a memory problem as much as a compute problem.”

“AI workloads are fundamentally memory-bound problems, and the industry has been underinvesting in architecture-level solutions,” said Heejin Chung, SVP, head of venture investment, SK Hynix America. “Semidynamics is one of the few companies that has built from first principles around this constraint.”

Founded in 2016, Barcelona-based Semidynamics develops memory-focused AI infrastructure that incorporates its Gazzillion memory subsystem technology, designed to overcome an issue known as the “memory wall” – the gap between processor speed and memory bandwidth.

To date, the company has secured €45 million ($53m) in funding from European and Spanish innovation programs to support its ongoing build-out of a full-stack AI infrastructure platform.

Earlier this year, SK Hynix warned that the current memory chip shortage would continue into 2027. However, speaking to reporters at Nvidia’s GTC event in San Jose, California, last month, SK Group chairman Chey Tae-won revised that prediction, saying the shortage is now likely to persist until 2030, with the company expecting to see a wafer shortage of more than 20 percent.