SK hynix unveils next-gen AI memory technologies
Article excerpt
Highlighted: the sentence this signal was extracted from
The company showcased rack-scale AI memory systems, high bandwidth flash and a faster, more efficient CXL memory product at Open Compute's global summit in California. SK hynix said Friday that it showcased its next-generation AI memory technologies aimed at expanding the boundaries of AI infrastructure beyond conventional high bandwidth memory. The company presented the technologies at the 2026 OCP Global Summit, held in California from Tuesday to Thursday. The summit is an annual conference organized by the Open Compute Project Foundation that brings together engineers and industry leaders. The Korean chip giant said it divided its exhibition space into two sections. The front section showcased the company's flagship products, while the larger rear section featured memory solutions designed to support next-generation server ecosystems. Kim Ho-sik, a senior vice president at SK hynix, introduced MARS (memory-centric AI rack-scale system), a disaggregated rack-scale architecture designed to coordinate computing and memory resources seamlessly. Kim highlighted progress in high bandwidth flash, a technology aimed at addressing the industry's growing demand for memory capacity. The company also unveiled its second-generation 256-gigabyte memory product based on Compute Express Link 3.2, offering more than twice the bandwidth of its predecessor and an improvement of about 30...
Keep reading with a free account
The rest of this article, and every signal for SK Hynix, is in your free account.
