SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Rambus (RMBS) - Eagle or Penguin
RMBS 59.55-8.8%Jul 17 3:59 PM EDT

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
From: Elroy6/3/2024 8:41:14 AM
  Read Replies (3) of 93622
 
Anybody understand how RMBS fits into this story??

SK Hynix races ahead in HBM with next-gen innovation

digitimes.com

SK Hynix is swiftly advancing in the High-Bandwidth Memory (HBM) arena, aiming to develop next-generation HBM that incorporates computing, communication, and other advanced functions. This strategic move is designed to leverage differentiated technologies, further extending its lead over competitors.

According to industry insiders cited by South Korean media outlet ET News, SK Hynix plans to enhance HBM with new capabilities, including computation, caching, and network memory. To support this development, SK Hynix has already invested in securing essential Intellectual Property (IP).

By 2025, SK Hynix targets the rollout of its sixth-generation HBM (HBM4), featuring integrated system semiconductors such as memory controllers. From the seventh generation (HBM4E) onwards, SK Hynix plans to embed computing, caching, and network memory functions.

Traditionally, memory controllers managing HBM functions are installed separately. SK Hynix intends to place the memory controller at the bottom layer of the HBM stack using a base die, which will enhance power efficiency and signal transmission speed by integrating the controller within the HBM.

Furthermore, SK Hynix is exploring the optimization of HBM by adding computing, caching, and networking capabilities, potentially embedding some GPU or CPU functions within HBM.

Sources indicate that the inclusion of caching and network memory aims to boost AI semiconductor computational speed. These enhancements are expected to be customized based on specific customer requirements.

Currently, AI chips typically position the GPU at the center, surrounded by HBM, with system semiconductors such as memory controllers added separately. This configuration can lead to signal transmission delays and higher power consumption due to inter-chip gaps.

Integrating various system semiconductor functions within HBM is anticipated to lower latency, reduce power consumption, and enhance overall spatial efficiency, resulting in more compact and efficient designs. With Samsung Electronics and Micron entering the HBM market, SK Hynix aims to sustain its market leadership through a differentiation strategy.

SK Hynix has reportedly secured several system semiconductor IPs for HBM. The manufacturing of the base die will be outsourced to wafer foundries like TSMC. SK Hynix also plans to leverage relevant IP from IC design clients and wafer foundry partners to bolster its offerings.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext