Rapidus develops glass panel-level packaging for high-performance artificial intelligence chips

Japanese foundry Rapidus has built a glass-based panel-level packaging prototype aimed at high-performance artificial intelligence hardware and plans to push it toward mass production by 2028, positioning itself as a potential rival to TSMC.

Japanese foundry Rapidus is positioning itself for stronger competition with TSMC by developing a panel-level packaging prototype based on a glass interposer, with the company planning to move towards mass production by 2028. The prototype, which will be showcased at SEMICON Japan 2025 in Tokyo, replaces conventional round silicon wafers with large square glass panels designed for high-performance artificial intelligence packages that combine multi-chip GPUs and more than a dozen of HBM dies. This strategy is presented as a way to reduce material waste and enable larger, denser multi-chip assemblies required for next-generation artificial intelligence accelerators without breaking the traditional laws of 300 mm wafer.

Rapidus is working on 600×600 mm glass panels, which are described as providing a significantly larger usable area than 300 mm silicon wafers and allowing more interposers to be produced from a single sheet. The company highlights that glass substrates can offer advantages in electrical performance and flatness when compared to organic package materials, making them attractive for dense interconnect structures. The shift from circular wafers to large square panels is also framed as an efficiency gain, as more packaging units could be patterned per panel, potentially lowering costs and improving throughput for complex chip assemblies.

At the same time, the article notes that glass introduces new manufacturing challenges, including fragility and potential warpage as panel sizes grow. These hurdles underscore why the technology is still at the prototype stage and why careful process development will be required before mass production. The article points out that Intel has notably discontinued its in-house development of glass substrates and instead has chosen to license the technology to others. The company can thus monetize its innovations, but large-scale Intel-based technology production is not anticipated until after 2030, suggesting that widespread deployment of glass substrate packaging across the industry remains several years away.

58

Impact Score

OpenAI launches Artificial Intelligence deployment consulting unit

OpenAI has created a new consulting and deployment business aimed at helping enterprises build and roll out Artificial Intelligence systems. The move mirrors a similar push by Anthropic and signals a broader effort by model providers to capture more of the enterprise services market.

SK Group warns DRAM shortages could curb memory use

SK Group chairman Chey Tae-won warned that customers may reduce memory consumption through infrastructure and software optimization if DRAM suppliers fail to raise output. Demand from Artificial Intelligence data centers is keeping the market tight as memory makers weigh expansion against the long timelines for new fabs.

BitUnlocker bypasses TPM-only Windows 11 BitLocker

Intrinsec disclosed BitUnlocker, a downgrade attack that can bypass TPM-only Windows 11 BitLocker protections with physical access to a machine. The technique abuses a flaw in Windows recovery and deployment components and relies on older trusted boot code.

Micron samples 256 GB DDR5 9200 MT/s RDIMM server modules

Micron has begun sampling 256 GB DDR5 RDIMM server modules built on its 1-gamma technology to key ecosystem partners. The company positions the new modules as a higher-speed, more power-efficient option for scaling next-generation Artificial Intelligence and HPC infrastructure.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.